Assess property enquiries on WhatsApp
A system that replies to property enquiries and scores the conversation to help a salesperson decide whom to follow up with. AI assesses the messages; code applies the scoring weights and business rules.
What it had to do
Property enquiries arrive on WhatsApp at all hours, from portals, ads and forwarded numbers. Most are not buyers. Some are brokers checking prices, some are the same person enquiring 3 times, and a few are not real at all.
Supply in this market is tight, so the expensive problem is not generating enquiries. It is finding the few people who will actually transact, before a salesperson has spent a day on the rest. The brief was to answer one question for each enquiry, in the minutes after it arrives: is this worth a call today?
How it decides
Every conversation is scored out of 100 on 5 things, from how specific the person is about what they want to whether the budget matches what the project costs. The model assesses the conversation; code applies the weights, thresholds and business rules. That second half is ordinary arithmetic — readable, arguable, and impossible to talk past.
Where the model's judgment is too close to a threshold to trust, the enquiry goes to a person rather than quietly down a tier.
The product
This one became a product of its own, with its own site. What it does in full is explained there rather than here.
Need software for a job like this? Tell me what your team does now and what you want to change.
Email me about the work Book a call Twenty minutes, on a call