Skip to main content

​
What you will learn

Most organisations track logins. That tells you almost nothing. This page shows you what real adoption looks like, how to measure it, and how to use Pathfindr’s maturity scale to place your team honestly and plan the next move.

Spot the signals

Recognise the six signs that AI is genuinely woven into how your team works

Track what matters

Use quantitative and qualitative measures that go beyond usage data

Run the health check

Follow five steps to assess your organisation’s adoption health

Place your team

Use the maturity scale to identify where you are and what to do next

​
Six signals of real adoption

Login frequency is a directional signal, not the full picture. These six signals tell you whether AI is actually changing how work gets done.
AI is the first thing people think of before starting a task. They reach for it automatically, not as an afterthought. This is the strongest leading indicator of genuine adoption.
People share wins and teach each other without being asked. When someone shows a colleague a better way to do something using AI, that is organic adoption. You cannot force this. You can only create the conditions for it.
People work together inside shared AI tools and projects. AI stops being a solo activity and becomes part of how teams operate, shared prompts, shared projects, shared knowledge bases.
People feel more capable over time, not less. If confidence is flat or declining, something is wrong with the rollout, not with the people.
Work is happening that simply was not possible before. Not just faster versions of old tasks, but entirely new outputs, analyses, or capabilities that AI enables.
A recurring piece of work now has a step that runs through AI. This is the most concrete signal. If you can point to a specific workflow that has permanently changed because of AI, adoption is real.
Pick one signal to focus on. Pick one measure to track it. Review regularly. That is all you need to start.

​
What to actually track

You need both quantitative and qualitative measures. Usage without confidence is forcing it. Confidence without change is wishful thinking.

Quantitative measures

Reasoning model usage. Are people using the more capable models for real work, or just the basics?Login frequency trending up. A directional signal. Useful when combined with other data, not meaningful on its own.Workflows with AI in them. Count the recurring processes that now include an AI step. This is your strongest quantitative indicator.

Qualitative measures

Confidence pulse. A simple 1 to 5 survey run quarterly. Are people feeling more capable over time?Peer sharing. Is knowledge flowing between people without you organising it? Are Champions teaching others unprompted?Workflow stories. Can people describe, in their own words, how AI changed a specific piece of work? Stories are more telling than numbers.

​
Five steps to AI adoption health

This is the process for running a proper adoption health check. Follow these steps to build a complete picture of where your organisation stands.
1

Run the sentiment pulse survey

Targeted questions at week 6 and week 12 of your rollout. The survey measures confidence and how AI fits into real work. Pathfindr will review your results and provide guidance on what to focus on next.
2

Pull usage data from your AI platform

Login frequency and feature depth give you a directional signal. Access your admin console: Microsoft Copilot (Admin Center, Reports, Usage), ChatGPT Enterprise (Workspace Settings, Workspace Analytics), Claude (Settings, Analytics), or Google Gemini (Admin Console, Generative AI, Gemini Reports). Take a snapshot over the last 30 days.
3

Track AI wins and opportunities

Capture where AI is creating value. Run this monthly using the Opportunity Tracker. Every use case, whether it worked or not, goes in the tracker.
4

Combine all sources in one project

Bring your survey results, usage data, and opportunity tracking into a single AI project for analysis. This gives you one place to see the full picture.
5

Run the analysis prompt

Use the Pathfindr adoption health analysis prompt with your uploaded data. It produces a one-page leadership update covering four dimensions: Confidence and Skill, Impact and Habit, Culture, and Business Outcomes.

​
Platform fitness score setup guides

Follow your tool’s respective guide to capture a consistent AI Fitness Score for your organisation:

ChatGPT

Read more

Copilot

Read more

Gemini

Read more

​
The analysis prompt

Copy and paste this prompt into your AI tool along with your uploaded data from the previous steps. 
You are an AI adoption health analyst for a programme lead. I have uploaded my team's adoption data from three sources: our sentiment pulse survey results, AI platform usage data, and AI win/opportunity tracking.

Analyse this data and produce a one-page leadership update covering four dimensions:

Confidence and Skill
Rate as Strong, Developing, or Needs Attention. How capable is the team with AI and is that growing or flat?

Impact and Habit
Rate as Strong, Developing, or Needs Attention. Is AI changing how work actually gets done, or is it still used for one-off tasks?

Culture
Rate as Strong, Developing, or Needs Attention. Are people sharing, learning from each other, and comfortable being open about AI use?

Business Outcomes
Rate as Strong, Developing, or Needs Attention. What measurable value has AI created? Include specific examples where the data supports them.

Then provide:
- Overall maturity placement on the scale: Curious, Exploring, Embedded, or Integral. Justify in one sentence.
- Overall health summary in one sentence
- Top three wins this period with evidence
- Top five AI use cases ranked by frequency or impact
- Top risk that leadership needs to be aware of
- One recommendation that requires senior support or a decision

Rules:
- Use plain language, no jargon
- Be honest. If the data shows a problem, say so clearly
- Keep the entire update to one page
- Base every claim on the data provided. Do not infer what is not there
You may need to iterate a few times to get the output you want. Adjust the prompt based on the data you actually have available. Pathfindr works alongside you on all five steps.

​
The maturity scale

Place your team honestly. This is not about where you want to be. It is about where you are right now.

Level 1: Curious

People are aware of AI but have not tried it. They are interested but unsure where to start. Most tasks still happen the old way.

Level 2: Exploring

People are trying AI for one-off tasks. Some enthusiasm, some hesitation. Usage is inconsistent and mostly individual.

Level 3: Embedded

AI is part of recurring workflows. Teams are sharing knowledge. Confidence is growing. Managers are visibly using AI in their own work.

Level 4: Integral

Removing AI would break named deliverables. The team is expanding beyond original use cases. New starters learn from the team, not from training materials.
The biggest risk is complacency at Level 3. Things are faster, but nothing has fundamentally changed. The quarterly “what if” refresh prevents this.

​
Moving between levels

Each transition has a prerequisite and three actions. Check the prerequisite first. If it is not met, focus there before moving on.
First, check this: Ensure your team has clear guidelines on what is approved, what data is safe to use, and that AI use is actively encouraged.Then: assign one real task per person that starts in AI each week. Pair each person with a peer one level ahead for a 15-minute weekly check-in. Address hesitation directly by making it clear that starting slow is expected and supported.It is working when: people stop asking “am I allowed?” and start asking “how do I do this better?”
First, check this: Your managers need to be visibly using AI in their own work. If the team never sees their manager use it, adoption will stall here.Then: have each manager share one AI workflow they personally use with their team. Each team member selects one recurring workflow and rebuilds it with AI handling one step. Run a fortnightly show-and-tell with 30-minute sessions, screen shares, and real examples only.It is working when: the conversation shifts from “I tried AI for this” to “this is how I do it now.”
First, check this: Confirm that your team is expanding beyond their original use cases. If the same workflows have been unchanged for a month, you are still at Level 3.Then: select one team deliverable and rebuild it as a shared AI workflow the whole team contributes to. Each person identifies the specific deliverable that would not be possible without AI. Shift measurement from usage metrics to outcome metrics, focusing on what work is now possible that was not before.It is working when: removing AI would break named deliverables.

​
Quick checkpoint

Signals known

You can name at least three of the six adoption signals

Measures planned

You know what quantitative and qualitative data to collect

Health check ready

You have the five-step process and the analysis prompt

Maturity placed

You can place your team on the maturity scale honestly