Skip to main content
Assessment & Diagnostics/ai-maturity-org

AI Maturity Org

You need an organization-level AI maturity profile.

Use this when you need to assess a client organization's readiness to adopt AI practices — at engagement kickoff, during a quarterly review, or when organizational friction is blocking AI adoption. Evaluates 10 dimensions and outputs a maturity profile with scores, weakest-link analysis, and a prioritized adoption roadmap.

Related skills: To assess platform team effectiveness (not org readiness), use /ai-platform-maturity-assess. To design the platform operating model, use /ai-platform-operating-model.

Process

Step 1: Set context

Ask the assessor:

  1. What organization are you assessing? (Client name, team, or business unit.)
  2. What is the context? (Engagement kickoff, quarterly review, or specific adoption challenge.)
  3. How many people did you speak with to form your assessment? (Ideally 2-3 across tech, product, and leadership.)
  4. Are there any known constraints? (Regulatory environment, recent leadership changes, active transformation programs.)

Step 2: Walk through the 10 dimensions

For each dimension, present the 5-level rubric and ask the assessor to score the organization. Go through them in order:

Core dimensions (1-6):

  1. Prompt & Interaction Quality — Are there shared prompting practices, or is it ad hoc individual experimentation?
  2. Evaluation Discipline — Does the org review AI outputs systematically, or is it hit-or-miss?
  3. Workflow Integration — Are AI-assisted workflows documented and repeatable, or scattered individual usage?
  4. Context & Knowledge Management — Is organizational context structured for AI consumption, or ad hoc?
  5. Governance & Bounded Autonomy — Are there clear policies on what AI can do autonomously, or informal rules?
  6. AI Foundations — Does the org understand AI capabilities and limitations, or are misconceptions common?

Organizational factors (7-10): 7. Tool Access & Infrastructure — Are AI tools provisioned and integrated, or do people use personal accounts? Behavioral indicators:

  • Level 2: People use personal accounts for AI tools; no org-wide provisioning
  • Level 3: Org provides developer-tier tools (Cursor, Claude Code, Copilot); may also provision prompt-to-app tools (Lovable, v0) for PMs and designers
  • Level 4: Different tool tiers provisioned by role -- engineering amplifiers for developers, vibe coding tools for product/design, agent orchestration platforms for complex workflows
  • Level 5: Tool strategy reviewed quarterly; new categories (agent orchestration, specialist builders) evaluated proactively as the market evolves
  1. Change Management & Culture — Is there psychological safety around AI experimentation, or fear and resistance?
  2. Data Governance & Privacy — Are there clear rules on what data flows through AI tools, or is it unclear?
  3. Leadership Alignment — Is AI a strategic priority with executive sponsorship, or a grassroots effort?

For each dimension, use this rubric:

  • 1 — Not Yet Started: No activity in this area
  • 2 — Growing: Individual experimentation, no shared practices
  • 3 — Meets Expectations: Documented, repeatable, team-level adoption
  • 4 — Exceeds Expectations: Systematic, integrated, cross-team alignment
  • 5 — Leading: Industry-leading, continuous innovation

Step 3: Generate the maturity profile

# Organizational AI Maturity Profile

**Organization:** (Name)
**Assessed by:** (Name and role)
**Date:** (Date)
**Context:** (Kickoff / Quarterly / Ad hoc)

---

## Dimension scores

| # | Dimension | Score | Level |
|---|-----------|-------|-------|
| 1 | Prompt & Interaction Quality | X/5 | (Level name) |
| 2 | Evaluation Discipline | X/5 | (Level name) |
| 3 | Workflow Integration | X/5 | (Level name) |
| 4 | Context & Knowledge Management | X/5 | (Level name) |
| 5 | Governance & Bounded Autonomy | X/5 | (Level name) |
| 6 | AI Foundations | X/5 | (Level name) |
| 7 | Tool Access & Infrastructure | X/5 | (Level name) |
| 8 | Change Management & Culture | X/5 | (Level name) |
| 9 | Data Governance & Privacy | X/5 | (Level name) |
| 10 | Leadership Alignment | X/5 | (Level name) |

**Overall maturity:** X/5 (weakest-link — determined by the lowest score)
**Average score:** X.X/5

## Weakest-link analysis

(Identify the 2-3 dimensions pulling the overall score down. For each, explain why it matters and what delivery risk it predicts.)

## Strengths

(Identify the 2-3 strongest dimensions. Acknowledge what the org is doing well.)

## Prioritized adoption roadmap

### Immediate (next 2 weeks)
- (Action targeting the lowest-scoring dimension)
- (Action targeting the second-lowest)

### Short-term (next 1-3 months)
- (Actions to bring weakest dimensions to level 3)

### Medium-term (next 3-6 months)
- (Actions to deepen strengths and address remaining gaps)

## AHI risk mapping

(For any dimension scoring below 3, flag the predicted delivery risk and recommended monitoring.)

| Weak dimension | Predicted risk | When it typically surfaces |
|----------------|----------------|--------------------------|
| (Dimension) | (Risk) | (Timing) |

Step 4: Review and discuss

Present the profile and ask:

  • Does this match your impression of the organization?
  • Are there contextual factors that explain low scores? (e.g., recent leadership change, pending tool procurement.)
  • Which roadmap actions feel most achievable in the near term?
  • When should we reassess? (Typically quarterly, or after major changes.)

