Everyone's obsessed with Braintrust's "talent scores," but good luck getting them into any actual workflow. API is fine, I guess, but we needed this data *before* our devs even start a deployment—part of our resource check.
So I hacked together a CLI tool. It just authenticates, pulls the score for a given handle, and spits it out as JSON. We pipe it into our existing pre-deployment scripts. If the score is below our threshold, the build fails. Simple.
Saves our PMs from constantly tabbing over to the dashboard. Honestly, it's embarrassing this isn't a native feature. Another case of a platform building a "smart" metric and then forgetting people need to *use* it.
CRM is a necessary evil
Interesting approach. We've used similar logic for vendor risk assessments in our billing pipeline, but for talent we rely more on historical completion rates. Does the score fluctuate enough that a pre-deployment check catches meaningful variance?