Softlander Book a call
Agent identity, authorisation, and delegation

Agent authorisation modelling

Define which resources and actions an agent may use in context

Role names no longer express the constraints needed for agent actions. The decision is how to define which resources and actions an agent may use in context. 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 authorisation architects and application 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 identity and delegated authority relationship

  1. Starting evidenceResource models

    Establish the basis for acting subject.

  2. Decision boundaryRequested operation

    Distinguish read access from modification authority.

  3. Decision artefactAuthorisation model

    Retain the scoped finding and its unresolved dependencies.

Conceptual model. Technical findings do not settle a disputed mandate.

Decision

Frame the decision around acting subject

Questions and the evidence they need

Criterion / evidenceActing subject
Inspect resource models for the relevant constraint or observation.
Criterion / evidenceRequested operation
Inspect action catalogues for the relevant constraint or observation.
Criterion / evidenceEnforcement context
Inspect subject attributes 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 define which resources and actions an agent may use in context. This is relevant to authorisation architects and application owners when role names no longer express the constraints needed for agent actions. 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 a broad role permits changes to any customer record rather than the assigned case. 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

Permission depends on subject, action, resource, and context

Relationships within identity and delegated authority

Source or conditionActing subject
Resolve attributes used by the policy.
Crossing or choiceRequested operation
Distinguish read access from modification authority.
Consequence or decisionEnforcement context
Locate the decision before a resource changes.
Follow the check to the accepting resource. Each relationship needs evidence in the actual scope.

What this specialty examines

An authorisation model should represent the action and resource precisely enough for the business rule. Read, search, update, export and administer should not become interchangeable merely because one API groups them together. Context may include assignment, tenant, purpose or approval state.

The review should model allowed and denied subject-action-resource-context combinations and locate enforcement points. 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, richer context supports precise decisions while increasing policy complexity and testing effort. 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 acting subject

Illustrative working format

BasisResource models
Match this description to the deployed or proposed boundary being examined.
ContextAction catalogues
Retain the observed conditions and distinguish gaps from contrary evidence.
ObservationSubject attributes
Name the custodian and the conditions requiring a refreshed version.
DependencyPolicy decision records
Confirm the applicable revision and who resolves disputed interpretation.
Illustrative preparation record. Availability, revision, and handling need confirmation.

Start with resource models, action catalogues, subject attributes and policy decision records. 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 identity control owner, agreed expectations for a representative subject-action-resource case, and the authority to inspect the relevant boundary. Specify what evidence would distinguish the failure in which a broad role permits changes to any customer record rather than the assigned case 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 acting subject through to an accountable finding

From the scoped material to a decision record

  1. Establish conditionsResource models

    Locate the source of the claimed authority.

  2. Examine the relationshipRequested operation

    Apply the scoped check to enforcement context.

  3. Retain the findingAuthorisation model

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

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

Map principals, resources and authority

Reconcile resource models, action catalogues, subject attributes and policy decision records with the workflow described by authorisation architects and application owners. Identify the initiating actor, executing workload, target resource and component that grants access. Prepare paired cases that differ in one material authority condition, such as assignment, tenant, expiry or approval. The preparation output is an action map showing which permission is expected at each boundary and which records can verify it.

Exercise and review the authority path

Build decision tables before encoding policy. Include absent attributes, conflicting grants and resources at boundary conditions. Explain which component makes the decision and which actually enforces it; an advisory policy response does not constrain an unchecked tool.

Select the identity control change

Consolidate the checked paths into an authorisation model with decision tables and unresolved policy semantics. Distinguish an identity-verification problem from an excessive permission, lost delegation context or delayed withdrawal. Assign the proposed change to the component that controls the affected action. Before recommending wider use, identify the allowed case and denied counterpart that would demonstrate the revised boundary without breaking legitimate work.

Indicative artefact

Make the authorisation model usable by its owner

Authorisation model

Illustrative working format

Annotated fieldAuthorisation model
Identify the assessed workflow, revision, and decision boundary.
Annotated fieldDecision tables
Link supporting and contrary observations to inspectable evidence.
Annotated fieldUnresolved policy semantics
Map source meaning to receiving meaning and record unresolved interpretation.
Illustrative artefact structure, not a completed engagement or an observed result.

A useful decision artefact is an authorisation model with decision tables and unresolved policy semantics. 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 enforcement context through a constructed exception

A constructed case and its decision paths

Illustrative scenario

  1. SituationActing subject

    A broad role permits an operation outside its intended resource scope.

