Bite 1 / 7
Lesson 4.1

Why assess maturity

You'll be able to Explain why a fixed start point is needed before planning the AI journey.

You can't plan a journey without a start point

A maturity assessment is the honest 'you are here'. Microsoft's own warning stands: without clear success criteria, organisations cannot tell whether Copilot delivers value. Score the start, then plan the route.

Check what you know

What is the main reason to assess maturity first?

What is the main reason to assess maturity first?
Takeaway

No start point, no journey. Score 'you are here' before planning.

Lesson 4.2

Building a maturity model

You'll be able to Name the five dimensions to score and the levels Microsoft's model uses.

Five dimensions, five levels

Microsoft's Agentic AI adoption maturity model scores five dimensions — AI strategy and experience, business strategy, AI governance and security, technology and data, and organisation and culture — across levels from 100 (initial) up to 500 (efficient).

Score these five

  1. Strategy & experience — is there a formal AI strategy with an executive sponsor?
  2. Business strategy — are business metrics and human-agent teaming defined?
  3. Governance & security — is Responsible AI embedded, not bolted on?
  4. Technology & data — are platforms and data foundations ready and reusable?
  5. Organisation & culture — are champions and enablement active?

Check what you know

How many dimensions does Microsoft's maturity model score?

How many dimensions does Microsoft's maturity model score?
Takeaway

Score five dimensions (strategy, business, governance, technology, culture) on a 100–500 scale.

Lesson 4.3

Feeding the assessment with real Microsoft signals

You'll be able to Match each scoring dimension to a real Microsoft assessment that informs it.

Use real instruments, not guesses

Score honestly by feeding the model with evidence: the SAM Content Management Assessment, Purview DSPM data risk assessments, and the admin-center Copilot readiness report.

Signal to dimension

Check what you know

Which is a real Microsoft signal to feed the assessment?

Which is a real Microsoft signal to feed the assessment?
Takeaway

Score with evidence: SAM assessment, Purview DSPM, the readiness report.

Lesson 4.4

Current-state baseline

You'll be able to Produce an honest, evidence-based current-state score across the five dimensions.

Be honest, cite evidence

For each dimension, record the level (100–500) and the evidence behind it. Resist scoring high to look good — the assessment only helps if it is honest.

Check what you know

What makes a current-state baseline useful?

What makes a current-state baseline useful?
Takeaway

Score honestly with evidence — a flattering baseline is a useless one.

Lesson 4.5

Target-state definition

You'll be able to Define a realistic target state at 6 and 12 months, tied back to the Stage 1 strategy.

Where this organisation needs to be

Set target levels for 6 and 12 months for each dimension. Anchor them to the outcomes you defined in Stage 1 — the target is in service of the strategy, not a score for its own sake.

Check what you know

What should the target state be tied back to?

What should the target state be tied back to?
Takeaway

Set 6- and 12-month targets tied to the strategy, not to the scoreboard.

Lesson 4.6

Gap analysis to roadmap

You'll be able to Turn the gap between current and target state into a sequenced plan that drives Stages 5–9.

The delta is your plan

The distance between current and target, per dimension, is the work. Sequence it into a roadmap: what to do in Envision, Deploy, Adopt, Run and Innovate. The biggest, highest-risk gaps go first.

From gap to roadmap

  1. List gaps — current vs target, per dimension.
  2. Prioritise — by value and risk.
  3. Sequence — map onto Stages 5–9.
  4. Own — assign each item to a person and a date.

Check what you know

What is the output of gap analysis?

What is the output of gap analysis?
Takeaway

The gap is the roadmap — prioritise by value and risk, then assign owners and dates.

Wrap-up

Sources & notes

Everything here traces to Microsoft Learn, retrieved June 2026.

Accuracy flags

  • The original taxonomy said Microsoft does not publish a maturity model. As of 2026, Microsoft DOES publish the 'Agentic AI adoption maturity model'. This stage now cites it.
  • A standalone, single 'Copilot Optimization Assessment' page was not retrieved this run. The grounded instruments used here are the SAM Content Management Assessment, Purview DSPM data risk assessments, and the admin-center Copilot readiness report. Verify the exact name 'Copilot Optimization Assessment' before quoting it.
Fonts (Anton, Archivo Black, Inter) stand in for the licensed LearningBites faces — swap before publishing. Microsoft's Copilot and agent surface changes monthly; re-verify before reuse.