Test one whole journey

A score finds surface gaps. A journey shows where the actual job breaks.

Is Agenticis-agentic.com
The score is an at-a-glance summary; the individual checks and evidence are the useful part of the report.
MethodologyRead docs ↗

Notes from the Is Agentic checks and report format.

Run one concrete goal

Toggle the client-only answer, then run the journey.

Score aloneBad
92/ 100 readiness

Looks healthy. The task has not run.

checks passed: 46 / 50
Observed journeyGood
  1. Open the pricing page
  2. Find the team plan limit
  3. Open API documentation
  4. Find the matching endpoint
  5. Explain it with sources

Ready to run.

goal + steps + evidence + outcome
test('agent can answer a pricing question', async () => {
  const run = await observeAgent({
    goal: 'Find the Team plan limit and the API that reports usage.',
    startUrl: baseUrl,
  });

  expect(run.outcome).toBe('completed');
  expect(run.answer.sources).toHaveLength(2);
  expect(run.steps).toContainEqual(
    expect.objectContaining({ status: 'evidence' })
  );
});

Why it matters, pros and cons

Why it matters

A score is a map of likely friction. I still want to watch one ordinary agent attempt one ordinary job before I call the surface ready.

Pros

  • Failures point to an actual user goal.
  • The trace connects cause to fix.
  • Rerunning proves the fix changed behavior.

Cons

  • One journey cannot represent every agent or goal.
  • Runs vary, so keep evidence and timestamps.
  • Private and geographic states need separate coverage.