Output location

Present the maturity profile as formatted text in the conversation. The assessor can copy it into their engagement documentation or present it to leadership.

Example Output

Input

  • Organization: Meridian Health Partners — Digital Products team (~120 people across product, engineering, and clinical informatics)
  • Context: Engagement kickoff; brought in to accelerate AI adoption after a failed GitHub Copilot rollout 6 months ago stalled momentum
  • Interviews conducted: 3 — VP of Engineering (Marcus T.), Director of Product (Sela R.), and a Senior Clinical Informatics Analyst (unnamed)
  • Known constraints: HIPAA and state-level health data regulations; recent CISO change creating policy uncertainty; an active EHR migration program competing for engineering bandwidth

Output (abbreviated)

Organizational AI Maturity Profile

Organization: Meridian Health Partners — Digital Products Assessed by: Jamie Okafor, Engagement Lead Date: June 12, 2025 Context: Engagement kickoff


Dimension scores

#DimensionScoreLevel
1Prompt & Interaction Quality2/5Growing
2Evaluation Discipline2/5Growing
3Workflow Integration2/5Growing
4Context & Knowledge Management2/5Growing
5Governance & Bounded Autonomy1/5Not Yet Started
6AI Foundations3/5Meets Expectations
7Tool Access & Infrastructure3/5Meets Expectations
8Change Management & Culture2/5Growing
9Data Governance & Privacy1/5Not Yet Started
10Leadership Alignment3/5Meets Expectations

Overall maturity: 1/5 (weakest-link — determined by lowest score) Average score: 2.1/5


Weakest-link analysis

Governance & Bounded Autonomy (1/5) and Data Governance & Privacy (1/5) are the critical blockers. In a regulated healthcare environment, the absence of any policy defining what AI can do autonomously — or what data can flow through AI tools — is not a gap, it's a risk exposure. The failed Copilot rollout traces directly here: engineers used it without clarity on whether PHI-adjacent code or comments were in scope. The CISO transition has created a vacuum, and until it's filled, informal workarounds will accumulate. This predicts compliance incidents and will trigger leadership to freeze AI programs entirely if an incident surfaces.

Change Management & Culture (2/5) is the third drag. Post-rollout failure, there is visible skepticism in the engineering org ("we tried this and got burned"). Sela R. noted that PMs are quietly using ChatGPT on personal accounts because requesting enterprise tools "feels like it goes nowhere." Without psychological safety and a visible win, even well-designed interventions will face low participation.


Strengths

  • AI Foundations (3/5): The clinical informatics team has solid baseline literacy — they understand model limitations and are not over-indexing on AI capabilities. This is rare in regulated orgs and gives a credible anchor for broader education.
  • Leadership Alignment (3/5): The VP of Engineering and CPO are aligned that AI adoption is a 2025 priority. There is budget allocated and genuine sponsorship — the organization wants this to work, which is meaningfully different from orgs where adoption is purely grassroots.
  • Tool Access & Infrastructure (3/5): Despite the rollout failure, Copilot licenses remain provisioned. There is willingness to add Cursor for senior engineers and explore PM-tier tools. The infrastructure foundation is present; the problem was never tooling, it was governance.

Prioritized adoption roadmap

Immediate (next 2 weeks)

  • Convene a one-day Data Governance sprint with incoming CISO, VP Engineering, and Legal to produce a one-page AI Data Use Policy — classify what is permissible (non-PHI code, internal documents) vs. prohibited (patient records, identifiers) in AI tools. Ship it as a decision, not a committee draft.
  • Run a retrospective on the Copilot rollout with the 5-6 engineers who were most active. Surface what actually failed — the goal is to defuse the narrative that "AI tools don't work here" and replace it with "we lacked guardrails, now we're building them."

Short-term (next 1–3 months)

  • Draft and ratify an AI Governance Charter covering bounded autonomy — define what AI can do without human review (boilerplate code suggestions) vs. what requires sign-off (clinical decision logic, external-facing copy).
  • Establish a fortnightly AI practice session (30 min) for product and engineering leads to share prompting patterns — move Prompt & Interaction Quality from ad hoc to documented.
  • Relaunch Copilot with the new data policy, limited to 2 squads not involved in EHR migration, with explicit evaluation checkpoints.

Medium-term (next 3–6 months)

  • Build a shared context library: product briefs, user research summaries, and clinical terminology guides structured for AI consumption — addresses Context & Knowledge Management gap.
  • Design and run an AI Evaluation Playbook workshop — teams learn to review AI outputs systematically before promoting to production.
  • Expand tool provisioning to PM and design tiers (e.g., Notion AI, v0 for prototyping) with role-appropriate guardrails.

AHI risk mapping

Weak dimensionPredicted riskWhen it typically surfaces
Data Governance & Privacy (1/5)PHI inadvertently passed to AI tools via code comments or pasted documents; compliance breach triggers program freezeWithin 60 days of any expanded rollout without policy
Governance & Bounded Autonomy (1/5)AI-generated clinical-adjacent content published without review; incident erodes trust and invites regulatory scrutinyFirst time an AI workflow touches patient-facing output
Change Management & Culture (2/5)Low adoption of any new initiative; teams comply in meetings, revert to prior habits in executionWeeks 3–6 of any rollout, after initial enthusiasm fades
Workflow Integration (2/5)Productivity gains remain individual and invisible; leadership loses confidence in ROI and reduces sponsorshipQuarterly business review where AI impact cannot be demonstrated