# Serviti AI — reviewed business guides Owner: Serviti AI. These guides describe Serviti and general business workflows; they are not facts about another tenant. Product capability depends on the plan and configured integrations. Current pricing: https://serviti.ai/pricing. No live booking or automatic outbound customer messaging is claimed. Scenarios marked hypothetical are illustrations, not customer results. --- ## Missed customer enquiries: the revenue leak you can fix A missed enquiry is not always a missed call. It can be a website question nobody answers, a message buried under supplier chats, or a customer who receives a price but no useful next step. For a small business, those gaps matter because you have already done the work of attracting the person. Before spending more on ads, look at what happens between the first question and your team's reply. ### Find the gaps before buying more traffic Check enquiries from your website, campaign links, WhatsApp and offline touchpoints separately. Do they reach a shared record, or live on one person's phone? Who handles a question after closing time? Can someone else continue when the original employee is away? A visitor may only need your opening hours, delivery area or service options. If those answers are available only when the owner is free, a simple question becomes a delay. Do not count every missed conversation as a lost sale. Some visitors are researching; others are not a fit. Track the number of useful enquiries, the contacts captured and the outcomes your team records. ### Run a one-week enquiry audit You do not need a tool to start. You need one week of honest records. 1. **List every channel an enquiry can arrive through.** Website form, website chat, WhatsApp, phone, Instagram, email, QR codes, walk-ins. Include the ones nobody officially owns. 2. **For each channel, name the person responsible** and the hours they cover. Write "nobody" where that is the truthful answer. 3. **Log every enquiry for seven days** with four fields: when it arrived, which channel, when a person first replied, and whether contact details were captured. 4. **Mark each one** as answered within the hour, answered later the same day, answered the next working day, or never answered. 5. **Separate the useful from the rest.** A supplier pitch, a job application and a wrong number are not lost revenue. 6. **Count how many useful enquiries ended without a recorded next step** — no callback agreed, no owner, no note. That last number is usually the uncomfortable one. It is also the one you can act on without spending anything. #### What to look for in the results - **Clustered gaps.** Enquiries arriving after closing, at weekends, or during your busiest trading hours often share one cause. - **One person carrying a channel.** If every WhatsApp enquiry depends on one phone, that is a single point of failure, not a process. - **Answered but not captured.** A helpful reply with no contact detail recorded is work you cannot follow up. - **Repeated questions.** The same five questions appearing all week are the content your first response should already cover. ### What the research says about response speed The most cited study on this is *[The Short Life of Online Sales Leads](https://hbr.org/2011/03/the-short-life-of-online-sales-leads)*, published in Harvard Business Review in March 2011. It found that firms contacting a potential customer within an hour of receiving a query were "nearly seven times as likely to qualify the lead" as those who tried an hour later, and more than 60 times as likely as those who waited 24 hours or longer. Among companies that responded within 30 days at all, the average first response took 42 hours. Read that with its limits in mind. The research covered U.S. companies, it was published in 2011, and it measured whether a lead could be *qualified* — reaching a decision-maker and establishing fit — not whether a sale closed. Your market, your price point and your buyers' habits may differ. What travels across those differences is the direction: the gap between a reply in minutes and a reply tomorrow is not a small one, and most businesses are slower than they believe. The audit above tells you where you actually sit, which is more useful than any benchmark. ### Give the first response a clear job A useful first response does three things: - Answers the question using current business information. - Asks a relevant follow-up when the request is unclear. - Offers a practical next step, such as a callback request or a link to the right service. An assistant should not guess a price, stock level or delivery date to keep a conversation moving. A specific request for confirmation is more useful than a confident promise your team cannot honour. #### An illustrative example *The following is a hypothetical scenario written to show the difference, not a customer case study.* A furniture retailer runs ads that send people to a product page. A visitor asks at 9:40 PM whether a sofa comes in grey. **Without a first response:** the question sits until the shop opens. The buyer messages two competitors the same evening and has an answer from one of them before breakfast. **With one:** the assistant answers from the retailer's own product information, asks which room size they are working with, and offers to take their number for a callback about delivery to their area. The next morning a salesperson opens a lead that already contains the product, the room size and the question still outstanding. The second version has not made a sale. It has produced a conversation a person can continue — which is the realistic thing to aim for. ### How Serviti helps Serviti's website AI assistant answers from the business content you provide. You can share its agent link through ads, WhatsApp messages and QR codes, or embed the widget on your website. When a visitor provides contact details and enquiry intent, Serviti can capture a lead with conversation context. Your team uses the lead CRM to manage status, ownership and follow-up. A QR scan or page visit alone is not a captured lead. Notifications depend on the plan. Starter includes a daily missed-lead summary and weekly digest email; Growth adds real-time email and WhatsApp lead alerts and human WhatsApp handoff. Lead Intelligence and deeper insights are Scale features. Check [current plan details](https://serviti.ai/pricing) before choosing. These are team workflows. Serviti does not send automatic outbound nurture messages to your customers, and a WhatsApp alert reaching your team is not the same as an AI replying on your WhatsApp number. ### Put a number on the opportunity Use the [missed enquiry calculator](https://serviti.ai/missed-enquiry-calculator) with your own enquiry volume, missed-response rate, conversion rate and customer value. The result is an estimate of revenue exposure, not a forecast or a promise of recovered sales. Then review one week of actual conversations. Look for repeated unanswered questions, late replies and leads without an owner. Fix one gap and measure again. ### Start with your most useful information Before you share an assistant link with anyone, make sure it can answer what people actually ask: - Services or products, described the way customers name them - Prices, with the conditions attached to them - Opening hours, including exceptions - Locations, service areas and delivery limits - How to reach a person, and when - The five questions from your audit that came up most Then test it yourself with the awkward questions — the ones about price, availability and timing where a wrong answer costs you. If it guesses, fix the underlying information rather than the wording. ### Common questions **How quickly should we reply to an enquiry?** Faster is better, and the honest answer for most small businesses is "sooner than we do now." Rather than adopting someone else's target, measure your current median first-response time in the audit and improve it in stages. **Does every unanswered message represent lost revenue?** No. Some enquiries are researchers, some are out of area, some are not buyers at all. This is why the audit separates useful enquiries from the rest — inflating the number makes it easier to dismiss. **We reply quickly already. Is there anything here for us?** Possibly the capture step. Businesses that answer fast often still lose the enquiry because no contact detail was recorded and no owner was assigned, so nothing happens after the first exchange. **Will an AI assistant replace our front desk?** It should not. The useful division is that an assistant handles the repetitive, always-on part — answering known questions, capturing details, recording context — so people spend their time on the conversations that need judgement. **What should we fix first?** The channel with the most useful enquiries and the least clear ownership. That is usually where a small process change returns the most, and it costs nothing to test. Read next: [organise WhatsApp enquiries](https://serviti.ai/blog/whatsapp-ai-lead-capture-small-business) and [build a follow-up routine](https://serviti.ai/blog/automated-lead-follow-up-small-business). Source: https://serviti.ai/blog/missed-customer-enquiries-cost-small-business --- ## Real estate lead leakage: where buyers disappear before a site visit A property enquiry can go cold long before a buyer rejects the project. The buyer asks about a configuration after office hours, receives an unrelated brochure, or waits for a callback that never comes. The problem is not always lead quality. Sometimes the process loses the information that would help the sales team recognise a useful enquiry. ### Follow the enquiry across channels List the routes into your project: website, ad landing pages, channel-partner links, brochures and hoardings. Check where each route sends the buyer and where the resulting enquiry is recorded. A contact number without the project, configuration or original question is a weak handover. It makes the buyer repeat themselves and leaves pre-sales guessing where to start. Use tagged agent links or QR codes for distinct campaigns. Attribution records the source when that information is passed through; it does not automatically identify every offline interaction. ### Trace one enquiry end to end Reading a funnel report tells you the totals. Walking a single enquiry through your own process tells you where it breaks. Do this once per route, with a real enquiry or a test you submit yourself. 1. **Submit the enquiry the way a buyer would** — through the hoarding QR, the ad landing page, the channel-partner link, the website. Use a real question, not "test". 2. **Note the timestamp** and what the buyer sees immediately afterwards. Is there any acknowledgement, or silence? 3. **Find where it landed.** A CRM record, a shared inbox, a spreadsheet, one executive's phone. Write down the actual destination, not the intended one. 4. **Check what arrived with it.** Project, configuration, budget, timeline, the original question, the campaign source. Tick only the fields genuinely present. 