Home / Examples

THE DECISION LAB / 08 PRACTICAL PATTERNS

Small decisions.
Real possibilities.

Explore what structured AI can do between the steps of your workflow. Pick a scenario, change the context, and see what Jev decides.

INPUT → DECISION → NEXT STEPLess theory.
More “try it.”

8 walkthroughs · 16 scenarios
Choice, Score, and Yes / No

Find your first example ↓

NEW TO JEV?

Start with a familiar decision.

Route a support ticket, inspect the probabilities, then turn the result into a shareable card.

Try support routing ↗

8 examples to explore

Choice

Customer operations

Support ticket routing

Build an AI support ticket classifier with Jev. Route billing, technical, and sales requests using explicit choices and a review fallback.

01 / TWO SCENARIOSExplore ↗

Choice

Agents & workflows

Agent tool selection

Try AI agent tool routing with Jev: choose weather lookup, web search, a calculator, or no tool before an agent takes its next step.

02 / TWO SCENARIOSExplore ↗

Score

Search & quality

RAG relevance scoring

Score retrieved passages before generation. Try a Jev RAG relevance rubric, compare related versus useful documents, and evaluate retrieval quality.

03 / TWO SCENARIOSExplore ↗

Yes / No

Search & quality

Response instruction check

Use Jev Yes / No to check whether an AI response follows a stated instruction. Try passing and failing examples before designing a validation gate.

04 / TWO SCENARIOSExplore ↗

Choice

Customer operations

Email intent classification

Classify incoming email by intent using Jev Choice. Separate demo requests, pricing questions, account help, and messages needing manual review.

05 / TWO SCENARIOSExplore ↗

Choice

Customer operations

Customer feedback themes

Turn product feedback into useful themes with Jev AI classification. Distinguish bugs, feature requests, usability friction, and positive feedback.

06 / TWO SCENARIOSExplore ↗

Choice

Customer operations

Sales lead routing

Route inbound sales requests with Jev AI using stated product requirements. Try enterprise, team onboarding, self-service, and manual review paths.

07 / TWO SCENARIOSExplore ↗

Yes / No

Agents & workflows

Agent completion check

Check whether an AI agent has met a task’s completion criteria. Use Jev Yes / No with observable evidence, retry limits, and an explicit stop policy.

08 / TWO SCENARIOSExplore ↗

CHOOSE YOUR DECISION SHAPE

Choice

Pick a label from a defined set. Start here for routing tickets, classifying feedback, or selecting the next tool.

Explore classification ↗

Score

Judge a passage against ordered criteria. Use a clearly defined rubric when a useful answer lies on a spectrum.

Explore relevance scoring ↗

Yes / No

Check one condition. Keep the question narrow and provide observable evidence for instruction or completion checks.

Explore completion checks ↗

Original educational examples informed by public technical references, linked on each walkthrough. Expected outcomes are illustrative, not recorded benchmarks or customer success claims. Read the API guide ↗