SynthesisLogic

AI agent for real estate agents, brokers and investors

An agent that answers property enquiries from your portal, website and WhatsApp within minutes, qualifies buyers and tenants on budget and timing, books viewings in your agents' calendars and keeps in touch until they decide.

Book an intro call How it works
A morning with this agent
  1. Replied to 17 overnight enquiries about 6 listingsDone
  2. 4 leads qualified as ready to view this weekDone
  3. 3 viewings booked in the agent's calendarDone
  4. A buyer asks to negotiate the priceAsks a person

An example. Your agent is built around your own work.

How it works

  1. Catch the enquiry

    Messages from listing portals, website forms, WhatsApp and email land in one place with the property they asked about. The agent pulls that listing's details, such as area, price and status, so it can answer about the right unit.

  2. Qualify with a few questions

    It asks buy or rent, budget range, timing, financing status and preferred areas, and stores each answer in its own field. Enquiries outside your areas, and spam, are closed with a polite reply.

  3. Book the viewing

    For serious enquiries it offers open slots from the agent's live calendar, books the one the person picks and sends the address and the agent's contact details. A reminder goes out before the viewing.

  4. Keep in touch and hand over

    People who are not ready get follow-ups at the pace you set, and new listings that match their wishes. Questions about price negotiation, offers, legal terms or mortgages go to the agent, with the conversation summarised.

New to AI agents? How they work, in plain words

Real examples, setup steps and tool connections for this agent

An example conversation

An example of how a conversation runs. The grey lines show what the agent does in your systems meanwhile.

  1. CustomerHi, is the 2-bedroom on Marina Street still available? Saw it on the portal.
  2. Matched the enquiry to listing MS-204 in the CRM: 2 bedrooms, listed price 1,450,000, status available. Read the notes: viewings handled by Dana.
  3. AgentYes, the 2-bedroom at Marina Street is available at the listed price of 1,450,000. Are you looking to buy, and roughly when would you like to move?
  4. CustomerBuying. Within three months. Can it be less? Maybe 1,300,000?
  5. AgentI can only give you the listed price. A lower offer is something Dana can discuss with you directly. Do you have a budget range and a mortgage in place, so I can tell her what to prepare?
  6. CustomerBudget up to 1.4 million, mortgage pre-approved.
  7. Saved to the lead record: buy, budget up to 1.4M, three months, financing pre-approved. Read Dana's calendar: free tomorrow 11:00 and 16:30, Thursday 10:00.
  8. AgentThank you. Dana can show you the apartment tomorrow at 11:00 or 16:30, or Thursday at 10:00. Which suits you?
  9. CustomerTomorrow 16:30.
  10. Booked the viewing in Dana's calendar and sent a confirmation with the address and her number. Passed the summary and the offer question to Dana.
  11. AgentBooked for tomorrow at 16:30 at Marina Street. Dana has your question about the price and will talk to you at the viewing.

How it is built

Enquiries from every source
Website forms and email arrive through webhooks or the Gmail API and Microsoft Graph, WhatsApp through the official Business Platform. Most listing portals only share enquiries with approved partners, so we use the portal's own feed, its email notifications or a CSV export. We do not scrape portal pages.
The listing is the source of truth
Prices, area, status and features are read from your listing records or sheet at the moment of the question, not recalled by the model. Whatever the agent says about a property has to match the listing. If something is not in the listing, it says so and asks the agent.
A language model talks, your rules score
A large language model, the same kind of AI behind ChatGPT and Claude, understands the enquiry and keeps the conversation natural, in the buyer's language. Your rules decide the score from the answers: budget, timing, property type and area. Scoring never uses personal characteristics, and we review the fields with your broker.
Viewings only from live calendars
The agent books into Google Calendar or Outlook calendars through their APIs, and only into slots your agents have marked free. It checks the calendar again at booking time, so two viewings never land on the same hour.
A fixed list of actions
The agent may look up a listing, ask qualifying questions, update the CRM, offer and book a slot, send confirmations and reminders, and notify the agent. It cannot quote a price that is not listed, accept or pass on an offer, or give legal or mortgage advice.
Your agents stay in control
Every conversation and every action is logged and visible in the CRM. An agent can step into any chat, take it over and hand it back. We review the logs with you weekly at first, looking at response times, viewings booked and leads that went cold.
Where it runs, and buyer data
The system runs in the cloud or on your own servers. Buyer budgets and financing status are sensitive, so the agent, its database, its logs and the language model can all run on your servers, using open models such as Llama, Qwen or Mistral. Local open models are smaller than the largest cloud models. For answering from your listings, sorting enquiries and drafting replies they are usually enough, and we test on your own enquiries first. WhatsApp and email still pass through those services.