5. **Time the first human response** and record who made it. 6. **Ask that person what they knew** before they called. If they had to ask the buyer to repeat the question, the handover failed regardless of response time. Run this for every route and you will usually find one or two that lose the context entirely while still producing a phone number. #### The fields that decide whether a callback is useful | Field | Why the next executive needs it | |---|---| | Project and configuration | Prevents quoting the wrong unit type | | Original question | Lets the call start where the buyer left off | | Approximate budget | Only if the buyer offered it; do not infer | | Timeline and purpose | Separates an active buyer from early research | | Preferred contact time | Raises the chance the call is answered | | Campaign source | Makes spend comparisons meaningful | If fewer than four of these reach the lead record, the leak is in capture, not in follow-up. ### Ask enough to make the next call useful Qualification should feel like a conversation, not an application form. Start with the question the buyer asked, then collect relevant preferences: - Project or locality of interest. - Configuration and approximate budget. - Buying timeline and whether the property is for self-use or investment. - Preferred contact method and a convenient time. Do not interpret a short number as a budget or a survey reference without context. Ask what it means. #### An illustrative comparison *The following is a hypothetical scenario written to show the difference between two handovers, not a customer case study.* Two enquiries arrive on the same evening from the same hoarding QR code. **The first** reaches pre-sales as a name and a number. The executive calls the next morning, opens with "you enquired about our project", and spends the first two minutes establishing which project and what the buyer wanted. The buyer has since visited two other sites. **The second** reaches pre-sales with the buyer's question about east-facing 3BHK availability, their stated timeline of "within six months", a preferred callback on Saturday morning, and the hoarding location as source. The executive calls on Saturday and opens with the answer to the question that was actually asked. Neither enquiry has been sold anything. The second one has a conversation worth having; the first has to be rebuilt from nothing. ### Keep project facts attached to their project Prices, carpet areas and possession dates need clear ownership. Put the project, tower or configuration beside the fact in your content. A comparison with another developer's price is not your own price. A website knowledge base is not a live inventory feed, a title check or an authoritative property-record lookup. When current availability or a specific record needs checking, the agent should explain the boundary and direct the buyer to your team or an approved service. ### Where Serviti fits Serviti answers from approved project content, captures enquiry details and gives your team a lead record with conversation context. The team can manage follow-up from the lead CRM rather than relying only on an executive's personal notes. An agent can collect interest in a site visit and the buyer's preferred timing. **That is a visit request, not a confirmed booking.** Your team must check the schedule and confirm it. Growth includes real-time lead alerts and human WhatsApp handoff. Scale adds Lead Intelligence and supported integrations. Review [plans](https://serviti.ai/pricing) for the current capabilities. ### Measure the handover, not just the contact count Track response time, contactable enquiries, visit requests and visits actually confirmed by staff. Record outcomes consistently before comparing campaigns. A useful monthly view separates four numbers that are often collapsed into one: - **Enquiries received** — everything that arrived, including duplicates and wrong numbers - **Contactable enquiries** — those with usable details and a genuine property interest - **Visit requests** — buyers who asked for a time - **Visits confirmed by staff** — the schedule actually accepted them The gap between the third and fourth number is where most reporting quietly overstates performance. ### Common questions **Is lead leakage the same as poor lead quality?** No, and conflating them is expensive. Poor quality means the enquiries were never going to buy. Leakage means usable enquiries lost their context or their owner. The trace exercise above separates the two — if executives are calling people who asked real questions but arriving without the question, that is leakage. **We use a CRM already. Can we still be leaking?** Yes. A CRM records what reaches it. If the hoarding QR opens a chat that captures only a phone number, the CRM will faithfully store an incomplete lead. Check what arrives, not whether something arrives. **Should we ask for budget up front?** Only if the buyer volunteers it. Demanding a budget before answering a question tends to end the conversation, and an unexplained number in a chat is not reliable data. Capture it when offered and mark it as approximate. **How do we attribute enquiries from hoardings and print?** Use a distinct tracked link or QR per placement. That records the source when the buyer uses it. It will not capture people who see the hoarding and search for you separately — offline attribution is always partial, and treating it as complete will mislead your spend decisions. **What is the difference between a visit request and a booking?** A request is the buyer telling you when they would like to come. A booking is your team confirming that slot against the schedule. Showing a buyer a confirmation before staff have accepted it creates exactly the kind of broken promise that loses the sale. Read next: [WhatsApp lead management](https://serviti.ai/blog/whatsapp-lead-management-real-estate) and [real estate follow-up](https://serviti.ai/blog/real-estate-lead-follow-up). Source: https://serviti.ai/blog/real-estate-lead-leakage --- ## Missed enquiries at schools and clinics: a front-desk problem A school office or clinic reception handles people who are already there while new enquiries arrive on the phone and website. Both deserve attention, but one person cannot answer everything at once. The aim of an AI assistant is not to replace that person. It is to answer routine questions from approved information and preserve the requests that need a human response. ### Separate information from decisions Schools can publish clear information about curriculum, the admission process, documents, transport routes and campus visiting arrangements. Clinics can publish administrative information about services, locations, opening hours and how to request an appointment. Some questions still require confirmation. A school place is not guaranteed because eligibility guidance matches. A doctor's published hours do not prove an appointment slot is available. An assistant should make those differences explicit. ### Build your approved-answers list first Most front-desk assistants fail because nobody decided what the approved answer actually is. Before configuring anything, spend an hour building this list with the people who answer the phone today. **Step 1 — collect the real questions.** Ask reception to note every question for three days. Not categories. The actual wording, including the vague ones. **Step 2 — sort each one into three buckets:** | Bucket | Meaning | Who owns it | |---|---|---| | **Published** | The answer is fixed and already public | Assistant can answer directly | | **Conditional** | The answer depends on grade, branch, service or period | Assistant answers with the condition stated | | **Confirm** | Requires staff judgement or a live check | Assistant captures the request only | **Step 3 — write the conditional answers with their condition attached.** "Fees for Grade 6 at the Kondapur branch, academic year 2026–27" is usable. "Fees" is not. **Step 4 — name an owner per topic.** Transport routes, fee structure, admission steps, clinic timings, doctor availability. When something changes, that person updates the source. **Step 5 — write the refusal wording** for the confirm bucket. What the assistant says when it cannot answer matters more than what it says when it can. A short, accurate list beats a long one. Twenty correct published answers and a clean handover for everything else is a working front desk. ### Make the content usable Keep each fee beside the programme or service it belongs to. State the academic year or effective period where relevant. If branches have different hours or transport coverage, name the branch in the same section. A well-organised short set of accurate answers is more useful than a long document mixing old policies with current ones. Review what the assistant says when the answer is missing. It should identify what needs checking and offer an appropriate next step, not make up a rule. ### Capture a request without over-collecting For an admission enquiry, the intended grade, residential area and a parent or guardian's preferred contact may be enough to start. For a clinic appointment enquiry, collect the minimum administrative details the team needs to respond. Do not ask visitors to post detailed medical histories into a public sales chat. The assistant should not diagnose, prescribe or treat a symptom conversation as a sales opportunity. Clinical questions belong with qualified staff. #### Why less collection is better here Over-collecting creates three problems at once. It lowers completion, because people abandon long forms. It creates a record you then have to protect. And in a clinic context it invites people to type health details into a channel that was never designed to hold them. Ask for what the responding staff member genuinely needs to make the next contact useful. Everything else can be asked by a person, later, in an appropriate setting. ### How Serviti supports the front desk Serviti can answer from your school's or clinic's approved content, capture enquiry intent and contact details, and keep the context in the lead CRM. You can embed the assistant on your website or share its link through a campaign or QR code. Human WhatsApp handoff and real-time lead alerts are Growth features; supported BSP integration is a separate Scale capability. A campus visit or appointment collected in chat is a **request for staff to confirm**, not an automatically reserved slot. AI HelpDesk is a separate add-on for support enquiries, not a live student or patient-record system. ### Review the unanswered questions Look for recurring gaps: a missing transport area, an unclear document requirement or absent location directions. Ask the responsible staff member to update the approved content. Make this a fixed fortnightly review rather than an occasional clean-up. Fifteen minutes with the list of questions the assistant could not answer is the single most useful maintenance habit, because it is driven by what families and patients actually asked rather than what staff assumed they would ask. ### Common questions **Will an assistant make our front desk look impersonal?** It depends what you give it. An assistant that answers "what documents do I need for Grade 6 admission" accurately at 9 PM is more helpful than an unanswered phone. The risk is not automation — it is an assistant that guesses, or that makes a family feel processed. Test the refusal wording as carefully as the answers. **What happens when a parent asks something the assistant cannot answer?** It should say what needs checking and capture the request for staff. That is a successful outcome, not a failure. An assistant that produces a plausible-sounding rule about eligibility or fees is doing real damage. **Can it confirm an appointment or a campus visit?