Jev support resolution preference

Identify the resolution a customer explicitly requests

Read the customer message to distinguish a refund, replacement or troubleshooting request. Keep a described problem without a chosen resolution separate from an explicit preference.

Which remedy does the customer actually request?Live API
Try an example

Original synthetic input. Edit the task or evidence to make a new API request.

Source: Workflow inspiration on X

Source and example scope

The inputs, rubric and preview are original. This window demonstrates the stated decision only; it does not reproduce the source application or its measurements.

Read customerMessage.

Decision rules and labels
Refund requestedReplacement requestedTroubleshooting requestedNo chosen resolutionNeed preference
A real response from the Jev API

Edit any input to explore a different decision. No request is sent until you run it.

Your next decision

Not run yet

Choose an example or write your own input, then run it to see the ranked choices.

Results are not stored by this site. Your submitted text is processed by the API provider.

The decision

How does Jev support resolution preference use the supplied evidence?

Jev support resolution preference identifies the remedy a customer actually asks for. Jev separates an explicit refund, replacement or troubleshooting request from a reported symptom and an unresolved preference. The local output proposes what an agent should confirm, without granting a remedy.

A support-classification post warned that valid structured output can still contain the wrong judgment. This original demonstration keeps the customer wording, authored expectation and full returned distribution together so the remedy interpretation can be reviewed.

Support department routing and desired resolution answer different questions. The same broken lamp can produce a request for money back, another lamp or help fixing it. A complaint alone does not authorize the model to select a remedy the customer never chose.

All messages are fabricated and no order or policy record is attached. A real support application must establish identity, eligibility and permitted actions separately. This page reads an expressed preference only, and even a clear preference remains a pending review rather than an account change.

Original Jev support resolution preference workflow diagram showing supplied evidence, a typed decision and the local preview.
Original workflow illustration. The output is a local preview, not an executed action or source screenshot.
How it works

Try the example in four steps

Use the sample to understand the rule, then test your own input for customer resolution requests.

  1. Illustration of adding the text or evidence for a decision

    Add the information to check

    Read the customer message to distinguish a refund, replacement or troubleshooting request.

  2. Illustration of defining a question and its allowed choices

    Confirm the question and labels

    Open the decision rules and keep the intended distinction between Refund requested, Replacement requested, Troubleshooting requested, No chosen resolution, Need preference. Edit the choices only when testing your own rule.

  3. Illustration of sending one input and comparing the returned choices

    Run the decision

    Use a sample first, then try your own input. Compare the selected label with the other returned probabilities.

  4. Illustration of reviewing a result and an uncertain item before using it

    Review the next step

    Review Need preference before using a result. The label is a decision about the supplied information, not an executed action.

Our API test

What the six original inputs returned.

What the six original inputs returned.
Original test inputFixture expectationObserved labelLeading probability
Money back, explicitly requestedRefund requestedRefund requested100%
Another item, explicitly requestedReplacement requestedReplacement requested100%
Help with the current itemTroubleshooting requestedTroubleshooting requested100%
A symptom without a preferred remedyNo chosen resolutionNo chosen resolution100%
Alternatives without a choiceNeed preferenceNeed preference100%
A paraphrased refund requestRefund requestedRefund requested100%
  • 6 of 6 original fixture expectations matched. Contrary outputs are preserved in the table and downloads; these checks are not held-out accuracy or production suitability.
  • Measured adapter round trips: 454-1187 ms. Estimated input cost: $0.00002423-$0.00002453 per request. These observations are not a latency guarantee or invoice.

6 actual Jev API calls on October 4, 2026. Original synthetic inputs check this rubric and handoff; they do not establish held-out accuracy. Probabilities are model outputs.Verified

Understand the results

Money back, explicitly requested

The first customer explicitly rejects another lamp and asks for the payment back.

Read the example notes

Refund requested proposes confirming that expressed preference. It does not verify a transaction, establish refund eligibility or calculate an amount. The distinction matters because recognizing intent and authorizing a financial action require different evidence. The inspectable packet asks whether an agent may review the order and applicable refund policy. That wording is a fixed authored string, not generated advice or an executed refund.

Another item, explicitly requested

The second message asks for another item and explicitly excludes money back.

Read the example notes

Replacement requested preserves the chosen remedy. A real replacement workflow still needs stock, order identity and policy checks. The demonstration cannot infer any of those facts from a clearly worded message.

The flickering lamp message asks for diagnosis and repair while keeping the item

The flickering lamp message asks for diagnosis and repair while keeping the item.

