Skip to content
All products

Guided UAT

Available · Workday

User acceptance testing that doesn't collapse

Consultants curate a versioned test library; client testers execute it one step at a time, guided by an agent that knows the tenant design, the personas and the known failure modes.

The problem

UAT is where ERP programmes quietly lose their schedule. Testers are business users who test twice a year, scripts drift from the tenant, and defects arrive as “it didn't work” three weeks after the fact.

Who it’s for

System integrators and internal programme teams running Workday UAT cycles with business-user testers.

How it works

  1. 01

    The consultant uploads, maps, validates and publishes a test library. Once published, the version is immutable.

  2. 02

    A tester opens their assignment and sees one step at a time, in dependency order — and nothing from another engagement.

  3. 03

    The agent guides and diagnoses against the tenant design, personas and ranked failure modes.

  4. 04

    The human tester judges pass or fail. A named consultant confirms. No model response ever sets a result.

What it does

Grounded by construction

The agent answers from what it retrieved for this engagement, not from what it remembers. Retrieval is visible to the tester alongside the reasoning.

Immutable library versions

Import, map, validate, publish. A published version cannot be edited, so a defect can always be traced to the exact script that produced it.

Diagnosis against known failure modes

Ships with a curated content base — areas, personas, and failure modes with ranked probable causes — so a failing step arrives with a hypothesis attached.

Model where judgement is needed, code where it isn't

The rules engine handles validation and lifecycle deterministically. The model handles explanation and diagnosis. There is no code path where the model writes a result.

Hard role boundaries

Admin, consultant and client tester see different systems. The cross-engagement boundary is covered by end-to-end tests, not just middleware.

Degrades honestly

Without a model key the runner falls back to a manual checklist and everything else keeps working. Health reports ok, degraded or unhealthy.

Built, not mocked up

  • 80 unit tests, 36 integration tests and 6 end-to-end role-boundary tests
  • A fixed agent evaluation set run as part of the suite
  • Content base of 23 areas, 8 personas, 8 failure modes with 33 ranked causes, and a 47-term glossary
  • A fully worked reference engagement ships with the product

Where it sits in the method

Guided UAT is step E in one product: the tool and the testing process change together, because a guided runner is a different process, not a faster old one.

Read the OUTCOME Method

Next step

See Guided UAT against your own workflow.

We will run it on a real example from your business rather than a canned demo, and be straightforward about where it does not fit.