** No. It captures the preferred time as a request. Staff check the schedule and confirm. Anything that shows a family a confirmation before a person has accepted it will eventually produce someone arriving to an empty slot. **Is it safe for a clinic to use this?** For administrative questions — services, timings, locations, how to request an appointment — yes, provided the boundaries hold. It must not diagnose, must not prescribe, and must route clinical questions to qualified staff. Avoid designing any flow that invites people to describe symptoms. **How much content do we need before launching?** Less than most people expect. The published and conditional buckets from your three-day question log, written accurately, are enough to start. Add to it from the unanswered-question review rather than trying to anticipate everything. Read next: [WhatsApp enquiries for schools and clinics](https://serviti.ai/blog/whatsapp-ai-schools-clinics) and [a practical follow-up process](https://serviti.ai/blog/enquiry-follow-up-schools-clinics). Source: https://serviti.ai/blog/missed-enquiries-schools-clinics --- ## WhatsApp AI lead capture: turn chats into useful enquiries WhatsApp is a familiar place for customers to ask a business about prices, opening hours or a product they have seen. The challenge is turning that interest into a record your team can act on. Before buying automation, separate three different needs: answering questions, handing a conversation to a person, and recording what happened. ### A WhatsApp link is not a native WhatsApp bot A link shared in a WhatsApp message can open a hosted AI conversation in the browser. A website handoff button can open a conversation with your team on WhatsApp. A Business Solution Provider integration can capture supported WhatsApp activity. Those are different flows. Do not assume a WhatsApp icon means an AI assistant automatically replies inside every WhatsApp chat. #### The routes, side by side | Route | Where the conversation happens | Who replies | What gets captured | |---|---|---|---| | Native WhatsApp AI add-on *(paid, Growth/Scale)* | WhatsApp | Your configured AI agent, with team takeover | Conversation and captured enquiry details; replies use the plan allowance | | Tracked agent link shared in WhatsApp | Browser, hosted agent | The assistant | Question, details the visitor shares, campaign source | | Website widget | Your website | The assistant | Same, plus page context | | Human WhatsApp handoff *(Growth)* | WhatsApp | Your team | Whatever your staff record | | Supported BSP integration *(Scale)* | WhatsApp | Depends on the supported workflow | Supported events only | A tracked agent link opens browser chat; human handoff opens a chat with your team; the native add-on answers inside WhatsApp. Choose the route you need rather than treating every WhatsApp icon as the same capability. ### Start with a clear route Choose the route that matches your team: 1. Share a tracked Serviti agent link when customers need business information before speaking to someone. 2. Let the website assistant answer from approved content and offer a human handoff when available. 3. If you use a supported BSP, confirm the integration and its supported events before relying on it for conversation capture. A click-to-WhatsApp ad opens WhatsApp. With the WhatsApp AI Agent add-on connected to that business number, Serviti can reply inside WhatsApp. Without the add-on, a person or your existing system handles the message. A tracked agent link remains a separate route to Serviti's browser-based chat. #### A pre-launch test script Run this on a phone, not a desktop, before you put a link in any campaign. Fifteen minutes here prevents the most common launch failures. 1. **Open your own link from a WhatsApp message.** Does it load quickly on mobile data? Is it obvious what this is? 2. **Ask a price question with a condition** — a product that costs different amounts in different sizes. Does the answer keep the condition attached? 3. **Ask about something not in your content.** Does it say what needs checking, or does it produce a confident guess? 4. **Type a single vague word**, like "price" or "delivery". Does it ask a useful clarifying question? 5. **Share your details when asked.** Then check what your team actually received — the question, the product, the contact detail, the source. 6. **Trigger the human handoff** if you are on Growth. Does the visitor understand they are moving to WhatsApp and talking to a person? 7. **Ask for something it cannot do**, such as cancelling an order. Does it decline clearly and offer a real next step? Failures on steps 3 and 7 are the ones that cost you customers. A wrong confident answer is worse than no answer. ### Capture context, not just a number When a visitor is willing to share details, preserve the question, product or service of interest, preferred next step and contact information. Ask one useful question at a time. A person typing "price?" may need a particular product, not a generic price list. If the product is unclear, clarify it before quoting. Prices and delivery promises should come from current business content or be confirmed by staff. A chat transcript is not proof that stock has been reserved. ### What Serviti provides Serviti combines a website or hosted AI assistant with lead capture and a lead CRM. Tracked links can be shared through WhatsApp campaigns and QR codes. Starter includes a daily missed-lead summary and weekly digest email. Growth adds real-time email and WhatsApp lead alerts and human WhatsApp handoff. Scale adds supported bidirectional BSP integration and Lead Intelligence. Native replies require the separate WhatsApp AI Agent add-on on Growth or Scale and a connected business number. Alerts, handoff and BSP integration do not enable native AI replies by themselves. Nor do team follow-up reminders mean automatic outbound customer messages. See the [WhatsApp feature guide](https://serviti.ai/whatsapp-lead-capture) and [plan comparison](https://serviti.ai/pricing). #### Matching the plan to how you work The plan difference that matters most is not a feature list, it is how quickly a person finds out a lead exists. On **Starter**, captured leads are available in the dashboard, with a daily missed-lead summary and weekly digest email. These are scheduled summaries, not instant alerts. If enquiries need an immediate staff response, consider Growth or Scale. On **Growth**, real-time email and WhatsApp alerts reach the team as enquiries arrive, and the assistant can hand a conversation to a person on WhatsApp. If your enquiries are time-sensitive, this is the line worth crossing. **Scale** adds Lead Intelligence and supported BSP integration — relevant once you have enough enquiry volume that prioritising between them is a real problem. ### Keep the human handover practical Tell visitors who will contact them and avoid promising a response time your team cannot meet. Give leads an owner and record the outcome after the conversation. A stated response window you consistently meet builds more trust than an ambitious one you miss. "We reply during business hours, usually within a few hours" is better than "instant response" if instant is not true. ### Common questions **Can Serviti reply automatically on our WhatsApp number?** Yes, with the separate WhatsApp AI Agent add-on on Growth or Scale and a connected WhatsApp Business number. The AI replies inside WhatsApp, and your team can take over. Website chat, tracked links, handoff and alerts remain distinct capabilities; the base plan does not include native replies. **Our ads use click-to-WhatsApp. Does that feed the assistant?** A click-to-WhatsApp ad reaches the assistant when the destination business number is connected with the WhatsApp AI Agent add-on. Otherwise it reaches your existing WhatsApp conversation flow. A tracked Serviti link can instead open the website agent. Test the actual ad destination and do not assume every DM includes campaign attribution. **Will it send follow-up messages to customers who go quiet?** No. Reminders go to your team, prompting a person to follow up. There is no automatic outbound customer sequence, and you should not describe one to your customers. **What if the assistant gets a price wrong?** Check the source content, retrieval and configuration as well as the answer. An old page or a price without its conditions can cause errors, but the model can also misread valid evidence or retrieve the wrong subject. Correct the cause and re-run steps 2 and 3 of the test script. **Do we need a QR code as well as a link?** Only where people are physically present — a counter, a package, a vehicle, a standee. The QR is just a way of delivering the same tracked link. Use a distinct one per location if you want to compare them. Read next: [stop losing enquiries](https://serviti.ai/blog/missed-customer-enquiries-cost-small-business) and [organise follow-up](https://serviti.ai/blog/automated-lead-follow-up-small-business). Source: https://serviti.ai/blog/whatsapp-ai-lead-capture-small-business --- ## WhatsApp lead management for real estate: from enquiry to a visit request Property buyers often want project details they can read later or share with family. WhatsApp can be useful for that exchange, but messages alone do not make a managed sales process. A useful system keeps the buyer's question, project interest and next step together, even if a different executive handles the follow-up. ### Avoid the brochure-and-silence pattern Sending a large brochure may not answer the buyer's actual question. Start with the configuration or project they asked about, then point them to the relevant information. Before offering a price, confirm which project and unit type it refers to. Before discussing possession, use the approved project-specific source. Do not infer current availability from an old marketing page. ### Distinguish the routes into WhatsApp A click-to-WhatsApp ad opens a WhatsApp chat with your business. A tracked Serviti link opens Serviti's hosted AI conversation. A WhatsApp handoff from the widget opens a human conversation. Plan your campaign around the route you actually use. A QR code can link to your agent, but scanning it alone does not capture the person's contact details or book a visit. #### Choosing the route per campaign Different placements deserve different routes, and the decision is usually about who needs to be available. **Hoardings and print** work well with a tracked agent link behind a QR. People scan these in the evening and at weekends, when your site office is closed, and the assistant can answer configuration and location questions immediately. **Performance ads for an active launch** can use click-to-WhatsApp with the WhatsApp AI Agent add-on on Growth or Scale. Connect the destination business number for native replies and lead capture, with pre-sales available to take over. Without that add-on, staff or your existing system must reply. Check campaign attribution on the actual route rather than assuming every DM contains its source. **The project website** should use the widget, with human handoff available on Growth for buyers who want a person. **Channel-partner links** need distinct tracking per partner, or you will not be able to tell which relationship is producing enquiries. Decide this before the campaign goes live. Changing the destination mid-flight makes the source data uncomparable. ### Build a useful buyer handover Capture preferences that help the team respond: - Project and configuration. - Budget range, if the buyer wants to share it. - Buying timeline and purpose. - Preferred visit time or callback request. - Contact details and the original question. Keep uncertainty visible. "Interested in Sunday" is different from "Sunday confirmed." #### A handover template Whatever system you use, the record an executive opens before calling should answer five questions without them having to ask the buyer: ``` Project / configuration : .............................. What they actually asked : .............................. Timeline & purpose : ............ (as stated, not inferred) Preferred contact time : .............................. Source : .............................. Status : request | contacted | visit requested | visit confirmed by staff | closed ``` The status line matters more than it looks. "Visit requested" and "visit confirmed by staff" being the same field value is how developers end up reporting site visits that never happened. ### What Serviti can do Serviti answers from the project content you supply and can capture a buyer's enquiry in the lead CRM. Campaign links help retain source context. On Growth, the agent can offer human WhatsApp handoff and the team can receive real-time lead alerts. Scale includes Lead Intelligence and supported BSP integration. A BSP connection does not by itself turn every WhatsApp conversation into a native AI conversation. **Serviti can collect a site-visit request; it does not automatically reserve a live calendar slot.** Staff must check availability and confirm the visit. Follow-up reminders for your team are not an automatic customer messaging sequence. #### An illustrative exchange *The following is a hypothetical conversation written to show the boundary between a request and a confirmation, not a customer transcript.* A buyer asks on Saturday evening about 3BHK availability and says they could visit on Sunday. **Overpromising:** "Your site visit is confirmed for Sunday at 11 AM." Nobody has checked the schedule. If Sunday at 11 is already full, the buyer arrives to a queue or an apology. **Honest:** "I've noted your interest in a 3BHK and a preferred visit on Sunday morning. Our sales team will confirm the time with you — they work from the live schedule. Is this the best number to reach you on?" The second version sets an expectation the team can meet. It also gives pre-sales a reason to call on Sunday morning with something useful to say. ### Keep the company record up to date Assign the lead, record whether the buyer was reached, and update the outcome after a visit. That makes the next interaction useful and reduces dependence on one executive's memory. For post-sale support, the AI HelpDesk add-on can capture issues and answer approved FAQs. It should not claim to have checked a buyer's payment or title records without an actual supported connection. ### Common questions **Should we run click-to-WhatsApp ads or send traffic to the agent?** It depends on the connected channel and your campaign. Click-to-WhatsApp can reach the native AI agent when its add-on is enabled for your business number; otherwise your team or existing system must reply. A tracked agent link opens a browser conversation. Test answers, lead capture and available attribution on the route you choose. **Does a BSP integration mean AI replies on our WhatsApp number?** BSP integration alone does not activate AI replies. Native customer conversations require the separate WhatsApp AI Agent add-on on Growth or Scale and a connected business number. Your team can take over when the buyer needs a person. **Can the assistant tell a buyer which units are still available?** Only if that is in your approved content and current. A website knowledge base is not a live inventory feed. For anything that changes daily, the assistant should say it needs confirmation and route the buyer to your team. **How do we stop leads living on one executive's phone?** Route enquiries through a tracked link or the widget so they land in the CRM with context, and make the status field mandatory. The technical part is easy; the discipline of recording outcomes is what actually fixes it. **What should we measure?** Four separate numbers: enquiries received, buyers actually reached, visit requests, and visits confirmed by staff. Collapsing these into "leads" is what makes campaign comparisons meaningless. Read next: [real estate lead leakage](https://serviti.ai/blog/real-estate-lead-leakage) and [the follow-up checklist](https://serviti.ai/blog/real-estate-lead-follow-up). Source: https://serviti.ai/blog/whatsapp-lead-management-real-estate --- ## WhatsApp AI for schools and clinics: give enquiries a clear path Parents and patients may message outside office hours, when the front desk cannot reply. A useful enquiry flow gives them approved information and a way to leave a request without implying that someone has already confirmed it. The first step is deciding where the AI conversation happens and where the human conversation begins. ### Choose the right channel flow A Serviti agent link shared in WhatsApp opens a hosted AI conversation in the browser. The website widget can also offer human WhatsApp handoff on Growth. Native replies are available through the separate WhatsApp AI Agent add-on on Growth and Scale. Connect your eligible business number so the agent can answer inside WhatsApp and staff can take over. Scale's supported BSP integration is a different capability; it does not activate this add-on. ### For schools: answer, then capture intent Start with the questions families ask repeatedly: curriculum, application steps, documents, published fees, transport and campus visits. Keep grade-specific and branch-specific details together in the knowledge base. Ask for the intended grade or branch when necessary rather than giving a rule from a different context. Collect only the details needed for the admissions team to respond. A visit preference is a request for staff to confirm, not an accepted admission or reserved appointment. ### For clinics: keep the role administrative The assistant can explain published services, opening hours, locations and the appointment-request process. It must not turn a published schedule into a claim of live availability. Staff should confirm the clinician, date and time. Clinical questions need qualified staff; the agent should not diagnose or prescribe. Avoid collecting sensitive health information just to arrange a callback. ### Keep the enquiry out of a personal inbox Serviti can capture contact details and enquiry context in the lead CRM. Give the request an owner, record the response and leave notes needed for the next staff member. Starter includes a daily missed-lead summary and weekly digest email. Growth adds real-time email and WhatsApp lead alerts and human WhatsApp handoff. Scale adds Lead Intelligence and supported integrations. The AI HelpDesk add-on handles support requests separately. Team notifications and reminders support staff follow-up. They are not automatic appointment reminders or outbound messages to every parent or patient. ### Test the awkward cases before launch Try questions about an unlisted branch, an old fee, a full appointment schedule and a missing transport area. Check that the assistant asks for clarification or explains what staff need to confirm. Also test the handoff on a phone. The visitor should understand whether they are leaving details in Serviti or opening WhatsApp. #### The pre-launch test script Have two people run this — one who knows the content and one who does not. The second person finds more problems. **For a school:** | Test | What a good response looks like | |---|---| | "What are the fees?" | Asks which grade and branch before answering | | "Is my son eligible for Grade 6?" | Explains the published criteria; does not decide eligibility | | Ask about a branch you do not have | Says it does not cover that location; offers the right contact | | Quote last year's fee back at it | Does not confirm an outdated figure | | "Can I visit on Saturday?" | Captures it as a request for staff to confirm | | Ask about transport to an area not covered | Says it needs checking rather than inventing a route | **For a clinic:** | Test | What a good response looks like | |---|---| | "What are your timings?" | Answers from published hours, names the location | | "Is Dr X free tomorrow at 4?" | Explains published hours are not live availability | | Describe a symptom | Does not diagnose; routes to qualified staff | | "Can you book me in?" | Captures an appointment request for staff to confirm | | Ask about a service you do not offer | Says so plainly | | Offer unprompted medical history | Does not encourage it; asks only for what is needed to respond | The symptom test is the one to run most carefully. An assistant that engages with a symptom description — even helpfully — is operating outside its role, and the failure mode is serious rather than merely embarrassing. ### Handle personal information with care Both settings involve information about people who did not choose to be in your marketing system: a child in a school enquiry, a patient in a clinic one. Three habits keep this proportionate: - **Ask for the minimum.** A grade and a parent's contact is enough to return an admissions call. A date of birth is not needed at enquiry stage. - **Do not design flows that invite health details.** If a field or a question can be answered with a symptom, someone will answer it with a symptom. - **Give the record an owner and a purpose.** An enquiry captured by nobody, for no stated reason, is a liability rather than a lead. Where your institution has its own policy or a formal process for handling family or patient data, the assistant sits inside that policy rather than alongside it. Involve whoever owns that policy before launch, not after. ### Common questions **Can the assistant reply on our clinic's WhatsApp number?** Native WhatsApp replies need the separate WhatsApp AI Agent add-on on Growth or Scale and a connected business number. A shared website-agent link still opens in the browser; human handoff and BSP capture are separate capabilities. Keep the agent focused on approved administrative information, with staff handling clinical or admissions decisions. **Does it confirm appointments or campus visits?** No. It captures a preferred time as a request. Staff check the schedule and confirm. This distinction is the single most important thing to get right, because a family or patient arriving for a slot nobody accepted damages trust far more than a slower reply would have. **What stops it giving medical advice?** Its role is administrative, and the content it answers from should contain no clinical guidance. Test it with symptom descriptions before launch and check that it routes to qualified staff rather than engaging. **Parents message at 10 PM. Is an automated answer appropriate?** Answering a published question — timings, documents, transport routes — at 10 PM is helpful. Making a decision at 10 PM is not. The line is between information and judgement, and the assistant should stay firmly on the information side. **How often should we update the content?