Here's a sentence I've heard in some form at least a dozen times this year: "AI is central to our firm's future."
And here's the thing - it's true. It's also true of every other firm in your sector. Every firm in the sector next to yours. The 150-person consultancy down the road, the Magic Circle firm, the two-partner outfit in Cheltenham. Saying AI is central to your future is like saying electricity is central to your operations. Technically correct. Strategically meaningless.
I'm not trying to be glib. I know the intent behind it is genuine. You've seen what's coming. You've sat through the demos, read the McKinsey reports, maybe even attended one of those vendor dinners where someone shows you an AI tool doing something that genuinely makes you think right, we need to be on this. And you probably went back to your partners and said something along the lines of: we need AI in our strategy.
So you put it in.
We're committed to being an AI-enabled firm. We will use AI to deliver better outcomes for our clients. AI differentiation is a strategic priority.
And then... nothing materially changed.
If that stings a bit, stay with me. Because you're not alone, and this isn't an individual failure. It's a structural pattern I see across almost every mid-market professional services firm I talk to. The ambition is real. The specificity is absent. And the gap between those two things is where twelve months of potential progress quietly disappears.
Let me describe something and you tell me if it sounds familiar.
Your firm's strategy document - or maybe just the slide deck from the last partners' retreat - includes a section on AI. It says something about exploring use cases, investing in capability, staying ahead of the curve. The language is confident. The partners all nodded when it was presented. Someone might have even used the phrase "AI-first" without anyone in the room asking what that actually means in practice.
Here's what it doesn't include: a named application. A timeline. A budget. An owner. A success metric. A definition of what "differentiation" looks like once you've achieved it.
McKinsey's 2025 State of AI report found that 71% of organisations are now regularly using generative AI. But only about a third are successfully scaling it across the organisation. Just 42% believe their AI strategy is actually ready for execution. Meanwhile, 88% of broader transformation programmes fail to meet their original ambitions.
So the vast majority of firms are using AI in some form. The vast majority are also failing to turn that usage into anything structured, measured, or strategically meaningful. There's a word for this gap. It's not a strategy gap - it's a specificity gap.
I was in a conversation a few months ago with the managing partner of a mid-sized professional services firm. Good firm, strong reputation, about 250 people. He told me, quite proudly, that AI was one of their three strategic pillars for the year. When I asked what specific AI initiative they were delivering in Q2, he paused. Then he said, "Well, we're exploring a few things."
That's the vague ambition trap in one sentence. "Exploring a few things" isn't a strategy. It's a holding pattern.
I'll be honest - I've been guilty of this myself. Early on, when we were first building out our own AI thinking at Distinction, I had a version of that same conversation from the other side of the table. Someone asked me what we were actually doing with AI, and I gave a pretty polished non-answer about capability-building and staying close to the tools. It sounded fine. It meant nothing. The difference between then and now is that someone eventually pushed back hard enough to make me uncomfortable, and I had to get specific. Most managing partners never get that push.
I've watched this pattern enough times now to see the mechanics clearly. There are a few ways that vague AI ambition converts into no AI progress, and they tend to operate simultaneously.
The most obvious one: if "AI differentiation" doesn't have a named application attached to it, there's no project to resource. No team to own it. No result to measure. It stays permanently in the category of "strategic priority" - the kind that gets a warm discussion at the annual planning session and then gets deferred because nobody could agree on where to start. I've seen firms have the same AI conversation three years running. Same enthusiasm, same vagueness, same outcome.
Closely related is the absence of any success metric - and this one's subtle but devastating. Without a defined outcome, any result can be interpreted as progress. A partner uses ChatGPT to draft a client email and mentions it at a team meeting? Progress. Someone attends an AI webinar? Progress. The firm buys a licence for a tool that six people use twice and then forget about? Progress. There's no forcing function for the honest conversation about whether anything has actually changed at a firm level. And so the honest conversation never happens.
But the thing that really kills it is accountability - or the absence of it. When AI differentiation is a partnership-level strategic priority, owned collectively, it's owned by nobody. You get to the next strategy day, someone asks "where are we on AI?", and three different partners each point to the same handful of individual experiments. Oh, Sarah's been using that document review thing. Tom's team tried something with client comms. Everyone nods. The box gets ticked. And the firm is in exactly the same position it was twelve months ago, except now the competitors who actually committed to something specific are twelve months ahead.
But we have AI in our strategy. Our partners know it's a priority. We're exploring use cases and investing in the capability. That's a strategy.
I hear you. And I'm going to push back - because what you've described is aspiration, not strategy. Aspiration says "we believe in this." Strategy says "here's what we're doing about it, who's responsible, what success looks like, and when we'll know." The gap between those two things isn't semantic. It's the gap between firms that will have meaningful AI capability in two years and firms that will still be "exploring."
Let me give you three examples. Not hypothetical moonshots - these are things a mid-market professional services firm could realistically commit to and deliver within a quarter.
Example one: "We will use AI to reduce the time our fee earners spend on first-draft document review by 40%, measured across all due diligence matters in Q3, with adoption confirmed by our operations lead by the end of the quarter."
Example two: "We will deploy AI-assisted knowledge search across our matter archive by the end of H1, with adoption above 60% among associates confirmed by a usage audit in month three of operation."
Example three: "We will use AI to generate the first draft of all standard client update emails by the end of Q2, with a quality score above 4/5 from reviewing fee earners."
Each one has a specific outcome. A timeline. An accountable owner. A measurable success criterion. None of them requires a massive investment or a twelve-month programme. They're achievable, bounded, and the kind of thing you can actually deliver and then point to and say: that worked, or it didn't, and here's what we learned.
