Framework

HCD operating model baseline

Establish how human-centered work enters the organization, becomes evidence, influences delivery, receives oversight, and demonstrates value.

Connected MyKMHub content

Purpose

When to use this baseline

Use it when establishing an HCD capability, joining a new program, diagnosing inconsistent engagement, scaling across teams, or resetting a practice that produces artifacts without reliably changing decisions.

This is a starting structure, not a universal organization chart. The appropriate roles, review depth, and documentation depend on mission, risk, regulation, delivery model, and team maturity.

Operating model

Seven capabilities that must connect

Mandate and outcomes

Why does the HCD capability exist, and what must it improve?

Name the organizational outcomes, user outcomes, service boundaries, and constraints the practice is accountable for influencing.

Demand and prioritization

How does work enter, and what receives attention first?

Use visible intake criteria, capacity assumptions, risk, reach, evidence needs, and decision deadlines rather than relying on proximity or urgency alone.

Evidence lifecycle

How does evidence become a finding, decision, and reusable record?

Define permitted sources, traceability, confidence, validation, retention, access, and the path from research evidence to delivery decisions.

Decision rights

Who recommends, decides, reviews, and resolves disagreement?

Clarify HCD, product, program, accessibility, technical, policy, and operational authority before difficult decisions expose ambiguity.

Delivery integration

Where does HCD change planning and implementation?

Connect discovery, requirements, design, acceptance criteria, implementation review, release decisions, and learning after launch.

Accessibility and responsible practice

Which obligations cannot be traded away silently?

Set accessibility, privacy, consent, ethics, inclusion, security, and human-review expectations as explicit delivery constraints.

Measures and learning

How will the organization know the practice is useful?

Combine experience, adoption, accessibility, delivery, risk, capacity, and organizational-learning evidence without reducing HCD value to activity counts.

Working artifact

Create a minimum operating agreement

Capture one reviewable page before building a detailed playbook. It should state:

  • the mandate, desired outcomes, scope, and accountable owner;
  • how requests enter, are assessed, prioritized, and declined;
  • required evidence and accessibility checkpoints;
  • decision roles and an escalation path for unresolved risk;
  • where HCD connects to planning, delivery, and release;
  • which records must be retained and who may access them;
  • the measures reviewed, their cadence, and resulting actions;
  • the date and evidence that will trigger the next model review.

Treat the agreement as versioned operational knowledge. A model that exists only in presentations or individual memory cannot govern work consistently.

Leadership review

Questions for a Director-level review

  • Can teams explain when and why to involve HCD?
  • Can leaders see demand, capacity, priority, and blocked work?
  • Can a consequential finding be traced to appropriate evidence?
  • Are decision rights clear when user, mission, policy, and delivery needs conflict?
  • Do accessibility findings change requirements and acceptance decisions?
  • Are teams learning after release, or only validating before it?
  • Do measures show changed outcomes, reduced risk, or improved capability?
  • Can another person continue the work from the retained record?

Warning signs

Look beyond the presence of HCD activity

  • Research begins after major commitments are already fixed.
  • Requests depend on personal relationships instead of visible intake.
  • Teams collect evidence but cannot trace decisions back to it.
  • Accessibility is deferred to a final compliance review.
  • Artifacts accumulate without owners, lifecycle, or reuse.
  • Success is reported only as interviews, workshops, or deliverables completed.
  • AI-generated synthesis is accepted without source review or accountable interpretation.

These are operating-system problems. Adding another template or workshop may increase activity without correcting the conditions producing inconsistent outcomes.