Softlander Book a call
Agent evaluation and assurance

Autonomy-level evaluation

Choose the permitted level of agent action for a particular workflow

An assistant is moving from suggestions to executing business changes. The decision is how to choose the permitted level of agent action for a particular workflow. 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 executives and control owners. A useful outcome is an explicit account of what was examined, what remains uncertain, and which decision the evidence can support.

Book a 30 minute readiness callBrowse this capability group

The release evidence relationship

  1. Starting evidenceAction inventories

    Establish the basis for read and propose.

  2. Decision boundaryApproval boundary

    Name who may authorise the consequential action.

  3. Decision artefactAutonomy decision matrix

    Retain the scoped finding and its unresolved dependencies.

Proposed working model. The customer retains the release decision.

Decision

Frame the decision around read and propose

Questions and the evidence they need

Criterion / evidenceRead and propose
Inspect action inventories for the relevant constraint or observation.
Criterion / evidenceApproval boundary
Inspect reversibility for the relevant constraint or observation.
Criterion / evidenceExecution scope
Inspect approval rules for the relevant constraint or observation.
Read each criterion against the available evidence before deciding that additional work is needed.

The decision is how to choose the permitted level of agent action for a particular workflow. This is relevant to product executives and control owners when an assistant is moving from suggestions to executing business changes. 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 authority is expanded for an entire agent even though only some actions have adequate evidence. 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

Autonomy belongs to an action and its consequence

Relationships within release evidence

Source or conditionRead and propose
Separate information access from business commitment.
Crossing or choiceApproval boundary
Name who may authorise the consequential action.
Consequence or decisionExecution scope
Tie unattended action to evidence and reversibility.
Keep case coverage linked to the release question. Each relationship needs evidence in the actual scope.

What this specialty examines

Autonomy should be allocated per action and condition. Reading a record, proposing an edit and committing it expose different consequences. Record who can authorise each step, what evidence is required and whether the action can be reversed or stopped once initiated.

The review should evaluate read, propose, approve and execute permissions separately by action and consequence. 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, more autonomy can remove handoffs while increasing the consequences of an undetected mistake. 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 read and propose

Illustrative working format

BasisAction inventories
Identify provenance and the specific decision this material can inform.
ContextReversibility
Check its relationship to the relevant workflow and expose missing context.
ObservationApproval rules
Connect the accepted action parameters to grant, amendment, expiry, and withdrawal.
DependencyOperator intervention paths
Identify provenance and the specific decision this material can inform.
Illustrative preparation record. Availability, revision, and handling need confirmation.

Start with action inventories, reversibility, approval rules and operator intervention paths. 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 authority is expanded for an entire agent even though only some actions have adequate evidence 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 read and propose through to an accountable finding

From the scoped material to a decision record

  1. Establish conditionsAction inventories

    Separate hard constraints from graded preferences.

  2. Examine the relationshipApproval boundary

    Apply the scoped check to execution scope.

  3. Retain the findingAutonomy decision matrix

    Record the evidence, open dependencies, and the owner’s next decision.

Proposed sequence; resolve interpretation or access gaps before relying on the next step.

Prepare the evaluation cases

Work with product executives and control owners to assemble action inventories, reversibility, approval rules and operator intervention paths. 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

Examine escalation as part of the autonomy design. An agent with limited permissions still needs a way to return useful partial work and explain what remains unauthorised. Avoid expanding authority merely because blocked tasks reduce a headline completion rate.

Choose the release or evaluation change

Use the reviewed cases to produce an autonomy decision matrix with prerequisites and escalation boundaries. 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 autonomy decision matrix usable by its owner

Autonomy decision matrix

Illustrative working format

Annotated fieldAutonomy decision matrix
Identify the assessed workflow, revision, and decision boundary.
Annotated fieldPrerequisites
Link supporting and contrary observations to inspectable evidence.
Annotated fieldEscalation boundaries
Name the unresolved condition, next action, and accountable owner.
Illustrative artefact structure, not a completed engagement or an observed result.

A useful decision artefact is an autonomy decision matrix with prerequisites and escalation boundaries. 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 execution scope through a constructed exception

A constructed case and its decision paths

Illustrative scenario

  1. SituationRead and propose

    An assistant recommends suppliers while purchase approval remains with a person.

