Product Updates
George Rekouts
Co-Founder & CEO
AI TAM Assistant: Built from Real Searches
We took thousands of successful TAM searches created by hundreds of GTM teams using DiscoLike and had Claude Code extract the knowledge.
The data became the foundation of our AI TAM Query Assistant. You give it one sentence — the rough ICP description that lives in every founder’s head — and it returns a fully configured dense search vector, ready to return precise matches.
The Training Data Was Already There
A few hundred GTM teams use DiscoLike to build their target account lists. When they find the right results, they save and export them. Every saved ICP search by a serious GTM team contains hard-won knowledge.
All our searches are anchored by the ICP Text. This maps the search directly to the ideal customer profile text, instead of taxonomy labels.
Hundreds of teams independently converged on descriptions that surface real matches across almost 70 million business domains.
From Corpus to Assistant
That corpus is extraordinary training material:
- ICP descriptions plus sample domains that anchored the search
- Best phrase matches and proper industry groups
- All of it validated by the simple act of a satisfied customer hitting export
The knowledge was already there, distributed across thousands of successful searches our customers ran. We made it accessible to anyone using our product.
Or keep clicking tired database filters.
Related posts:
