Results
The challenge
The developer sells villa projects in Bengaluru, with a busy in-house sales team and steady paid-ads spend. Lead supply was not the problem. Follow-up was.
An enquiry landing at 6 PM was usually called back the next morning, more than fourteen hours later. Missed calls sat in the phone system with no owner. Days passed between the first call, the project walkthrough and the site visit, and buyers forgot earlier conversations or toured two or three rival projects the same weekend. A single sale could take 45 to 60 days, and managers had no clear view of who had spoken to whom.
They wanted every lead answered fast, every conversation remembered, and the whole team working from one system, not a telephony dashboard plus spreadsheets.
What we built
Arthat built a dedicated AI sales system around how the team already sells, and shipped it in release waves.
- One pipeline for every source. Google and Meta lead forms, website enquiries and inbound calls land in one CRM, are scored and are routed to the right salesperson automatically.
- Every call understood. The cloud-telephony line is connected end to end. Each call is matched to the right lead and employee, and recorded conversations are transcribed, translated and summarised. Missed calls become leads with an owner instead of disappearing.
- An assistant for the team. Salespeople and managers ask for daily briefings, answers about projects, inventory and deals, or add leads from a photo or a list. Exports stay admin-only.
- Follow-up that doesn't wait. A first reply is built in, ready to go out the moment an enquiry arrives. Managers build their own When → If → Then automations across email, WhatsApp and tasks, with quiet hours, opt-out and per-customer limits.
- Human in the loop. AI in the sales cycle starts in suggest-only mode with human approval, and each customer-facing automation is switched on deliberately.
The system runs on its own isolated server and database, and every release passes more than 10,000 automated app, server and database checks.
How it shipped
From a 45-day chase to a managed journey
Before, a buyer's path looked like this: enquiry at night, a call the next morning, a week of missed calls, a walkthrough, another week before a visit, and then a comparison with rival projects. Each gap gave the buyer time to forget and gave competitors time to step in.
Now every enquiry has an owner from the first minute. The call history and summary sit on the lead record, so the salesperson picks up from what was actually said last time. Managers see the whole pipeline, calls included, in one place.
What's live today
The system runs in production for the sales team. Leads from ads, the website and calls arrive in one CRM and are routed to a salesperson. Calls on the sales line are matched to the right lead, then transcribed, translated and summarised. Site visits, the pipeline and each salesperson's performance sit in one view, and an AI assistant briefs the team every day and answers questions on projects, inventory and deals. Customer-facing automations, from the first reply to visit reminders, are built in and switched on project by project as the team approves them.
What made it work
- Built around their calls, not a generic CRM. Most real-estate selling happens on the phone, so call understanding came first.
- Their languages. Buyers and staff mix English with Kannada and other Indian languages, so drafts and call handling cover five languages.
- Safe by default. Nothing reaches a customer until the team turns that automation on.
Your turn
Losing buyers between the enquiry and the site visit? Talk to Arthat and we'll show you where AI can close the gaps in your sales process within weeks.
Arthat AI
Want results like these at your company?
We built this for them. We can build something similar — or something completely different — for you. Start with a 30-minute discovery call and we'll tell you honestly whether AI is the right tool for your situation.
