Orchestration layer

Clay's own framing of itself makes the distinction clean: it isn't primarily a data vendor, it's an orchestration layer, a workflow builder that calls other companies' data and runs logic on top of what comes back, enriching, verifying, scoring, routing to a CRM, triggering a sequence. The orchestration layer doesn't necessarily hold the data itself. It decides which source to call first, what to do when that source comes back empty, and what happens next once a record is enriched.

That's the same role a waterfall plays inside an ABM motion: Clay is described as "the orchestration layer where this lives," cascading through a contact-level waterfall like FullEnrich underneath it, so a miss on the first provider falls through to the second instead of leaving the field empty. The orchestration layer is what turns a set of separate tools into one system with a decision at each step, instead of a person manually deciding which tool to check next.

The orchestration layer can sit at more than one level

Orchestration doesn't stop at data enrichment. In a fully code-run stack, an agent like Claude Code can be the orchestration layer over the whole system: research that used to live in a browser tab runs as agent sessions, enrichment runs as per-task waterfalls in Clay underneath it, and sends go out through a sequencer over its API. Each layer coordinates the layer below it, which is why "orchestration layer" describes a role in a stack, not one specific product.