Building with TypeSafe AI's Jev in Logic Apps: from triage to a router agent

Description and use

Typed Jev responses from TypeSafe AI enable Logic Apps to perform deterministic branching without parsing free text, implementing triage and task routing in real time.

Problem

Integrating a typed AI model into Logic Apps workflows to replace free‑text parsing with deterministic, schema‑based decisions for triage and routing.

Outcome

The typed responses from Jev enabled deterministic branching in Logic Apps, eliminating free‑text parsing. The triage workflow ran in 2.4s, and the router agent correctly dispatched math and general tasks, demonstrating practical, end‑to‑end integration.

Workflow

• 1. Add three Jev questions (urgent, category, sentiment) to a single API call.
• 2. Use Compose_Triage to evaluate urgent.noul, category.choice, and sentiment.score with @if expressions.
• 3. Deploy triage workflow in Logic Apps Automation and observe 2.4s total run time.
• 4. Create RouterAgent that sends a single choice question (math vs general).
• 5. Switch on the answer and dispatch to BodmasAgent or GeneralAgent.
• 6. Resolve call‑workflow limitation in Automation by using Http POST to the target workflow’s trigger URL.
• 7. Test both branches: math expression and general question, confirming correct dispatch and responses.

Evidence

TypeSafe AI's Jev model answers only in a shape you define. You send it state (your text) and a set of typed questions, and it returns typed values with calibrated probabilities instead of free text to parse.

Source

This run took the default (general) branch: ReqTask → Classify_Task (808 ms) → Route_By_Task_Type → Dispatch_To_GeneralAgent (7.9 s) → Response_General (132 ms), succeeded in 10 seconds total.

Source

The math branch works too, tested separately: Math: (12 + 8) * 3 - 5 / 5 → classified math (confidence 1.0) → dispatched to BodmasAgent → "Answer: 59". Correct.

Source