Framework · AI strategy
Prerequisites for Strategic AI Adoption
Most AI adoption stalls because teams skip the groundwork. This is the six-part framework we work through with every leadership team before any serious AI investment. Use it as a self-audit or as the basis of a workshop with your leadership team.
Most AI adoption stalls for the same reason. Teams buy licences, turn on Copilot, run a lunch-and-learn, and then wonder why, six months later, nothing has really changed. The workflows look the same. Staff feedback is mixed. Leadership asks where the ROI went.
The problem is not the tool. The problem is that AI adoption was treated like a software rollout, when it is really a change programme.
This framework captures the six prerequisites we work through with leadership teams in every AI Adoption Masterclass. If you tick them off, your AI investment has a real chance of producing results. Skip them, and you are gambling.
The six prerequisites
Readiness assessment
Before you can plan where you are going, you need an honest picture of where you are.
What to look at
- Tools in use. What AI tools does your team have access to today? Copilot for Microsoft 365, ChatGPT Enterprise, sector-specific tools like LEAP, Actionstep, or NetDocuments AI features.
- Actual usage. Of the people with licences, how many are using it weekly? Microsoft and Anthropic usage data shows most licensed users touch AI for less than five hours a week, and many not at all.
- Where value is already coming from. Which tasks are getting faster? Which teams are adopting fastest? What have they worked out that could be shared?
- Where value is leaking. Security concerns, accuracy issues, wasted time on prompts that do not produce useful output.
- Data readiness. Is your documentation where AI tools can find it? Is sensitive data properly classified? Are permissions correct?
Why it matters
You cannot write a strategy for a fleet if you have not counted the cars.
Output of this step
A one-page summary of current AI usage, adoption rates, and any known quality or compliance issues.
Objectives setting
A strategy without objectives is a wish list.
Too many AI strategies read like a tour of the technology. "We will adopt Copilot, implement an internal chatbot, and explore Custom AI." That is a shopping list, not a strategy.
What good objectives look like
- Tie each objective to a business outcome the leadership team already cares about: billable hour recovery, faster matter turnaround, reduced compliance exposure, better client responsiveness.
- Make it specific enough to measure. "Reduce the time lawyers spend on file notes by 40 per cent by end of financial year" beats "use AI to save time".
- Make it realistic given your readiness assessment. If only 10 per cent of your team is using Copilot weekly, a goal that assumes full adoption in 90 days will fail.
- Prioritise ruthlessly. Three objectives you will deliver beat twelve that sit in a deck.
Why it matters
Objectives are the filter you run every future AI decision through. If a proposed tool or project does not move one of them, it is not worth your attention.
Output of this step
Three to five measurable AI objectives, agreed by the leadership team, with timeframes.
Strategy and prioritisation
With a readiness baseline and a set of objectives, you can decide what to actually do.
Strategy at this stage is not "what AI to buy". It is which workflows to touch, in what order, with what level of investment.
Questions to answer
- Which workflows, if we applied AI well, would move our priority objectives the most?
- Which of those workflows is the lowest risk to start with? High value plus low disruption is the sweet spot for building momentum.
- Which workflows are off limits for now? Client advice that cannot be reviewed. Regulated processes needing sign-off change. Systems due for replacement.
- What is the minimum viable change that could prove value in the next 30 days? Not the full transformation. The smallest useful thing.
Why it matters
Without a priority order, every team pursues their own AI wishlist and nothing gets the investment needed to succeed.
Output of this step
A sequenced roadmap of AI initiatives, each with a clear owner, timeframe, and definition of success.
Governance and guardrails
The single most common reason AI adoption stalls in professional services is that nobody is sure what is allowed.
Staff are cautious about using AI on client work because they have heard about data leakage, accuracy problems, or privilege risks. Leadership is cautious about approving broader use because they have not set the rules. Everyone defaults to using AI for low-value tasks, which is exactly where it produces the least benefit. Governance fixes this by making the rules explicit.
What governance needs to cover
- Approved tools. Which AI tools the firm has approved. Which are not. How a team requests approval for something new.
- Data handling. What information can be pasted into which tools. Where client-identifiable information can and cannot flow. How this intersects with existing confidentiality and privacy obligations.
- Human oversight. Which outputs need human review before they go to a client. Which do not. Who is accountable for errors.
- Disclosure. When AI use needs to be disclosed to clients. When it does not. What the firm’s position is on AI-assisted work for billable time.
- Training and competence. Minimum expected training before someone uses AI on client work. How this is tracked.
Why it matters
Without governance, the most cautious staff avoid AI entirely and the least cautious use it in ways that expose the firm. With governance, everyone knows the rules and operates with confidence.
Output of this step
A written AI usage policy, approved by the leadership team, and a communication plan to roll it out.
Culture and capability
AI adoption is mostly a people problem. The technology is the easy part.
Staff need to believe that using AI well is expected, supported, and safe. If the signals are mixed, people default to caution and the investment does not return what it should.
What culture needs to look like
- Permission. The leadership team is visibly using AI itself. Not just asking others to.
- Learning. Time is allocated for staff to learn. Five hours a week of licensed-but-unused AI access is a learning problem, not a tool problem.
- Sharing. Teams share what works. Prompts that landed, tools they found, things that failed. Internal channels, short Loom videos, lunch-and-learns. All of it compounds.
- Feedback. There is a mechanism for staff to flag AI outputs that were wrong, misleading, or inappropriate. Those get reviewed and the lessons shared.
Practical moves
- Run hands-on training for every team, built around their actual workflows. Generic "here is Copilot" training does not move the needle.
- Assign AI champions in each team. Give them time and visibility.
- Make AI skills part of performance conversations. Not as a stick, as a signal that this matters.
Why it matters
The highest-performing firms with AI are not the ones with the most tools. They are the ones where using AI well is part of how work happens, top to bottom.
Output of this step
A capability-building plan, including training, internal champions, and ongoing enablement.
Leadership alignment
The final prerequisite, and often the hardest to close. The leadership team needs to genuinely agree on what AI is for, what the priorities are, and who is accountable.
What alignment means in practice
- The CEO, COO, IT lead, and practice leaders can describe the AI strategy in roughly the same words.
- There is a single owner for the AI programme, not a committee that never quite decides.
- Budget is allocated for the priority initiatives, including training, not just licences.
- There is a review cadence. Monthly or quarterly, the leadership team looks at progress against the objectives set in step 2 and decides what to change.
- Bad news travels up. If something is not working, the firm adjusts rather than persisting with a failing approach.
Why it matters
AI adoption touches every part of the business. Without executive alignment, middle management gets conflicting signals and staff wait for clarity. A year later, you have spent money and have very little to show for it.
Output of this step
A signed-off AI programme brief, a nominated owner, an agreed review cadence, and an allocated budget.
How to use this framework
You can run this framework as a self-audit with your leadership team. Use the six headings as workshop topics. Give each one 60 to 90 minutes. At the end, you will have a clear picture of where you are ready, where you are not, and what to do about it.
If you would rather not build and facilitate that yourself, the AI Adoption Masterclass is our structured version of this framework. Four sessions over four weeks, with a customised roadmap and working documents at the end. Investment: $3,999.
Either way, the point is to do the work. AI that gets real value out of a professional services firm is not the result of a licence purchase. It is the result of a leadership team that worked through these six prerequisites before spending.
Ready to work through this with us?
The AI Adoption Masterclass is the four-session structured version of this framework. Or book a free consultation and we will help you work out the right next step for your organisation.