** Whenever a fee, a schedule, a route or an intake changes, and on a fixed fortnightly review of questions it could not answer. Stale content is the most common source of wrong answers in both settings. Read next: [missed front-desk enquiries](https://serviti.ai/blog/missed-enquiries-schools-clinics) and [follow-up for admissions and appointments](https://serviti.ai/blog/enquiry-follow-up-schools-clinics). Source: https://serviti.ai/blog/whatsapp-ai-schools-clinics --- ## Automated lead follow-up: build a routine your team can keep "I'll think about it" does not always mean no. It may mean the customer needs a detail, another person's input or a better time to talk. Follow-up works best when the team remembers that context. Automation can help with capture, reminders and prioritisation, but it should not turn every enquiry into a generic message sequence. ### Separate reminders from customer messages A reminder tells your team that a lead needs attention. An outbound message contacts the customer. Those are different capabilities with different expectations. Serviti supports lead management and team follow-up workflows. **Serviti's team reminders are not an automatic outbound customer messaging service.** Do not promise a personalised message sequence merely because reminders are enabled. ### Give each lead a next step After the first conversation, record the customer's need, the open question, the person responsible and the intended next action. That action might be a phone call, a staff-confirmed visit or sending an answer the assistant could not verify. "Follow up later" is too vague to manage. An enquiry without usable contact details is not ready for a callback. An unanswered question should not be marked high intent just because the assistant offered one. ### Why the first reply matters more than the tenth The most cited research here is *[The Short Life of Online Sales Leads](https://hbr.org/2011/03/the-short-life-of-online-sales-leads)*, published in Harvard Business Review in March 2011. Firms that contacted a potential customer within an hour of receiving a query were "nearly seven times as likely to qualify the lead" as those who tried an hour later, and more than 60 times as likely as those who waited 24 hours or longer. Among companies that responded within 30 days at all, the average first response took 42 hours. Two caveats before you act on it. The study covered U.S. companies and was published in 2011. And it measured *qualifying* a lead — reaching someone and establishing fit — not closing a sale. The practical reading is not "reply within an hour or lose everything." It is that the decay is steep early and shallow later, so effort spent shortening your first response usually returns more than effort spent adding a fifth follow-up message. Most teams have this backwards. ### A practical routine Use this as a staff checklist, not a preconfigured automation schedule: - After an enquiry, answer the question and confirm what happens next. - If the customer is undecided, follow up with useful information relevant to their request. - If they propose a visit, confirm the details with the person managing the schedule. - If they decline further contact, record that and stop. - If they remain interested, agree on a sensible next contact rather than repeatedly asking for an update. Never invent a stock shortage, discount deadline or delivery promise to create urgency. #### The weekly follow-up review Twenty minutes, same time each week, with whoever owns leads. The agenda is fixed: 1. **Leads with no owner.** Assign them or close them. An unowned lead is not pending, it is abandoned. 2. **Leads with no next action.** Either write a specific action with a date, or mark it closed with a reason. 3. **Leads where the next action date has passed.** Do it now, reschedule it deliberately, or close it. 4. **Leads contacted three times with no reply.** Decide your stopping rule and apply it consistently. Continuing indefinitely is not persistence, it is noise. 5. **Closed-lost reasons.** Read them. If "no response" dominates, your problem is reachability, not persuasion. The purpose is not to work the list in the meeting. It is to make sure nothing is sitting in an undefined state, because that is where leads quietly die. #### An illustrative example *The following is a hypothetical scenario written to show the difference, not a customer case study.* A services business receives an enquiry asking whether they cover a particular area and what a typical job costs. **The generic sequence:** the customer receives a templated "thanks for your interest" message, then a case study two days later, then an offer on day five. None of these answer the question they asked. They do not reply, and are recorded as not interested. **The specific follow-up:** a staff member replies confirming the area is covered and gives a price range with its conditions. When there is no reply after two days, they send one message answering the question they expect comes next — availability — and stop there. The customer replies a week later, having discussed it at home. The second approach sends fewer messages. It works because each one is about the customer's actual question rather than the business's calendar. ### How Serviti supports the routine The AI assistant answers from your approved content and can capture enquiry details in the lead CRM. Staff can use the record, conversation context, ownership and status to continue the discussion. Growth adds real-time lead alerts and follow-up reminders for the team. Scale adds Lead Intelligence to help prioritise enquiries. These features support judgement; a score does not guarantee that a visitor will buy. Share tracked agent links through campaigns and QR codes so source information reaches the conversation. The agent responds in its supported chat surface; this does not mean it initiates a new WhatsApp conversation when a visitor disappears. ### Measure whether follow-up helps Review contactable leads, time to first human response and outcomes your team records. Compare similar periods and account for changes in campaign mix. Two measurements are worth more than the rest: - **Median time to first human response.** Use the median, not the average — one lead answered eleven days late will distort a mean and hide a decent typical performance. - **Share of leads with a recorded outcome.** If a third of your leads end in no status at all, your funnel numbers describe a fraction of reality. Fix those two before investing in anything more sophisticated. ### Common questions **Does Serviti send follow-up messages to customers automatically?** No. Reminders go to your team. There is no automatic outbound customer sequence, and you should not describe one to customers or prospects. **How many times should we follow up before stopping?** There is no universal number, but you should have one and apply it consistently. Three genuine attempts with something useful to say is a defensible rule. What matters more is recording the stop, so the lead is closed rather than silently abandoned. **Should we follow up leads that gave incomplete details?** If there is no usable contact method, there is nothing to follow up — that is a capture problem to fix upstream, not a follow-up task. Do not leave them in the pipeline inflating your numbers. **Is a lead score a prediction?** No. Lead Intelligence on Scale helps prioritise when you have more enquiries than time. It orders your queue; it does not tell you who will buy, and treating a score as a forecast leads to ignoring people who convert. **What if we simply do not have time to follow up?** Then reduce the number of leads you accept as open, rather than carrying a list you cannot service. Ten leads worked properly beat a hundred that each receive one unanswered message. Start with [the lead capture process](https://serviti.ai/blog/missed-customer-enquiries-cost-small-business), then decide whether your next bottleneck is response speed, missing information or staff ownership. Source: https://serviti.ai/blog/automated-lead-follow-up-small-business --- ## Real estate lead follow-up: a system for the weeks after enquiry A buyer may compare projects, discuss finance with family and revisit a location before deciding. One brochure and one missed call are rarely enough to understand where they are in that process. A good follow-up system helps your team continue the right conversation. It should not manufacture urgency or claim to know a buyer's intent from a single message. ### Stage one: understand the enquiry Record the project, configuration, approximate budget and questions still unanswered. Clarify whether a location or price refers to your offering before sharing details. Keep the original question with the lead. A phone number alone tells the next executive very little. ### Stage two: arrange a visit with a person Ask about a convenient time, then check it against the team's schedule. Send a confirmation only after staff have accepted the booking. Serviti can collect a **site-visit request**, including a preferred time. It does not automatically reserve a live slot. The distinction matters when several buyers ask for the same time. ### Stage three: record what happened After a visit, update the lead's status and note the configuration discussed, open questions and agreed next step. Do not leave a no-show marked as a completed visit. Do not infer a booking from a positive chat. Reporting is only useful when the underlying outcomes are recorded accurately. ### Stage four: follow up with relevant information Staff can answer the outstanding question, offer a second visit or share a genuine project update. Avoid generic broadcasts that ignore the buyer's requirements. A payment plan or possession date should come from current approved project information. Never invent limited availability or a price increase to force a decision. The appropriate timing depends on what the buyer requested. Record their contact preferences and stop unwanted follow-up. ### The lead record that makes stage four possible Stage four fails when the record from stages one to three is too thin to act on. A different executive, three weeks later, should be able to open the lead and know what to say. | Field | Recorded when | Why it matters later | |---|---|---| | Project and configuration | Enquiry | Prevents discussing the wrong unit | | Original question | Enquiry | Gives the next call an opening | | Budget (as stated) | Enquiry, if offered | Mark approximate; never inferred | | Timeline and purpose | Enquiry | Separates active from researching | | Visit requested / confirmed | Stage two | Two distinct statuses, never merged | | Visit outcome | Stage three | Attended, no-show, rescheduled | | Units discussed | Stage three | The specifics they reacted to | | Open question | Stage three | The reason for the next contact | | Contact preference | Any stage | Time, channel, and any opt-out | | Owner | Always | An unowned lead is an abandoned one | If your team cannot fill "open question" after a site visit, the visit was not debriefed and stage four has nothing to work with. ### Handling the awkward cases **No-shows.** Record them as no-shows on the day. A same-day message asking whether they would like to reschedule is reasonable; three chasing messages are not. If your no-show rate is high, look at how visits were confirmed rather than at the buyers. **Buyers who go quiet after a visit.** Silence after a visit usually means an unresolved concern — budget, location, a family member who has not seen it. One message addressing the most likely concern is more useful than a general check-in. **Buyers who ask about a competitor's price.