Yusuf Ahmed
← Builds
Build · Video walkthrough

Personality Map

A small build that shows the pattern behind most of the larger ones: Clay as the enrichment and routing layer, a custom app as the logic layer and front end, model judgment in between, and the result written back to drive the copy.

Since2025StackClayReplitHTTP APIOpenAI modelsLinksBuild video ↗
The problem

What was actually hard

Clay tables are good at fetching and routing data and poor at holding application logic. The question was where the logic should live when a workflow outgrows a column formula.

How it works

In order

  1. Collect. Clay pulls each prospect's recent LinkedIn posts and tweets.
  2. Score. Clay posts them over HTTP to a Replit app that scores the Big Five traits.
  3. Dashboard. the app renders everyone on a filterable view for the seller.
  4. Write back. scores return to Clay and set each email's angle and tone.
Judgment calls

Decisions that made it work

  • Keep Clay as the bus, not the brain.Logic in a real app is testable, versionable and reusable; logic in a table column is none of those.
What it did

Outcomes

Published as a build video; the pattern reappears in every larger client system on this site.

Why it matters

The small version of a pattern I use everywhere: the table routes, the app decides, the model judges.