prompts.chatprompts.chat
Login
All classifiers
NoulIT & incidents

Customer impact signal

Check whether an alert and any incoming reports show customers are affected right now.

F@f

Problem

Internal alerts don't say whether customers notice. Support tickets and status checks that would say so arrive in other tools.

Goal

Know early whether an incident is customer-facing, so communication starts on time and severity is proposed with that fact in view.

Input / state

The alert plus recent support tickets and synthetic check results for the same service.

alert
The alert that fired.
recent_tickets
Subjects of support tickets for this service in the last 30 minutes.
synthetic_checks
Results of customer-path checks.

Decision question

NoulQuestion key customer_impact

Do these signals indicate that customers are affected right now?

What counts as yes and no

  • trueCustomers report problems, customer-path checks fail, or the alert is on a customer-facing path.
  • falseThe problem is internal, for example a batch job or staging, with no customer reports.

Allowed outcomes

Decided in your code from the answer and its confidence. Thresholds are yours to set and test.

  • Customer impact likelyProbability of yes is above your upper threshold.
  • UnclearProbability falls between your two thresholds.
  • No customer impact seenProbability of yes is below your lower threshold.

Example scenarios

Predefined demonstrations. No model is called; the outcome shown is an illustrative judgment.

Example state

{
  "alert": "checkout-5xx-rate 12%",
  "recent_tickets": [
    "Can't pay",
    "Checkout error 500",
    "Payment page broken"
  ],
  "synthetic_checks": "checkout path failing"
}

Outcome

Illustrative
Customer impact likely

Probability of yes is above your upper threshold.

Why

Customer reports and failing customer-path checks.

Using it well

  • Absence of tickets isn't absence of impact; customers often notice before they write. Keep the lower threshold low.
  • Use this to start status page drafts early, not to decide whether to fix the problem.

Request definition

POST https://api.typesafe.ai/v1/systemone
{
  "model": "jev-latest",
  "state": {
    "alert": "<alert>",
    "recent_tickets": "<recent_tickets>",
    "synthetic_checks": "<synthetic_checks>"
  },
  "questions": {
    "customer_impact": {
      "type": "noul",
      "instructions": "Do these signals indicate that customers are affected right now?",
      "criteria": {
        "true": "Customers report problems, customer-path checks fail, or the alert is on a customer-facing path.",
        "false": "The problem is internal, for example a batch job or staging, with no customer reports."
      }
    }
  }
}

Request shape from the System One API reference; a Kev server accepts the same body. Replace the placeholder state with your data. Requests need your own API key or server.

Version History

v1Current Versionabout 8 hours ago@f

Initial version

Comments (0)