Skip to main content

Accessibility Planner

The Accessibility Planner is a phase-based conversational agent that guides your team through a structured accessibility assessment across five frameworks: WCAG 2.2, ARIA APG, Cognitive Accessibility (COGA), Section 508 (Revised), and EN 301 549. It produces framework selections, success-criterion mappings, evidence-register entries, plan-risk classifications, tradeoff logs, and a dual-format ADO + GitHub backlog handoff.

The goal is not to replace accessibility expertise. It is to make sure the right frameworks get selected, the right success criteria get mapped, and the resulting work items land in your backlog with enough evidence to act on.

Purpose​

  • Select the frameworks that apply to the product's surfaces and regulated markets, then capture conformance levels and license posture.
  • Map in-scope surfaces against the selected frameworks and resolve each success criterion to a target compliance state with evidence pointers.
  • Classify plan-level risk, document mitigation versus accept-with-tradeoff decisions, and produce work-item seeds with evidence references.
  • Hand off prioritized backlog items in both Azure DevOps and GitHub formats, with autonomy tiers and a pinned planning disclaimer.

When to Use​

ScenarioRecommended entry mode
Fresh start with no upstream artifactCapture mode
Existing PRD with audience scope and surface inventoryFrom-PRD mode
Business requirements with market and procurement driversFrom-BRD mode
Security plan already produced for the same projectFrom-Security-Plan mode
RAI plan flagging AI-generated UIFrom-RAI-Plan mode

Entry Modes​

Five entry modes determine how Phase 1 (Discovery) begins. All modes converge at Phase 2 once discovery completes.

ModeTriggerInput sourceBehavior
captureFresh startConversationExploration-first questioning to build surface inventory, audience scope, and regulatory drivers
from-prdPRD exists.copilot-tracking/prd-sessions/Pre-populates Phase 1 from audience scope, in-scope surfaces, devices, and accessibility commitments
from-brdBRD exists.copilot-tracking/brd-sessions/Extracts market geographies, procurement obligations, and contractual VPAT or ACR commitments
from-security-planSecurity plan existsSecurity Planner artifactsReuses surface inventory, AI/ML flags, and external-input touchpoints; sets securityPlanRef
from-rai-planRAI plan existsRAI Planner artifactsReuses AI-generated UI flags and audience-impact signals; sets raiPlanRef

The Six Phases​

flowchart LR
  A["Phase 1<br/>Discovery"] --> B["Phase 2<br/>Framework Selection"]
  B --> C["Phase 3<br/>Standards Mapping"]
  C --> D["Phase 4<br/>Plan Risk Assessment"]
  D --> E["Phase 5<br/>Impact & Evidence"]
  E --> F["Phase 6<br/>Backlog Handoff"]

Phase 1: Discovery​

Captures project context, audience scope, surface inventory, regulatory drivers, and existing accessibility artifacts. In capture mode, the agent uses exploration-first questioning to build the inventory from scratch.

Phase 2: Framework Selection​

Presents the five supported frameworks (wcag-22, aria-apg, coga, section-508, en-301-549) using a host-aware multi-select, with wcag-22@AA and section-508 pre-checked as defaults. Captures conformance levels, atomic disabled bundles, and license posture acknowledgements.

Phase 3: Standards Mapping​

Maps in-scope surfaces against the selected frameworks. Each success criterion resolves to a target compliance state, attaches evidence pointers, and emits cross-references for any criterion shared across frameworks.

Phase 4: Plan Risk Assessment​

Classifies the assessment risk tier (low, medium, or high), enumerates plan-level risks such as audience scope versus tested coverage and automation-only coverage, and raises escalations. The agent re-enters capture coaching when discovery gaps surface.

Phase 5: Impact and Evidence​

Produces the evidenceRegister, tradeoffLog, and workItemSeeds arrays in state.json. Documents mitigation versus accept-with-tradeoff choices for each unresolved gap, cross-links to RAI, SSSC, and Security Planner artifacts when present, and captures VPAT or EAA evidence references.

Phase 6: Backlog Handoff​

Renders Phase 5 outputs into dual-format ADO + GitHub backlog files, applies the review rubric, attaches autonomy tiers, sanitizes content, and emits the planning disclaimer block.

Outputs​

ArtifactDescription
Framework selectionsThe frameworks in scope, with conformance levels and disabled-bundle records
Control mappingsSuccess criteria mapped to surfaces with target compliance states
Evidence registerStable-URI evidence records reusable by Security, SSSC, and RAI planners
Tradeoff logMitigation versus accept-with-tradeoff decisions for unresolved gaps
Backlog handoffDual-format ADO + GitHub work items with autonomy tiers and disclaimer

State persists at .copilot-tracking/accessibility/{project-slug}/state.json, validated against scripts/linting/schemas/accessibility-state.schema.json.

Cross-Planner Integration​

The Accessibility Planner shares evidence and signals with the Security, RAI, and SSSC planners through reference fields and a shared evidence-register schema.

Outbound flowWhat is shared
Accessibility → SecurityEvidence records carry stable URIs that security reports cite under external evidence
Accessibility → RAIPhase 4 inserts humanReviewControl entries when the profile declares AI-generated UI, alt text, or captions
Accessibility → SSSCSection 508 and EN 301 549 evidence records feed SSSC procurement gates such as VPAT and EAA conformance

See Cross-Planner Integration for the full integration matrix, including the inbound flows from the RAI and Security planners.

File typeLocation
Agent.github/agents/accessibility/accessibility-planner.agent.md
State schemascripts/linting/schemas/accessibility-state.schema.json
Instructions.github/instructions/accessibility/

Next Steps​

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.