** Answer about your own project. A comparison with another developer's number is not your price, and quoting it as context invites a conversation you cannot control. **Buyers who say no.** Record the reason and stop. A clean closed-lost with a real reason is worth more to your next campaign than a lead left permanently open. ### Stage five: know when to stop Every follow-up system needs a defined end. Without one, leads accumulate until the list is meaningless and executives ignore it. Agree a stopping rule the team applies consistently: a number of genuine attempts, a period of no response, or an explicit decline. Record the stop with a reason. A buyer who returns six months later is a new enquiry with useful history, not an unresolved task that has been open since March. Honour opt-outs immediately and permanently. Someone who asked not to be contacted and is contacted anyway will tell other people about it. ### Where Serviti fits Serviti's assistant handles questions from approved business content and captures enquiry context for the lead CRM. Tagged links and QR codes can preserve campaign source information. Growth includes real-time lead alerts, human WhatsApp handoff and team follow-up reminders. Scale adds Lead Intelligence and supported integrations. These are **team workflows, not automatic outbound buyer-nurture sequences**. Supported BSP integration can bring WhatsApp activity into the workflow, but it does not mean Serviti independently replies to every message on any WhatsApp number. ### Track a small, honest funnel Review enquiries received, buyers reached, visit requests, confirmed visits and outcomes entered by staff. Keep these separate rather than calling every contact a qualified visit. The ratio worth watching most closely is visit requests to confirmed visits. A widening gap there means your scheduling or confirmation step is failing, and no amount of additional lead generation will fix it. ### Common questions **How soon should we call a new property enquiry?** Sooner is better, and evenings and weekends matter in this category because that is when buyers research. Rather than adopting a target, measure your current median time to first human contact and improve it in stages. A tracked link that answers questions immediately buys you time; it does not remove the need for the call. **How many times should we follow up?** Agree a rule and apply it. Three genuine attempts with something useful to say, then a recorded stop, is defensible. Repeated "just checking in" messages train buyers to ignore your number. **Can Serviti send follow-up messages to buyers automatically?** No. Reminders go to your team. There is no automatic outbound buyer-nurture sequence, and staff remain responsible for every message a buyer receives. **A buyer says they are "interested in Sunday". Is that a booking?** No. It is a request. Staff check the schedule and confirm. Treating it as a booking is how buyers end up arriving at a site office that is not expecting them. **What should we do with leads from a campaign that has ended?** Work them to your normal stopping rule and close them properly. Do not recycle an old list into a new campaign as though it were fresh demand — your conversion figures will mislead you and the buyers will remember. Read next: [WhatsApp lead management for property teams](https://serviti.ai/blog/whatsapp-lead-management-real-estate) and [how to evaluate a real estate chatbot](https://serviti.ai/blog/real-estate-chatbot-vs-ai-sales-agent). Source: https://serviti.ai/blog/real-estate-lead-follow-up --- ## Enquiry follow-up for schools and clinics: helpful, not pushy A family leaves a campus with questions to discuss at home. A patient asks about an appointment but needs to check their schedule. Both interactions need a clear next step, not repeated sales messages. The front desk needs a reliable record of the request and who should respond. ### Start with a shared enquiry list Capture the enquiry, contact preference and unresolved question. Assign a staff member and update the status when they respond. Do not mix an enquiry with a confirmed appointment, admission or clinical record. Those represent different events and should not be inferred from a conversation. ### Decide who owns what, before you need to know Most front-desk follow-up fails at the handover, not the first reply. An enquiry arrives, someone answers the easy part, and the part needing a decision sits unclaimed. Write an ownership map once and put it where reception can see it. | Enquiry type | First response | Owns the outcome | Must not handle | |---|---|---|---| | Admission — general information | Reception or assistant | Reception | Eligibility decisions | | Admission — eligibility or exceptions | Admissions staff | Admissions head | — | | Fees and payment terms | Accounts | Accounts | Informal discounts | | Transport and routes | Transport coordinator | Transport coordinator | — | | Campus visit request | Reception | Whoever holds the schedule | Confirming without the schedule | | Appointment request | Reception | Reception | Clinical triage | | Clinical question | Qualified staff only | Qualified staff | Anyone else, including the assistant | | Complaint or service issue | Named staff member | Named staff member | — | Adapt the rows to your institution. The value is in two columns being different: the person who replies first is often not the person who owns the outcome, and writing that down prevents the enquiry stalling between them. ### A school follow-up checklist After an admission enquiry, provide the relevant grade and branch information. If the family wants to visit, ask for a preferred time and have staff confirm it. After the visit, record outstanding questions and point the family to the approved application process. Share a deadline only if it is current and applies to that grade or intake. Avoid collecting unnecessary information about a child before it is needed for the school's formal process. ### A clinic follow-up checklist For an appointment enquiry, clarify the administrative request and route it to reception for confirmation. A published doctor schedule is not live slot availability. Clinical follow-up, test results and treatment advice belong with qualified staff and the clinic's appropriate systems. Do not use a sales assistant as a substitute for them. Respect the person's contact preferences and avoid asking for sensitive medical details in a public enquiry form. ### What "not pushy" means in practice Both settings involve decisions people make carefully — where a child studies, which clinic they trust. Pressure tactics that might be tolerated elsewhere do real damage here, and families talk to each other. Some concrete boundaries: - **One follow-up per open question, not per day.** If they have not replied, they are deciding, not waiting to be reminded. - **No manufactured deadlines.** Share an intake cut-off only if it is real and applies to their grade. An invented "only two seats left" will be checked. - **No follow-up on a declined enquiry.** Record the decline and stop, permanently. - **No contact outside reasonable hours.** Answering at 10 PM is helpful; initiating at 10 PM is not. - **Nothing that implies a decision has been made.** "Your admission is progressing well" from someone with no authority to say so creates an expectation the school then has to withdraw. A family that felt respected during an enquiry will consider you again next year. One that felt chased will not. ### What automation should actually do Useful automation can keep the request from being forgotten: capture the lead, preserve the context and remind the responsible team. Automatic outbound messages require a separate supported workflow. A team reminder is not a patient appointment reminder, and a lead score is not a clinical priority assessment. ### How Serviti helps Serviti's assistant answers administrative questions from approved business content. It can collect admission or appointment interest and preserve contact details and conversation context in the lead CRM. Growth includes real-time lead alerts, human WhatsApp handoff and team follow-up reminders. Scale adds Lead Intelligence and supported integrations. AI HelpDesk is a separate add-on for support requests. Serviti does not automatically confirm appointments or send a scheduled series of messages to families or patients. Staff remain responsible for those next steps. ### Learn from the gaps Review questions the assistant could not answer. Ask the responsible team to update transport information, application steps, branch details or appointment instructions. Measure whether enquiries are being answered and handed over clearly, not just how many messages were generated. Three numbers are enough to run this well: enquiries received, enquiries with a recorded response, and enquiries closed with a reason. If the second and third numbers are much smaller than the first, the problem is ownership rather than volume. ### Common questions **How quickly should a school or clinic respond to an enquiry?** Within your stated hours, reliably, is better than fast and inconsistent. Publish a response expectation you can actually meet and hold to it. Families and patients tolerate a wait they were told about; they do not tolerate silence. **Can the assistant confirm an appointment or a campus visit?** No. It captures a preferred time as a request. Staff check the schedule and confirm. This is the boundary that matters most, because someone arriving for an unconfirmed slot loses trust immediately. **Does Serviti send appointment reminders to patients?** No. Reminders go to your team, prompting staff action. Patient-facing appointment reminders are a separate supported workflow and should not be promised on the basis of team reminders. **Should we follow up a family that visited but did not apply?** Once, with something useful — an answer to the question they left with, or the application steps. Then stop unless they respond. Repeated chasing about a child's education is the fastest way to be remembered badly. **What do we do with sensitive details someone volunteers anyway?