Use cases
For a personal project, a new app, or your team’s recurring work. Start with a clear question and examples whose answers you can check.
See the training workflowExplore a model that organizes research using the distinctions you care about.
“Is this note about methods, findings, or limitations?”
Train a prediction model on editorial judgments, using examples you have scored. It evaluates the headline; it does not write it.
“How well does this headline communicate the article’s main finding?”
Add a specialized classifier to your product through the API, using categories that match your application.
“Is this submission a question, a bug report, or a feature request?”
Learn from historical opportunities and their outcomes, then assess new deals over a defined period.
“Will this deal close in the next seven days?”
Connect the context of a support interaction to the rating the customer later gave.
“Will this customer leave a five-star review?”
Use past incidents and their confirmed severity to help assess a new report.
“Which severity category fits this incident?”
Use only information that would have been available before the outcome. Set aside independent test data and check the model on examples it hasn’t learned from.
These are illustrative prediction tasks, not claims of measured performance. Available connectors are Attio, Pylon, and Fireflies. Talk to us about sourcing data for your use case.