AI evaluation
Ask typed questions about a record and get probabilities back: a choice among options, a score on a scale, or whether something is true. Nothing to configure or authorise.
What the connection reaches
- Choice, score and probability questions in one call
- Runs in managed pipeline runs
- No model to pick and no key to manage
- Status
- Built into the pipeline SDK
- Connection class
- EvaluationConnection
- Credential
- Nothing: SegmentStream chooses the model and holds the credential
- In use
- In production at an AI video company, where it grades every demo request A to D within hours.
Connect AI evaluation
Three steps, the same for every connection. The credential goes to the platform, never into the repository or pipeline code.
- 1
Declare it in connections/connections.py
connections/connections.py from segmentstream.connections import EvaluationConnection connections = ( EvaluationConnection(key="evaluation"),) - 2
Deploy
$ segmentstream deploy - 3
Connect
Nothing to configure. The connection works as soon as the workspace deploys.
Use it
Code asks for the connection by its key. The platform adds the credential on the way out.
import segmentstream as s client = s.evaluation_client("evaluation")result = client.evaluate(state=lead, questions=questions)grade = result["answers"]["grade"]Or ask your agent
“Grade each new lead A to D by how likely sales is to accept it.”
Part of these solutions
Want AI evaluation set up for you?
Our forward-deployed engineers build one solution on your data, in your own workspace, as code you own. Tell us what you run today and what you want; your note goes straight to our founder.
- One solution, on your data, in your own workspace
- First version live in two weeks
- Parity proven before anything is switched off
- Fixed scope and price, credited to your contract if you continue
Not ready for a pilot? Send the form anyway with your question. An engineer will reply, not a sales rep.
How the pilot worksConnect what your company runs.
Our engineers build the first version on your data in two weeks, at a fixed price.
Or join the waitlist for self-serve:
Self-serve opens to the waitlist first.