A generic assistant
Knows the name and the industryAsk a question and this side answers with nothing but the business name.
Portal enquiries arrive without a budget, a timeline, or any sign of whether the person can actually transact. Qualifying that before Saturday is the difference between four productive viewings and four wasted ones.
Portal leads · WhatsApp · Missed calls · Viewing coordination · Listing updates
Not one dramatic failure — a few at each step, in the specific places this kind of business loses them.
The same portal enquiry goes to several agencies at once. Whoever replies first has the conversation; everyone else is following up on someone already talking to a competitor.
A name and a listing id. No budget, no timeline, no configuration, no idea whether they are buying or renting — so the first call is spent on questions that could have been answered on arrival.
Viewings arranged before the requirement is understood produce polite no-thank-yous. Agent time is the scarcest thing an agency has and it is routinely spent on guesses.
Someone moving in eight months is not a bad lead — they are an early one. Chased hard they disappear; ignored entirely they buy through whoever was still sending them useful listings.
Nothing burns trust faster than travelling to see something already gone. It also makes every other listing you show look uncertain.
The enquiry arrives from a portal, a social post or a board outside a building.
Within seconds rather than hours, because the same enquiry went to three other agencies.
Buying or renting, budget range, areas, configuration, and — the question that decides everything — when they actually want to move.
Genuinely available properties that fit. If the budget does not work for the area, that is said at the start rather than discovered on a Saturday.
Grouped into one trip, proposed against owner availability, and confirmed rather than promised.
Negotiation, legal questions, loan eligibility and anything involving money go to a person with the full requirement attached.
Requirement recorded, and matching listings sent at a frequency they agreed to. No chasing someone who is eight months out.
A buyer today is a seller or a landlord later. The requirement and the relationship both stay on file.
This is what the system has to know to be useful here — and, just as importantly, what it must refuse to do.
Anchor Realty is an illustrative real estate business. Ask it anything — the same model answers twice, once knowing only the name, once given the business. The second answer is the one your customers would get.
Ask a question and this side answers with nothing but the business name.
…and this side answers with everything in the panel on the left.
Both answers come from the same model, in real time. The tags under the second answer are the system events that would follow in a live deployment — this page cannot write to a CRM or a calendar.
These are the sentences that stall a sale in this business. A system that has never been told how you answer them will answer them badly.
Enquiries from portals, WhatsApp and missed calls answered in seconds, at any hour.
You are the agency having the conversation rather than the one leaving a voicemail.
Budget, area, configuration and timeline captured in the first conversation.
Agents spend their time on people who can transact, with the brief already written.
A budget that does not work for an area is flagged honestly at the start.
Nobody spends a Saturday discovering it, and the agency keeps its credibility.
Three or four matched properties arranged for a single visit, subject to owner confirmation.
More properties seen per hour of agent time, and a better experience for the buyer.
Requirement on file, matching listings only, at a frequency they consented to.
The eight-month buyer is still yours in month eight.
Negotiation, title questions, loans and payments routed to the area consultant.
Nothing that carries legal or financial weight is answered by software.
The numbers that tell you whether it is working in this business. No figures are shown until they come from your own systems.
How long a person waits for their first real answer.
Measured from Timestamp of inbound message to timestamp of first reply.
Which channel each enquiry came from — website, WhatsApp, Instagram, phone.
Measured from Source tagged on the conversation at the moment of capture.
The share of enquiries where enough was learned to know whether this is a real prospect.
Measured from Conversations reaching the qualification fields your profile defines.
Enquiries that became a booked appointment, viewing, table or call.
Measured from Confirmed calendar bookings ÷ qualified conversations.
Bookings that actually turned up. A booking nobody attends is not a result.
Measured from Attendance you confirm, against bookings made.
People who did not book, but where intent and consented contact details were captured and a follow-up is possible.
Measured from Conversations tagged as not-ready with consented contact details.
Conversations passed to your team, and why. A rising handoff rate is a signal, not a failure.
Measured from Escalation events, grouped by trigger.
What counts as a win here: A qualified viewing with a buyer who can actually transact — and an agent's Saturday well spent.
Limits are part of the build, agreed before anything goes live. In this industry they are these.
Working in a different industry? The approach is the same — we map how your business wins customers before recommending anything.
A free 30-minute look at how your business attracts, captures, converts, serves and retains customers. No obligation, and the recommendation is not always automation.