Task and capability evaluation
Determine which tasks an agent can complete within defined operating limits
A general capability claim is being used to justify a new business task. The decision is how to determine which tasks an agent can complete within defined operating limits. The work begins with your actual workflow and looks for the evidence that separates intended behaviour from a consequential exception. It is particularly relevant to product owners and domain reviewers. A useful outcome is an explicit account of what was examined, what remains uncertain, and which decision the evidence can support.
The release evidence relationship
- Task definitions
Establish the basis for task family.
- Accepted outcome
Define completion separately from fluent output.
- Task coverage map
Retain the scoped finding and its unresolved dependencies.
Decision
Frame the decision around task family
Questions and the evidence they need
- Task family
- Inspect task definitions for the relevant constraint or observation.
- Accepted outcome
- Inspect accepted examples for the relevant constraint or observation.
- Escalation case
- Inspect difficult exceptions for the relevant constraint or observation.
The decision is how to determine which tasks an agent can complete within defined operating limits. This is relevant to product owners and domain reviewers when a general capability claim is being used to justify a new business task. Name the affected workflow, the consequence of a wrong decision, and the owner who can act on the result. The initial discussion should establish which evidence is already available and which uncertainty is material enough to investigate.
The specific failure hypothesis is that success on familiar examples is generalized to tasks requiring different evidence or authority. It remains a hypothesis until examined in the deployment. Existing internal tests or operating records may already resolve it; identify the uncovered question before commissioning additional work.
Mechanism
Task competence depends on the required business end state
Relationships within release evidence
- Task family
- Group work by the ability and evidence it requires.
- Accepted outcome
- Define completion separately from fluent output.
- Escalation case
- Allow unresolved evidence to require a reviewer.
What this specialty examines
A task family should describe a recognizable job with an observable end state. Extracting an order number, resolving conflicting records and authorising a refund are distinct abilities even when one interface exposes them together. Report the conditions under which each was exercised.
The review should separate task families, define completion criteria and compare outcomes within each family. Identify the normal route, relevant exceptions and the authoritative record of the outcome. Inspect the relevant business-system record, enforcement event or process evidence alongside any reported status. Keep incomplete cases visible, with an explanation of what their absence prevents the owner from concluding.
The tradeoff to make explicit
For this decision, task specificity creates useful decisions while making headline capability scores less convenient. Compare the available design or operating choices against the same intended outcome. If one option removes a required control or changes the task boundary, describe that difference before treating its apparent simplicity, speed or lower effort as an improvement. The decision should expose the tradeoff instead of burying it in an aggregate score.
Preparation
Connect the working inputs to the boundary being examined
Evidence packet for task family
Illustrative working format
- Task definitions
- Identify provenance and the specific decision this material can inform.
- Accepted examples
- Keep expected behaviour and the limits of the constructed example explicit.
- Difficult exceptions
- Name the custodian and the conditions requiring a refreshed version.
- Domain scoring rubrics
- Separate required outcomes, consequential exceptions, and reviewer interpretation.
Start with task definitions, accepted examples, difficult exceptions and domain scoring rubrics. Identify the owner, revision and operating context of each source. Sanitised examples may support initial design, but differences from actual conditions should remain explicit. Where direct access is unsuitable, a customer-controlled evidence or execution route may be considered, with agreed handling and retention.
The scope also needs the release owner, agreed expectations for a representative task case, and the authority to inspect the relevant boundary. Specify what evidence would distinguish the failure in which success on familiar examples is generalized to tasks requiring different evidence or authority from acceptable behaviour. If stakeholders disagree about intended behaviour, resolve the interpretation before classifying the observed difference as a defect.
How the work proceeds
Carry task family through to an accountable finding
From the scoped material to a decision record
- Task definitions
Separate hard constraints from graded preferences.
- Accepted outcome
Apply the scoped check to escalation case.
- Task coverage map
Record the evidence, open dependencies, and the owner’s next decision.