Read the example notes

Troubleshooting requested proposes continuing with that preference. The model does not diagnose an electrical fault or provide repair instructions. It returns only the supported request type from the supplied text.

The symptom-only message describes flickering but chooses no remedy

The symptom-only message describes flickering but chooses no remedy.

Read the example notes

No chosen resolution proposes asking which outcome the customer wants. Reusing the refund result from a neighboring example would add information absent from this request. The exact symptom-only input and its recorded response remain downloadable.

Alternatives without a choice

The customer names two alternatives and explicitly says the choice is unresolved.

Read the example notes

Need preference preserves that uncertainty instead of taking the first mentioned option. It does not mean the complaint is invalid. It means a remedy preference has not been established by the supplied message.

The last customer asks for the money paid to be returned without using the word refund

The last customer asks for the money paid to be returned without using the word refund.

Read the example notes

The model should interpret the requested outcome from meaning. Refund requested still leaves transaction verification and policy outside the local packet, and this paraphrase check is not a claim of accuracy on unseen customer messages.

Source and example scope

Fixture expectations are authored before the calls. Any contrary model choice remains visible beside its expectation and complete distribution; no repeated judgment is used to obtain a preferred label.

Using the result in your application

Download the exact sample requests and saved observations below.

Read the example notes

Expected labels are original fixture checks and stay outside the request sent to Jev. The observation file preserves the complete probabilities and the hash of each exact request. Editing a task, rubric or candidate requires a fresh response; a matching sample name does not make an older answer current.

Using the result in your application

The JavaScript runner makes one independent SDK request for a chosen sample ID.

Read the example notes

It reads TYPESAFE_API_KEY privately from the environment and has no automatic retry. The Python consumer operates offline: it checks the rubric, labels, unique sample identities, request hashes and valid probability distributions before selecting a fixed local preview. A tied leading result remains pending for review.

Source and example scope

The page reports measured adapter round trips and estimated input cost for the recorded calls.

Read the example notes

Those values describe these requests, not a provider latency guarantee or a billing statement. Six deliberately authored fixtures can check the stated behavior, but they do not establish accuracy on an unseen dataset. Keep disagreements and missing-input examples when testing an integration.

Runnable offline example

Replay one request and inspect the saved preview.

View the code example
Replay commands
npm install @typesafe-ai/sdk@0.6.0
# Set TYPESAFE_API_KEY privately in your shell.
node run-choice-sample.mjs samples.json 1
python preview_choice_workflow_v2.py samples.json saved-api-results.json
Runnable offline example

The first request and its recorded local output.

View the code example
Exact first sample and local preview
{
  "input": {
    "task": "Review the supplied evidence under the explicit rule, preserving any missing or unresolved facts.",
    "customerMessage": "The lamp arrived cracked. I want my payment returned rather than another lamp."
  },
  "localOutput": {
    "sampleId": "1",
    "leaders": [
      "Refund requested"
    ],
    "probabilities": {
      "Refund requested": 1,
      "Replacement requested": 0,
      "Troubleshooting requested": 0,
      "No chosen resolution": 0,
      "Need preference": 0
    },
    "action": "confirm_refund_preference",
    "payload": {
      "expressedPreference": "Refund requested",
      "proposedConfirmationQuestion": "You requested money back. May the agent review the order and applicable refund policy?",
      "messageSent": false,
      "ticketChanged": false
    },
    "pending": true
  }
}
Scope of this demo

Where does this decision fit?

Useful with explicit supplied evidence

  • Identify the remedy explicitly preferred in a supplied customer message, preserving symptom-only and unresolved alternatives before any separate support policy or execution.
  • Inspecting the original task and a fixed output before integrating external tools.
  • Comparing an explicit expected label with the complete recorded distribution.

Needs separate implementation or evidence

  • Inferring facts absent from the supplied input.
  • Treating a semantic selection as authorization to act.
  • Executing an external action or generalizing six fixtures into an accuracy claim.
Example FAQ

Questions about this original decision workflow.

Keep relevance, compliance and execution separate.

Does recognizing a refund request authorize payment?

No. Identity, transaction, amount and policy require separate authoritative checks. The packet executes nothing.

Is this support department routing?

No. It identifies the customer’s preferred remedy, not which team owns the issue.

What if a customer has not selected a remedy?

Keep No chosen resolution or Need preference visible and request clarification. Do not invent the customer’s preference.

Are the messages real customer data?

No. The messages are original fabricated inputs. The stored Jev responses, SDK replay and edited browser call are real API observations.

Try a task from your own workflow

Change the supplied task or evidence here, then inspect the live response. Playground supports another explicit decision.