  2. Question to examineRequested operation

    Compare permitted and denied resource-action combinations for the same role.

  3. Conditional next actionEnforcement context

    Narrow the grant when its scope exceeds the business mandate.

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

Consider a fictional situation: a support agent should edit contact preferences for an assigned case but not export the entire customer collection. The scenario is illustrative, with no client engagement, executed test or measured improvement implied.

For assigned-case editing, test the same agent against assigned and unrelated records. A narrow allow result should depend on the actual assignment and action. The owner can then decide how exceptional cross-case work receives explicit authorisation.

Scope and next action

Keep the conclusion within named authority paths

Confirm fit before the next commitment

Included boundaryActing subject
This review is limited to named authority paths.
PrerequisiteResource models
The relevant owner must confirm access and the intended operating conditions.
Claim limitAuthorisation model
Technical findings do not settle a disputed mandate.
Bring the triggering change and existing evidence; resolve the dependency that could alter the decision.

Identity design and technical testing do not establish a legal mandate or resolve a disputed business authorisation. Qualified IAM review is needed for material control conclusions. 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 role names no longer express the constraints needed for agent actions; confirm which missing observation would change the owner’s next action.

Ways to engage

Related engagement scopes

Agent Release Readiness Review

The Agent Release Readiness Review is a fixed-scope, three-week independent test of one AI agent workflow that shows a release board what the agent does, what it is allowed to do and what happens when it is wrong, and ends in a signed recommendation.

3 weeksRead about this offer

AI Agent Authorisation Review

Make delegated authority inspectable. Trace who an agent acts for, what it can access and how that access can be withdrawn.

Read about this offer

AI Agent Runtime Security Review

Make every delegation explicit. Examine who the agent acts for, where it can execute and which tools it can reach when instructions alone are not enough.

Read about this offer

Continue exploring

Related insights and resources

Whitepaper15 MIN READ

Designing agent authority boundaries

Turn agent permissions into an enforceable action contract. Separate a model's suggestion from a resource owner's authorisation, test the boundary, and make delegation and revocation reviewable.

Agent security6 MIN READ

The permission boundary behind every agent action

An agent can only be evaluated in the context of what its tools allow. Examine identity, delegated authority and action boundaries before you put a workflow in motion.

Whitepaper5 MIN READ

Before your agent can act, define its authority

A working paper for the moment an AI system moves from advice to action. Map identity, permissions, approvals and recovery into an authority model your team can review and test.

A little more clarity

Questions to consider.

Start with the questions that shape a better decision.

Should authorisation be role-based, attribute-based or action-specific?

Choose the model that can express the actual business boundary and be operated reliably. Roles may suffice for broad duties; resource, assignment, tenant or approval conditions may require additional attributes or action-specific rules.

Which resource and action records inform authorisation?

Start with resource models, action catalogues, subject attributes and policy decision records. 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 missing or conflicting attributes be tested?

Model allowed and denied subject-action-resource-context combinations and locate enforcement points. Observe the authenticated principal, mandate and resource decision separately. Include a permitted action and a denied counterpart so the review can establish whether authority is narrow enough without making legitimate work impossible.

What should an authorisation model specify?

The proposed output is an authorisation model with decision tables and unresolved policy semantics. 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 is finer-grained context worth the complexity?

Finer context is justified when broad roles cannot express a material resource, tenant, assignment or approval boundary. Keep the model understandable and testable; additional attributes without reliable provenance can create ambiguity.

What if existing evidence already answers the question?

Review its scope, provenance and configuration before proposing new work. Existing work from authorisation architects and application 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 resource models, action catalogues, subject attributes and policy decision records 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

Delegated authority design

Make delegated authority explicit and bounded. The review connects delegation records, action limits, expiry rules and authority transfer diagrams to a bounded review of identity and authority, giving the identity control owner a clearer basis for the next action.

User-to-agent delegation controls

Preserve a person's intended limits when an agent acts for them. The review connects user journeys, approval records, consent text and action parameters to a bounded review of identity and authority, giving the identity control owner a clearer basis for the next action.

Agent-to-agent delegation controls

Prevent authority from expanding as agents delegate to one another. The review connects delegation chains, agent role definitions, resource policies and downstream tool logs to a bounded review of identity and authority, giving the identity control owner a clearer basis for the next action.