How HeadChannel Built an AI Acquisition Tool for Cohaesus Group

Cohaesus Group has been growing by acquiring niche digital agencies. As its ambitions grew, finding the right ones to approach became harder.

The Challenge

Scouting relied on manual research across scattered sources, and it was slow and prone to errors.

• Data gaps. Reliable, current information on small and mid-sized agencies was hard to find and pull together.
• Manual effort. Finding, evaluating and comparing candidates took a great deal of time.
• Inconsistent evaluation. With no standard framework, agencies were prioritised on strategic fit and commercial potential in different ways.

The Approach

HeadChannel built Teabase, an AI-led data collector, with one product owner and one AI developer. It runs on n8n, a low-code automation platform, with MS SQL and a PG Vector store.

• Finding agencies. OpenAI and Anthropic models identify digital agencies online, extract data from their websites and categorise them by services, technology and market positioning.
• Company records. Registration details and financial records come straight from the Companies House API.
• Consistent profiles. Large language models turn unstructured content into structured profiles in the same shape for every agency.
• HubSpot. Everything syncs to HubSpot, where the Cohaesus team uses the profiles for outreach, scoring and deal progression.

The Outcome

Research that took days now takes minutes, a 75 to 85% gain in efficiency. Teabase runs on its own with daily updates, so the pipeline stays current, and the team compares every target on the same structured data.