  2. Question to examineApproval boundary

    Map information access and purchasing to separate permissions.

  3. Conditional next actionExecution scope

    Keep purchase commitments behind accountable approval.

Fictional example. No executed test, customer engagement, or measured outcome is implied.

Consider a fictional situation: an internal assistant may summarise supplier options but purchase commitments require a separate accountable approval. The scenario is illustrative, with no client engagement, executed test or measured improvement implied.

For the supplier order, the agent may assemble a recommendation and supporting evidence while a person approves commitment. If approval is unavailable, the expected result is a prepared draft with an explicit pending decision, not an improvised purchase.

Scope and next action

Keep the conclusion within selected task cases

Confirm fit before the next commitment

Included boundaryRead and propose
This review is limited to selected task cases.
PrerequisiteAction inventories
The relevant owner must confirm access and the intended operating conditions.
Claim limitAutonomy decision matrix
The customer retains the release decision.
Bring the triggering change and existing evidence; resolve the dependency that could alter the 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 an assistant is moving from suggestions to executing business changes; 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.

Read about this offer

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.

Read about this offer

AI Long-Horizon Agent Evaluation

Examine what changes when an agent keeps working. Evaluate accumulated state, recovery and goal adherence across extended tasks.

Read about this offer

Continue exploring

Related insights and resources

Whitepaper16 MIN READ

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.

Whitepaper15 MIN READ

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.

Whitepaper5 MIN READ

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.

A little more clarity

Questions to consider.

Start with the questions that shape a better decision.

Can one autonomy level describe every action an agent performs?

Usually not usefully. Read, propose, approve and execute actions can require different evidence and constraints. Allocate autonomy by action and context, including reversibility, consequence and available intervention.

Which action records support an autonomy decision?

Start with action inventories, reversibility, approval rules and operator intervention paths. Record their revision, owner and relationship to the intended workflow. Missing context should be identified before interpreting outcomes; a sanitised fixture is useful only when its relevant differences from the operating case are understood.

How should read, propose and execute permissions be evaluated?

Evaluate read, propose, approve and execute permissions separately by action and consequence. Keep task correctness, permitted action and final state distinct. Review scoring disagreements against the agreed rubric. Include valid alternatives so the cases do not reward superficial sequence matching or rejection of every difficult request.

What belongs in an autonomy matrix?

The proposed output is an autonomy decision matrix with prerequisites and escalation boundaries. Include the selected conditions, supporting observations, unresolved differences and next decision. A reviewer should be able to follow each important conclusion back to its evidence and understand the limits without depending on an undocumented explanation from the author.

When should human approval remain mandatory?

Retain approval when the action's consequence, uncertainty, reversibility or mandate requires an accountable human decision. Evaluate whether the reviewer receives enough evidence and can actually refuse or stop execution.

What if existing evidence already answers the question?

Review its scope, provenance and configuration before proposing new work. Existing work from product executives and control owners may already be sufficient. In that case, capture the decision, unresolved limitations and the change that would require another review. Additional assessment should address a real gap.

What would make the findings inconclusive?

Missing action inventories, reversibility, approval rules and operator intervention paths can prevent a reliable interpretation, depending on the case. Record the specific missing observation and its implication. The owner can supply evidence, narrow the claim or defer the dependent decision; an incomplete case should not silently be counted as a pass.

Start with the decision you need to make.

Tell us about the agent workflow you want reviewed, and where you need greater clarity. Scope and availability are confirmed together.

Book a 30 minute readiness call Get the agent release checklist

Connected disciplines

Policy-compliance evaluation

Translate operational policies into testable agent behaviour. The review connects approved policy text, precedence rules, exceptions and action logs to a bounded review of evaluation evidence, giving the release owner a clearer basis for the next action.

Agent release-gate design

Make the release decision depend on explicit evidence and accountable exceptions. The review connects release criteria, critical failure definitions, test reports and exception records to a bounded review of evaluation evidence, giving the release owner a clearer basis for the next action.

Continuous production evaluation

Refresh evaluation evidence as live workflows and dependencies change. The review connects sanitised production samples, change events, drift signals and incident classifications to a bounded review of evaluation evidence, giving the release owner a clearer basis for the next action.