Trusted Real Estate: chatbot pre-qualification and automated contract drafting
Trusted Real Estate is a Dubai brokerage that needed to move faster. PHII Labs rebuilt their website on Tilda and added an AI layer that pre-qualifies buyers in chat and generates PDF contracts automatically.
Context
A real estate brokerage in Dubai operates on speed. A lead comes in from a web form or a listing portal, and the first agent to call wins. The problem is that most leads are not ready to buy — they are browsing, comparing, or asking about a property that is already sold. Agents spend hours calling leads who will never convert.
Trusted Real Estate came to PHII Labs with a Tilda website that captured leads into a form. The form dumped names and phone numbers into a spreadsheet. Agents called them in order. There was no pre-qualification, no contract automation, and no way to handle leads outside business hours.
Problem
Two bottlenecks constrained the business:
- Lead quality was opaque. Agents called every lead the same way, whether the person was a serious buyer with financing in place or a tourist curious about a penthouse. The time-to-first-call was the only metric, and it was too slow because agents were calling the wrong people.
- Contract drafting was manual. When a deal closed, an agent spent 30 to 45 minutes typing up a contract in a Word template — buyer details, property details, price, payment schedule, terms. This was dead time that could be automated, and the manual process introduced errors that required re-drafting.
Approach
We kept the website on Tilda — the team knew it, the SEO was established, and rebuilding the CMS would have cost time for no gain. The AI layer was built separately in Python and embedded into the Tilda site via a widget.
The chatbot is an OpenAI-powered agent that engages every site visitor who opens the chat. It asks a structured set of questions: budget, timeline, financing status, residency status, property type of interest. It classifies the lead — hot, warm, cold — and writes the result back to the CRM with the full conversation attached. Hot leads get flagged for immediate call. Cold leads get an automated nurture sequence.
For contracts, we built a PDF generation pipeline using PDFKit. When a deal is ready, the agent fills a short form with the deal parameters. The system pulls property data from the listing database, buyer data from the CRM, and generates a formatted PDF contract with all fields populated. The agent reviews it, not drafts it.
What was built
- Tilda website: redesigned for conversion, with the chatbot widget embedded on every listing page.
- Chatbot pre-qualification: an OpenAI agent that conducts a structured intake conversation, classifies the lead, and routes hot leads for immediate follow-up.
- CRM integration: lead data and conversation history written to the brokerage's CRM automatically.
- PDF contract generator: a Python service using PDFKit that produces a formatted, populated contract from deal parameters and listing data.
- SEO layer: content and metadata optimization on the Tilda site to improve organic search visibility for Dubai real estate queries.
Results
- Lead-to-call time down 17%. Pre-qualification meant agents stopped calling cold leads first. Hot leads, flagged by the chatbot, got called within minutes instead of hours.
- SEO traffic up 21%. The content and metadata work on the Tilda site improved organic rankings for target queries, increasing inbound lead volume without paid spend.
- Agents saved ~12% of weekly time. The contract generator eliminated the drafting step. Agents reviewed generated contracts instead of writing them from scratch, freeing time for prospecting and client calls.
Takeaway
The highest-leverage AI in a brokerage is not the flashy part — it is the paperwork. A chatbot that pre-qualifies leads is useful, but the contract generator is what agents actually thank you for. It removes 30 minutes of typing from every deal, eliminates the errors that come from manual data entry, and lets agents spend that time on the work that generates revenue: talking to clients.