Review: discovery call next step + technique

The job

Jev judge eval (1 choice + 1 score) over a fixed state : runnable via experimental_evaluate.

The prompt

Copy this into Muse and replace anything in [square brackets] with your own details.

// Review: discovery call next step + technique — Jev judge eval (typesafe-ai/jev).
// A discovery call ends with vague promises; the judge classifies the close and grades technique.
// Model: typesafe-ai/jev via the Vercel AI Gateway ($0.042/1M input tokens).
// Auth: set AI_GATEWAY_API_KEY (vck_…) in the environment — never paste the key here.
// For zero-data-retention / no-training workloads, add
// providerOptions: { gateway: { … } } to the evaluate call below.

import { experimental_evaluate } from 'ai';

const result = await experimental_evaluate({
  model: 'typesafe-ai/jev',
  state: "Rep: So, sounds like this could help. Let us circle back sometime? Maybe I can send some info.\nProspect: Sure, send something over.\nRep: Great, will do!",
  questions: {
    "close_type": {
      "type": "choice",
      "instructions": "Classify how this call closed.",
      "criteria": {
        "committed": "A dated next step with an owner was agreed.",
        "soft": "Interest expressed but no date or owner set.",
        "dead": "No interest or an explicit no."
      }
    },
    "technique": {
      "type": "score",
      "instructions": "Score the rep discovery and closing technique. Higher is better.",
      "criteria": [
        "No discovery, no close — just vibes.",
        "Some discovery but the close is wishful.",
        "Solid discovery, weak close without a date.",
        "Pain quantified, champion found, dated next step booked."
      ]
    }
  },
});

console.log(JSON.stringify(result, null, 2));