** Do not carry them forward into the marketing record. Route the enquiry to the appropriate staff member and keep the lead record to what is needed administratively. Where your institution has a data policy, that policy governs — involve whoever owns it before launch. Read next: [WhatsApp enquiry flows](https://serviti.ai/blog/whatsapp-ai-schools-clinics) and [choosing an assistant for your front desk](https://serviti.ai/blog/chatbot-vs-ai-agent-schools-clinics). Source: https://serviti.ai/blog/enquiry-follow-up-schools-clinics --- ## AI chatbot vs AI sales agent: look at the workflow, not the label "Chatbot" and "AI agent" are used differently by different vendors. The name alone does not tell you whether a product can capture leads, book appointments or contact customers later. For a business website, the useful question is simpler: what happens after a visitor asks something? ### Start with answer quality A rule-based bot can be effective for a narrow set of predictable questions. A knowledge-based assistant can handle more natural wording, but still needs accurate source information. Test your own content. Ask about two similarly named services, a price with conditions and a detail that is not published. The assistant should distinguish them and explain missing information rather than inventing an answer. ### Follow the enquiry beyond the answer Check whether the product can: - Ask a relevant clarification. - Capture contact details when the visitor wants a next step. - Keep the original question and conversation with the lead. - Let staff manage ownership and status. - Provide a clear human handoff. - Retain source information from tagged campaigns. A product called a chatbot may offer these features. A product called an agent may not. Evaluate the working flow rather than assuming a category-wide checklist. ### An evaluation scorecard Take this into every demo and ask the vendor to demonstrate rather than describe. The right-hand column is what a satisfactory answer looks like — anything vaguer deserves a follow-up question. **Answers** | Ask them to show | A good answer | |---|---| | A price that has conditions attached | The conditions appear with the price, not separately | | Two similarly named products or services | They are distinguished without you clarifying | | Something absent from the content | It says what needs checking; it does not improvise | | A one-word message like "cost?" | It asks a useful clarifying question | | The same question in a second language or mixed script | Either it handles it, or it says plainly that it does not | **Capture** | Ask them to show | A good answer | |---|---| | The lead record after a test conversation | Question, context, contact detail and source all present | | What happens when someone declines to share details | The conversation still ends helpfully | | Campaign source from a tagged link | It arrives on the record automatically | | Duplicate enquiries from one person | You can see it is the same person | **Actions — the section where claims usually inflate** | Ask them to show | A good answer | |---|---| | "Book a demo" or "book an appointment" | They demonstrate whether a live slot is reserved or a request is captured | | Order or account lookup | An actual working integration, or an honest no | | Follow-up messaging | Whether messages reach *your team* or *your customer* | | WhatsApp | Which of the four possible routes this is | **Failure behaviour** | Ask them to show | A good answer | |---|---| | The knowledge source being wrong or outdated | It does not confidently repeat the error as fact | | A request outside its capability | It declines clearly and offers a real next step | | A visitor becoming frustrated or abusive | It stays useful and offers a human | If a vendor answers the actions section with "yes" four times without showing you anything, that is your result. ### Separate requests from completed actions An assistant saying "I can request a demo" is not the same as reserving a calendar slot. Collecting an order question is not looking up an order. Ask the vendor to demonstrate the actual integration and confirmation path for any action you need. Test failures as well as successes. Similarly, team reminders are not automatic outbound customer follow-ups, and a WhatsApp link is not a native WhatsApp AI responder. ### What Serviti provides Serviti is an [AI chatbot for business websites](https://serviti.ai/ai-chatbot-for-website) that answers from approved business content and captures enquiries for the lead CRM. You can embed it, share its hosted link or use a QR code. Growth adds human WhatsApp handoff and real-time lead alerts. Scale adds Lead Intelligence and supported integrations. AI HelpDesk is a separate add-on for support requests. Native WhatsApp AI Agent and Instagram DM AI Agent are separate paid add-ons for Growth and Scale, excluded from the trial. They answer incoming messages inside your connected WhatsApp Business number or Instagram account using approved business content. These native replies are different from WhatsApp alerts to your team, human handoff and links that open a website chat. Account connection and messaging requirements apply. Serviti can capture a preferred visit or callback time for staff. It does not claim automatic live booking or unsolicited outbound customer-nurture sequences. ### Run a small acceptance test Use genuine customer questions before publishing the widget. Include an incomplete message, a follow-up referring to an earlier question and a request outside the assistant's capabilities. Check what the customer sees and what the team receives. Verify the captured details rather than judging only the quality of the prose. ### Choose for the job you actually have If your problem is missed enquiries, focus on useful answers, reliable capture and clear staff follow-up. More autonomous features are not automatically better. A useful way to decide: write down the last ten enquiries you lost and why. If most were lost because nobody replied in time, you need answers and capture. If most were lost because the reply was wrong, you need better source content, not more automation. If most were lost after a good conversation, your gap is follow-up and ownership — and no widget fixes that. ### Common questions **Is an "AI agent" more capable than an "AI chatbot"?** Not reliably. The terms are marketing labels applied inconsistently across the industry. Some products called chatbots capture leads, manage ownership and hand off to staff; some products called agents only answer questions. Judge the demonstrated workflow. **How do we know if it will make things up?** Test it against a gap. Ask about something deliberately absent from your content and see whether it says the information needs checking or produces a confident-sounding answer. Do this before launch and again after every significant content change. **Does it need to book appointments to be worth having?** For most small businesses, no. Capturing a preferred time for staff to confirm solves the same problem with far less that can go wrong. Live booking only pays off when you have a reliable calendar system and enough volume that manual confirmation is the bottleneck. **Will it replace someone on our team?** It should not be bought on that basis. It handles the repetitive always-on part — known questions, capture, context — so staff spend their time on conversations needing judgement. If it is sold to you as a headcount replacement, ask what happens to the enquiries it cannot handle. **What is the most common mistake in choosing one?** Buying on demo quality. A demo is a controlled conversation with clean content. Insist on testing with your own content, your own awkward questions, and your own phone. Read next: [where enquiries go missing](https://serviti.ai/blog/missed-customer-enquiries-cost-small-business) and [a follow-up routine your team can maintain](https://serviti.ai/blog/automated-lead-follow-up-small-business). Source: https://serviti.ai/blog/ai-chatbot-vs-ai-sales-agent --- ## Real estate chatbot vs AI sales agent: test the buyer journey A property website can collect plenty of phone numbers while giving the sales team little useful context. Adding the word “agent” to the widget does not fix that. Evaluate the journey from a buyer's specific question to a staff-confirmed next step. ### Test project-specific answers Give the assistant questions about different projects, configurations and locations. Check whether it attaches each price, carpet area and possession date to the right offering. Then ask about something missing from the approved content. A useful assistant should say what needs confirmation, not borrow a number from a nearby paragraph or another project. A property assistant is not automatically a land-record or title-verification tool. Ask the vendor to demonstrate any live lookup separately. ### Look for useful qualification A strong conversation helps the team understand the buyer's project interest, configuration, timeline and preferred contact. It should not demand a long form before answering a simple question. It should also avoid treating an unexplained number as a phone, budget or survey identifier without enough context. Inspect the lead record: can the next executive see what was asked and what is still unresolved? ### A demo script you can run in thirty minutes Vendor demos are rehearsed on clean content and friendly questions. Take control of the session instead: send your own project material in advance, then work through this list in order. Ask to see the screen, not to hear a description. **Minutes 0–10: answers** | Ask the assistant | What a satisfactory result looks like | |---|---| | The price of a specific configuration in a specific project | The figure arrives with its conditions — floor rise, parking, charges — not alone | | The same configuration in your second project | A different, correct answer; no bleed between projects | | Possession date for a phase still under construction | The approved date, or a statement that it needs confirmation | | A carpet area you deliberately left out of the content | It says what needs checking; it does not estimate | | "Is it a good investment?" | It stays factual rather than making a return claim | **Minutes 10–20: capture and record** Ask the vendor to open the lead record created by that conversation while you watch. Look for the project and configuration, the buyer's actual question, anything they stated about timeline or purpose, the contact detail, the campaign source and an owner. If the record shows a phone number and a timestamp, the widget is a form with a conversational interface. **Minutes 20–25: actions** | Ask them to show | What a satisfactory result looks like | |---|---| | "Book me a site visit on Sunday" | They demonstrate whether a slot is reserved or a request is captured, and say which | | A WhatsApp journey | They name the route: tracked link, human handoff, or BSP integration | | Follow-up messaging | They state plainly whether messages reach your team or your buyer | | Any inventory or availability claim | A working data source, or an honest no | **Minutes 25–30: failure behaviour** Give it a one-word message. Give it a message mixing English and the local language. Ask about a competing project by name. Ask the same question twice with contradictory budgets. You are looking for an assistant that stays useful and honest under pressure, not one that improvises. If the vendor asks to postpone the second half because the content is not loaded, that is a result too — the first twenty minutes are the easy part. ### Verify what “book a visit” means Ask whether the assistant reserves a live calendar slot or only collects a preferred time. Both can be useful, but they must be described honestly. Serviti collects site-visit interest and preferred timing for staff to confirm. It does not automatically reserve a live slot. Do not show a buyer a confirmed visit unless the team's actual scheduling process has confirmed it. #### An illustrative comparison *The following is a hypothetical scenario written to show what the difference costs, not a customer case study.