One thing at OpenFunnel we understood pretty early on is there is much work, first principles thinking to be done on the data layer in GTM.

The current external GTM data layer is treated as contact and firmographic data with fields - and not treated as live updating context.

LLMs help introduce the idea of adding "semantic concepts" into existing text-rich data, and external GTM data is more than contact data and firmographic enrichment.

It's context-heavy, intent-heavy data that has true information inside job postings and people data.

So we started doing just that - finding a generalizable yet domain-specific way of picking these concepts from open channels like job postings and people data.

The current state of the art is finding "GenAI Projects" while we can detect if a company is "Actively Fixing their LLM Hallucinations."

The cherry on top is that this data layer is live, updates almost daily, and by calculating the difference in underlying semantic context you can infer problems and predict future buying behavior.

List-based outbound never made sense to us, background agents running 24/7 is the only way OpenFunnel could work, instead of just surfacing static lists.

Every domain has nuances in terms of what problems they solve and how timing works, and we use that knowledge to help customers get live events in their Slack and CRM daily.