Prepare the evaluation cases
Work with product owners and domain reviewers to assemble task definitions, accepted examples, difficult exceptions and domain scoring rubrics. Select a normal task, a consequential exception and a case with insufficient information. Identify the relevant requirement and expected result for each, then mark which outcomes need human adjudication. The preparation output is a case packet that separates blocking behaviour from graded quality preferences before any comparison is run.
Run and review the comparisons
Use domain reviewers to define accepted answers where business interpretation matters. Include cases requiring abstention or escalation, not just tasks with a direct answer. Track whether a task was wrong, incomplete, unsupported or outside authority; those outcomes demand different product decisions.
Choose the release or evaluation change
Use the reviewed cases to produce a task coverage map and capability profile with unsupported generalizations called out. Distinguish a product defect from a missing test, an ambiguous requirement or an unreliable measuring method. For each, name the proposed correction and the case that would check it. The release owner can then decide whether to change the implementation, improve coverage or restrict the intended use while a specific gap remains.
Indicative artefact
Make the task coverage map usable by its owner
Task coverage map
Illustrative working format
- Task coverage map
- Retain selection rationale, accepted alternatives, and unexamined task conditions.
- Unsupported generalizations called out
- Link supporting and contrary observations to inspectable evidence.
- Accepted outcome evidence
- Link observations and counterevidence to the case, configuration, and conclusion limit.
A useful decision artefact is a task coverage map and capability profile with unsupported generalizations called out. Supporting material may include selected cases, configuration or process records, evidence references, a disagreement log and prioritised actions. Retain the case selection rationale and material disagreements alongside the result; otherwise the result may be mistaken for complete coverage.
Illustrative scenario
Examine escalation case through a constructed exception
A constructed case and its decision paths
Illustrative scenario
- Task family
A ticket summarizer encounters conflicting refund records.
- Accepted outcome
Vary whether the conflicting records can be reconciled.
- Escalation case
Permit escalation when the available facts cannot support a refund decision.
Consider a fictional situation: an agent summarizes tickets accurately but is then asked to reconcile conflicting refund requests. The scenario is illustrative, with no client engagement, executed test or measured improvement implied.
For conflicting refund requests, vary whether the records can be reconciled from available evidence. A useful agent may complete the clear case and escalate the unresolved one. Treating both as mandatory automatic completion would encode the wrong task expectation.
Scope and next action
Keep the conclusion within selected task cases
Confirm fit before the next commitment
- Task family
- This review is limited to selected task cases.
- Task definitions
- The relevant owner must confirm access and the intended operating conditions.
- Task coverage map
- The customer retains the release decision.
A result applies to the recorded configuration and selected cases. It does not establish universal agent competence or authorise release. Any engagement depends on an agreed scope, qualified team, authorised access and confirmed availability. Bring the affected workflow, its current configuration or process, the triggering change and the evidence already available. The immediate scope question is prompted by the situation in which a general capability claim is being used to justify a new business task; confirm which missing observation would change the owner’s next action.
Ways to engage
Related engagement scopes
AI Agent Evaluation Evidence Diagnostic
Find the gaps between your release decision and the evidence behind it. Start with one agent workflow and a bounded set of questions.
AI Agent Evaluation Foundation
Turn scattered tests into a repeatable release discipline. Connect the behaviours that matter to datasets, evaluation criteria and evidence your team can maintain.
AI Long-Horizon Agent Evaluation
Examine what changes when an agent keeps working. Evaluate accumulated state, recovery and goal adherence across extended tasks.
Continue exploring
Related insights and resources
Engineer the harness, not just the prompt
Define the contracts around an agent's decisions: what state it trusts, which actions can execute, how interrupted work resumes, and what evidence an operator needs to recover safely.
What would justify releasing this agent?
Replace a single pass rate with a reviewable chain of requirements, cases, observations and judgements. Distinguish task quality, control failures and uncertainty before increasing an agent's exposure.
The evidence behind the go-live decision
A working paper for leaders who need to explain why an agent workflow is ready for its next release. Includes a proposed evidence-pack structure, decision table and review agenda.