Finding every credible buyer for a sell-side deal

An investment bank running a sell-side process needed to know every company that might realistically buy its client, who inside those companies would make the call, and in what order to approach them. It needed all of it fast, and it needed it to the standard a banker would sign.

I started by putting research agents on comparable deals, to understand what buyers had actually paid for and why. From there I built a universe of 530 potential acquirers across 16 categories, ranked into three outreach waves, with the exact roles to contact at each one and a specification ready to load into the bank's enrichment tools.

Alongside it, I built a signal agent that reads acquisition press releases to find companies that are actively buying, and pulls out the decision makers on the buying side. Two different AI models check every match before it counts.

A buyer universe the bank could start working the same day, with the roles to contact at every one of 530 companies, and a repeatable way to spot acquirers while they're still active.

Claude research agents, Perplexity, DeepSeek, Clay.