Clay + HubSpot: the enrichment workflow we use for every new lead

Clay is one of the most powerful enrichment tools available, waterfalling a single lookup through dozens of data providers until it finds an answer, but getting that enrichment to sync cleanly into HubSpot without creating duplicate fields, broken mappings, or half-filled records takes real setup. Here's the workflow we run for every new lead: what data we pull, how we handle failures, and what we deliberately skip.

Why Clay and HubSpot don't sync cleanly out of the box

Clay is built to be flexible, run any lookup, chain any provider, write any output field, and that flexibility is exactly what makes a naive sync into HubSpot messy. Without a deliberate mapping layer, enriched fields land in the wrong property, overwrite data a rep already entered manually, or create near-duplicate fields that say almost the same thing. The fix isn't avoiding Clay's flexibility, it's building the sync logic (through n8n, wired into both platforms) so enrichment behaves predictably every time a new lead comes through.

Setting up a bidirectional Clay to HubSpot sync via n8n

The workflow runs in both directions. A new or updated lead in HubSpot triggers a lookup in Clay, and once Clay's waterfall resolves the enrichment, the result is validated and written back to the matching HubSpot property through n8n, never as a raw dump into a catch-all field. Keeping the mapping explicit, one Clay output field to one specific HubSpot property, is what prevents drift over time as either platform's schema changes.

• >New or changed lead in HubSpot triggers the Clay enrichment lookup
• >Clay waterfalls the lookup across multiple data providers automatically
• >n8n validates the result before it ever touches a HubSpot property
• >Each enrichment field maps to one specific property, never a catch-all

How do we handle enrichment failures and partial data?

Enrichment fails constantly, a provider has no record for a small company, a job title comes back ambiguous, an email fails verification. The workflow is built to write what it actually finds and leave the rest blank rather than guessing or writing a low-confidence value that looks authoritative. A partially enriched record with honest gaps is more useful to a rep than a fully filled record where half the fields are wrong.

Which fields are worth enriching vs looking up on demand

Not every field belongs in the automatic sync. High-frequency fields that every rep needs immediately, company size, industry, seniority, get enriched automatically on lead creation. Deeper, more expensive lookups (detailed tech stack, recent funding, hiring signals) are better run on demand for leads that actually clear a scoring threshold, not on every lead that enters the funnel. Enriching everything automatically wastes credits on leads that were never going to be worked anyway.

Why the waterfall matters more than any single provider

No single enrichment provider covers every company well. A provider that's strong on large enterprise accounts is often thin on small businesses, and the reverse is just as common. Clay's real value is running a lookup through several providers in sequence, moving to the next one automatically when the first comes back empty, so the workflow returns a usable answer far more often than any single data source could on its own. Building that waterfall order thoughtfully, cheapest and most reliable providers first, is what keeps enrichment cost reasonable without sacrificing coverage.

Enrichment fields we deliberately skip automating

We don't auto-write anything a rep is likely to have entered manually and trusted more than an automated source, like a personal note field or a manually confirmed decision-maker status. Overwriting human judgment with automated enrichment is one of the fastest ways to break a sales team's trust in the CRM. When a field is genuinely ambiguous between what Clay found and what a rep entered, the workflow leaves the rep's value alone and stores the enrichment result in a separate field for reference instead of forcing a silent overwrite nobody asked for.

mkdir builds this enrichment and scoring logic inside the Clay and HubSpot instances you already run, so every new lead arrives enriched and ready to work, not just enriched and messy.