Find out what's going wrong with your Clay workflows — and what to do next.
Pick the closest symptom category. Guided diagnostics narrow the likely cause; every result includes verification steps, fixes, and cautions before changes touch production data.
Enrichment failures
Providers erroring, timing out, or returning nothing — diagnosis by scope, error type, and run context.
Unexpected credit or action consumption
Spikes and structural overspend: what changed, what the usage data says, which levers cut cost.
Poor enrichment coverage
Low match rates and empty columns: input preparation, provider choice, and waterfall order.
Duplicate account/contact records
Dedupe keys, upsert design, and cleaning up duplicates without losing data.
CRM synchronization errors
Rejected, missing, misfielded, or duplicated CRM writes and how to trace each.
API failures and rate limits
Provider quotas, key scoping, retries with backoff, and concurrency control.
Broken workflow dependencies
When an upstream table, provider, or integration changes and downstream steps silently fail.
Conditional logic problems
Formulas and conditions that misfire: null handling, type mismatches, and hidden assumptions.
AI research quality issues
Claygent and Use AI output that is inconsistent or wrong — prompt structure and verification.
Outbound activation failures
Sequencer and export failures: action accounting, suppressions, and deliverability checks.
Data validation problems
Format, completeness and plausibility checks that stop bad data reaching the CRM.
Slow or unreliable workflows
Concurrency, batching, table hygiene, and retry behavior that keep runs predictable.
Permissions and access issues
Role gaps, token expiry, and integration-user permission errors.
Workflow governance problems
Undocumented workflows, missing owners, and uncontrolled changes — and the minimum standards that fix them.