An AI agent system that finds sales leads in public registries

The situation

The furniture manufacturer's sales team was searching for new commercial-space projects by hand. Planning records, building permits, developers' backgrounds and the right contacts were scattered across different sources and registries, and putting each lead together took hours. Much of that time went not to selling, but to gathering information, and the best projects often reached competitors before they were even found.

The solution

We built an AI agent system that does this background work itself. The agent monitors public registries, spots upcoming development projects, identifies the actual developer behind them, enriches the information with financial data and decision-makers' contacts, and presents it all on a clear dashboard.

At the core of the system is an AI agent that is connected to a database and the registries, and that works continuously rather than as a one-off.

How the system works

1. Monitors the registries. The agent goes through the public sources and finds new, upcoming development projects.
2. Identifies the developer. It links the project to the actual developer and adds background.
3. Enriches the data. It adds financial information and decision-makers' contacts, so the sales team has a complete picture.
4. Presents it on a dashboard. It pulls everything onto one clear dashboard with source references, reviewable at a glance.

The system is built with Claude Code and uses PostgreSQL and Supabase to store the data and Python to process the registries.

Why an AI agent, not a simple workflow

Finding a sales lead is not "take it from here, put it there." The agent has to understand which project is genuinely getting under way, who is behind it, and who the right decision-maker is. That calls for understanding the content, not just moving data around. That is exactly why this is an AI agent, not a simple automated workflow: the agent makes decisions that a rule-based workflow couldn't.

The result

Background work that used to take hours now reaches the sales team automatically and in verified form, complete with source references, so every lead is reliable and quick to review. The sales team can focus on selling instead of hunting for information, and reaches projects earlier.

The same approach can be used anywhere sales leads are scattered across different sources and gathering them takes time. If your team spends hours hunting for information, let's talk about it and we'll look at what could be automated the same way in your business.