Now compare them to "we will use AI to differentiate."
The contrast is pretty glaring, isn't it?
And here's the thing that I think most managing partners find quietly uncomfortable about these examples: they're not hard to write. The specificity isn't difficult to produce. What's difficult is being willing to commit to it. Because once you've written down "40% reduction in first-draft review time by end of Q3, owned by the operations lead," you've created a commitment that someone can fail to deliver. You've made yourself accountable for a specific outcome. And that, honestly, is what most vague AI strategies are designed to avoid. Not consciously. Not cynically. But structurally.
Keeping it vague protects everyone from the discomfort of a named, measurable shortfall. It also guarantees that nothing meaningful happens.
I've written about AI readiness in more detail separately - there's a checklist piece that goes deeper on whether your firm has the foundations for this kind of commitment. But the starting point is simpler than most people expect.
Three questions. If your CTO or managing partner can answer all three for a specific proposed AI application, you've crossed the line from aspiration to strategy. If they can't, you haven't.
Question one: Which specific task will take less time, produce better output, or create more value for a specific client if AI is applied to it?
This forces specificity about the use case. Not "client delivery" or "knowledge management" or "operational efficiency" - an actual task. Document review in due diligence matters. First-draft generation of standard client updates. Research summarisation for bid preparation. Something you could walk over to someone's desk and watch them doing right now.
I had a conversation recently with a CTO who told me their AI priority was "improving efficiency across the practice." I asked him to name one task. Just one. He thought about it for a good thirty seconds before landing on proposal generation. That pause told me everything - not because he's bad at his job, but because nobody had ever asked him to be that specific before. The strategy had never required it. And I found myself thinking: we've probably done the same thing. Talked about AI in terms of direction rather than destination, and called it a plan.
Question two: How will we know if this has worked - what will we measure, when, and who is responsible for reporting the result?
This forces specificity about success criteria and accountability. "It's working well" isn't a measurement. "Fee earners seem to like it" isn't a measurement. "40% time reduction confirmed by comparing average review hours in Q2 vs Q3, reported by the operations lead at the October partners' meeting" - that's a measurement. It has a number, a comparison period, a named person, and a date.
Question three: What needs to happen in the next 30 days to make this pilot possible?
This is the one that separates the firms that will actually do something from the ones that won't. It kills the "we'll get to it next quarter" reflex. If you can't name the first three things that need to happen in the next month - identify the tool, get sign-off on a pilot budget, brief the team who'll test it - then you're not ready to start. And if you're not ready to start, the twelve-month timeline you've given yourself is a fiction.
I've written separately about how to build confidence in AI without overpromising - there's a companion piece on designing a governed pilot that converts these specific commitments into something your team can actually run with.
I want to be careful here, because I'm not suggesting that AI is the only thing that matters. It isn't. Relationships still win work. Reputation still opens doors. Sector expertise, quality of advice, the human judgment that clients actually pay for - none of that gets replaced by a clever bit of automation. AI is one lever among several, and anyone telling you it's the only one that matters is selling something.
But.
The managing partner who preserves vague AI ambition to avoid the accountability of specific commitments is making a trade-off. Worth being honest about what that trade-off looks like.
The upside: no risk of a named, measurable failure. Nobody has to stand up at a partners' meeting and say "we committed to a 40% reduction and we got 15%." The strategy stays comfortable. Everyone can point to exploratory activity and call it progress.
The downside: you're trading another year. Another twelve months of non-specific AI activity while competitors who have made specific commitments - and, critically, delivered against them - build a capability advantage that compounds. In a sector where AI adoption is still relatively early, the gap between firms that are three years into structured, measured implementation and firms that are still designing their strategy isn't just a gap. It's a moat.
And moats get harder to cross the longer you leave them.
I was talking to a senior partner at a consulting firm last month who put it bluntly. He said: "We spent 2023 talking about AI. We spent 2024 exploring AI. If we spend 2025 still exploring AI, we've lost three years." He's right. The frustrating part is that the firms who moved early didn't move because they had bigger budgets or better technology. They moved because someone was willing to say: "This specific thing. This team. This quarter. This is what success looks like."
Specificity and accountability. That's it. Not complicated to understand. Just uncomfortable to commit to.
If you've read this far and you're thinking right, we need to sharpen this up, the natural next step is to take whatever AI ambition currently sits in your strategy and pressure-test it against those three questions. If it survives - if you can name the task, the metric, the owner, and the 30-day kickoff plan - you have the beginning of something real.
If it doesn't survive, that's not a failure. That's useful information. You know where the gap is, and you can close it.
I've written a more detailed piece on what an AI strategy document actually needs to contain - the structure, the governance, the specifics that turn aspiration into a plan your team can execute against. If you've accepted the argument and want the template, that's the logical next read. And once the strategy is specific enough to be a real commitment, the challenge shifts to making the case for the investment it requires and governing the programme that delivers it. Different conversation, but an important one.
If you want to take the AI ambition in your firm's strategy and turn it into three specific, measurable commitments - in a two-hour session - book an AI strategy sharpening session. We've run these with managing partners and CTOs who walked in with "AI differentiation" and walked out with a named pilot, a success metric, and a 30-day action plan. It's a bit awkward, honestly. You're essentially being asked to admit, in a room with your peers, that the thing you've been calling a strategy isn't one yet. But that's the conversation. And it's the one nobody else is forcing you to have.
"We'll use AI to differentiate" isn't a strategy. It's a placeholder for the strategy you haven't written yet. The longer it sits there unchallenged, the more expensive it gets.