What you can connect to it

Salesforce, including Real Estate Cloud
Creates leads, contacts and opportunities through the REST API, and logs each conversation and viewing on the record.
HubSpot
Receives form submissions, creates contacts, assigns the agent and updates the deal stage.
Listing portals
Takes enquiries through the portal's approved feed, email notifications or CSV exports, and links each to the listing.
Your listing database or website
Reads price, area, features and status for each property when a buyer asks.
Google Calendar and Outlook calendar
Reads each agent's availability, books viewings and sends reminders.
WhatsApp Business Platform
Answers enquiries in the chat and sends approved template reminders for viewings.
Gmail and Outlook
Reads enquiry mail, replies to it and sends viewing confirmations.
Property management software
For managers, reads tenant requests and lease dates and opens maintenance tickets or renewal reminders.
Google Sheets or Excel
Holds leads, listings and the viewing log for small agencies without a CRM.
Slack or Microsoft Teams
Tells the agent about a booked viewing, with the summary of what the buyer asked.
Book an intro call

Connects to your tools. You stay in control.

Works with what you have
Salesforce, HubSpot, listing portals (through approved feeds or exports), your listing database, Google Calendar, Outlook, WhatsApp Business Platform, Gmail, property management software, Slack or Teams, spreadsheets.
A person approves what matters
Payments, orders, discounts or anything you choose wait for approval. Every action is logged.
Can run on your own servers
The agent, its data, its logs and even the language model (open models such as Llama, Qwen or Mistral) can run on your servers, so documents and customer data stay inside your network. How that works

What it does not do

  • It does not negotiate price, accept or reject offers, or promise that a property will not be sold.
  • It does not give legal, mortgage or tax advice. Those questions go to a licensed person.
  • It does not state anything about a property that is not in the listing, and it does not guess prices or rents.
  • It does not use protected characteristics to score or steer buyers, or describe neighbourhoods by who lives there.
  • It does not publish AI-written marketing for off-plan units unless a sales manager approves it.

Questions

Does it work for developers and property managers too?

Yes. Developers use it for off-plan enquiries across several projects, writing to each project's own unit list and payment schedule, with a sales manager approving any change to price or availability. Managers use it for tenant requests, viewings for lets and renewal reminders.

Will it discuss prices?

Only listed prices and terms you give it. If a buyer asks for a lower price, the agent says so plainly and passes the question to a person, with the buyer's budget attached.

Can it help real estate investors?

Yes. It can screen new listings against criteria you set, such as area, price and yield, and send a short list with the numbers. Any rent or return figure is an estimate from the stated inputs, not a valuation, and the inputs are shown next to each result.

Is this an AI agent for real estate agents?

Yes. It takes over the first reply, the questions about budget and timing, and the booking, so the agent spends time at viewings. Individual agents can use it on their own WhatsApp number and calendar.

How fast will it answer an enquiry?

Within minutes, at any hour, if your systems let it see the enquiry at once. We measure your current first-response time before launch, so the logs show your own before and after rather than a figure from another agency.

What stops it from saying something wrong about a property?

It answers from the listing record read at that moment, and says it does not know when the listing is silent. Prices, availability and legal statements always come from the data or a person. Every reply is logged, and we fix the listing data or the rule when a mistake shows up.

Do we need a CRM?

No. Small agencies can run it on a shared spreadsheet for leads and a Google or Outlook calendar for viewings, and we can help you move to a CRM later.

Can the AI run on our own servers?

Yes. The agent, its database, its logs and the language model can run on your servers with open models such as Llama, Qwen or Mistral, so buyer finances stay with you. Those models are smaller than the largest cloud ones, but for answering from your listings they are usually enough, and we test on your own cases first. WhatsApp and email still carry the messages.

How much does an agent like this cost?

It depends on the task and the systems it connects to. After a first call we give you a clear estimate of cost and time.

Want this agent built around your work?

Tell us in a few lines what the job looks like today and we will set up an online intro call.

Type of organisation