* Two developers run similar hoarding campaigns with a QR code. **The first** uses a widget that promises booking. A buyer scans on Saturday night, is told a Sunday 11 AM visit is confirmed, and arrives to a site office running a pre-booked group tour. The executive apologises and offers to reschedule. The buyer leaves having formed a view about how the project is run. **The second** captures the same request honestly: interest in a 3BHK, preferred Sunday morning, contact number, source recorded as the hoarding. Pre-sales call at 9 AM Sunday, confirm 12:30 PM and mention the configuration the buyer asked about. The visit happens once instead of twice. Both widgets "captured a lead". Only one of them produced a visit, and the difference was a claim, not a feature. ### Check channel claims carefully A QR code or ad can open a tracked Serviti agent link. A WhatsApp handoff opens a conversation with the team. Supported BSP integration is a separate capability. A click-to-WhatsApp ad reaches your native AI agent only when the destination number is connected with the separate paid WhatsApp AI Agent add-on on Growth or Scale. Otherwise it reaches your existing WhatsApp workflow. Ask to see the exact route your campaign will use. ### Where Serviti fits Serviti answers from approved project content and captures enquiry details in the lead CRM. Growth includes real-time alerts, human WhatsApp handoff and team follow-up reminders. Scale adds Lead Intelligence and supported integrations. Those reminders support the team; they are not automatic outbound buyer-nurture sequences. AI HelpDesk, a separate add-on, can capture support requests without claiming access to private payment or booking records. ### Measure the result you care about Compare useful enquiries, successful human contacts and visits actually confirmed by staff. Record outcomes consistently before attributing a change to the widget. Keep those four numbers separate for at least one full campaign before judging the tool. If cost per enquiry falls while confirmed visits stay the same, compare total spend and cost per confirmed visit: lower spend for the same visits can be a useful improvement. More low-quality enquiries alone do not prove better results. A well-run workflow is more valuable than a convincing demo that promises actions it cannot complete. ### Common questions **Is an "AI sales agent" better than a chatbot for property?** The labels are applied inconsistently across vendors and tell you nothing on their own. What matters for property specifically is whether project-level answers stay separated, whether the record reaching your executive is usable, and whether a visit request is distinguished from a confirmed visit. **Can it tell buyers which units are available?** Only if that information is in your approved content and kept current. A website knowledge base is not a live inventory feed. For anything that changes daily, the honest behaviour is to say it needs confirmation and route the buyer to your team. **What about RERA numbers, approvals and legal details?** Publish them in the approved content and let the assistant quote them exactly. Do not let it paraphrase, summarise or infer regulatory information, and do not treat it as a substitute for your compliance process. **Our buyers message in a mix of languages. Does that matter in evaluation?** Yes, and it is routinely skipped in demos. Test the actual wording your enquiries arrive in, including transliterated text. An assistant that handles clean English and fails on how people really type will underperform in exactly the campaigns you are buying it for. **How long before we can judge whether it works?** Long enough to see confirmed visits, not just enquiries — for most projects that means at least one full campaign cycle. Judging on enquiry volume in week one measures the widget's ability to open a conversation, which was never the bottleneck. Read next: [real estate lead leakage](https://serviti.ai/blog/real-estate-lead-leakage), [WhatsApp management](https://serviti.ai/blog/whatsapp-lead-management-real-estate) and [follow-up after the enquiry](https://serviti.ai/blog/real-estate-lead-follow-up). Source: https://serviti.ai/blog/real-estate-chatbot-vs-ai-sales-agent --- ## Chatbot for schools and clinics: choose a useful front-desk assistant Schools and clinics need an assistant that gives people clear information without making promises the staff cannot keep. Do not choose only by whether the product calls itself a chatbot or an AI agent. Test the questions, handovers and records that matter to your front desk. ### Can it answer from your information? For a school, test curriculum, application steps, published fees, transport and branch-specific rules. For a clinic, test services, opening hours, locations and appointment instructions. Include a question whose answer is absent. The assistant should identify what needs checking rather than make a plausible guess. Keep similar branches and services separate in the source content. A correct answer for one location can still be wrong for another. ### Does it distinguish a request from a confirmation? A family asking to visit has not booked a slot. A patient sharing a preferred time has not received an accepted appointment. Ask the vendor to show whether the system has live scheduling access or merely captures a request. Serviti captures appointment or campus-visit interest for staff to confirm; it does not automatically reserve a live slot. ### Are the boundaries appropriate? A clinic assistant should stay with administrative information and direct clinical questions to qualified staff. It should not diagnose or prescribe. For school enquiries, collect only the details needed to respond. Do not make admission decisions or guarantee eligibility from an incomplete conversation. A publicly accessible assistant should not imply it has checked private student, patient or payment records. ### A procurement checklist Institutions rarely buy this on one person's judgement, and the questions that decide the outcome are usually asked too late. Work through the four groups below before you shortlist, and note who in your institution owns each answer. **Content and accuracy — owner: whoever maintains your published information** - Who updates the assistant's source content when a fee, timing, route or intake changes? - How quickly does a change take effect, and who verifies it? - What does the assistant do when the content is silent — say so, or improvise? - Can branch-specific and grade-specific or service-specific information be kept separate? - Is there a record of questions it could not answer, and who reviews it? **Records and data — owner: whoever is accountable for family or patient information** - What exactly is stored about an enquirer, and where? - Can a record be corrected or deleted on request, and by whom? - Who inside your institution can see enquiry records, and is that access reviewed? - What happens to the records if you stop using the product? - Does anything in the flow encourage people to volunteer health or personal details you did not ask for? **Workflow — owner: the front desk lead** - Does every enquiry get an owner, a status and a recorded outcome? - Can reception see, at a glance, which enquiries have had no response? - What does a handover to a staff member actually look like on a phone? - Are visit and appointment requests visibly distinct from confirmed ones? **Commercial and rollout — owner: whoever signs** - Which claimed capabilities are on your plan, and which require a higher one or a separate add-on? - What is demonstrated working today versus described as coming? - Who trains reception, and what happens when that person leaves? - What is the exit: can you take your content and your enquiry records with you? Bring the four owners into one thirty-minute conversation before the demo rather than after. Most failed front-desk rollouts are decided in that meeting whether it happens or not. ### What reaches your team? Inspect the lead record after a test enquiry. The team should receive the contact information the visitor chose to share, the request and the relevant context. Check the human handoff on a phone. A visitor should know whether they are requesting a callback or moving into WhatsApp. #### An illustrative example *The following is a hypothetical scenario written to show where a front-desk rollout usually goes wrong, not a customer case study.* A school launches an assistant before the admission season. The content is accurate on launch day, and enquiries rise immediately. Six weeks in, the transport routes change and the fee structure for one grade is revised. Nobody was named as the content owner, so the assistant keeps giving last term's answers. Reception starts correcting it on calls, then stops trusting the enquiry records, then goes back to the notebook. The product is blamed; the missing owner is the actual cause. The version that works is dull: one named person, a fortnightly fifteen-minute review of unanswered questions, and a rule that any published change to fees, routes or intake updates the assistant's content the same day. ### How Serviti fits Serviti provides a website or hosted AI assistant connected to approved business knowledge and a lead CRM. Agent links can be shared through campaigns and QR codes. Growth adds human WhatsApp handoff, real-time lead alerts and team follow-up reminders. Scale adds Lead Intelligence and supported BSP integration. AI HelpDesk is a separate add-on for support requests. A BSP integration is not a promise of native AI replies on every WhatsApp number. Native replies require the separate paid WhatsApp AI Agent add-on on Growth or Scale and a connected business number. Team reminders are not automatic appointment reminders or outbound messages to families. ### Test before the public launch Ask staff to try incomplete questions, missing facts and requests requiring confirmation. Check both the visitor-facing response and the staff record. Choose the tool that makes the next step clear, not the one that makes the broadest automation claim. ### Common questions **Should a school or clinic buy the same product as a retail business?** The underlying assistant may be the same, but the evaluation is not. Both settings involve decisions people make carefully and information about people who did not choose to be in a marketing system, so accuracy, boundaries and record-keeping matter more than conversational polish. **Who should own this internally?** One named person for content, one for the front-desk workflow. A tool with no content owner drifts out of date within a term, and an assistant giving confidently wrong answers about fees or timings is worse than no assistant. **Can it handle admission eligibility or clinical triage?** No, and you should treat a vendor who says otherwise as a warning. It explains published criteria and routes the decision to the people qualified to make it. That boundary protects the institution as much as the enquirer. **What should we test that most buyers skip?** Three things: a question whose answer is deliberately absent from your content, an outdated figure quoted back at it, and the handover viewed on a phone rather than a laptop. Each one finds problems a scripted demo never surfaces. **How do we know it is working after launch?** Track enquiries received, enquiries with a recorded response, and enquiries closed with a reason. If the second and third numbers are much smaller than the first, inspect answer quality, contact capture, notification delivery and staff follow-up. The gap alone does not identify which part failed. Read next: [missed front-desk enquiries](https://serviti.ai/blog/missed-enquiries-schools-clinics), [WhatsApp enquiry flows](https://serviti.ai/blog/whatsapp-ai-schools-clinics) and [helpful follow-up](https://serviti.ai/blog/enquiry-follow-up-schools-clinics). Source: https://serviti.ai/blog/chatbot-vs-ai-agent-schools-clinics