A feature on cutting through the hype
The question isn't whether. It's where.
Every business is being told the same thing right now: use AI, urgently, or fall behind. It is vague, anxious, expensive advice — and it leads straight to panic-bought tools that solve no real problem. The useful question is not whether to use AI but precisely where: which handful of tasks it transforms, and which it merely makes more complicated and costly. AI strategy is the clear-eyed work of finding the few genuinely high-value uses, ignoring the shiny distractions, and building a plan that pays back — hype removed, ROI restored.
A strategist at a board mapping AI use cases, each scored on a value-versus-effort grid with a few clear winners highlighted. The clear-eyed prioritisation behind a real AI plan. Square aspect ratio, dark studio, focused light, solar-yellow glow, a sense of rigour and judgement.
Strategy as prioritisation — the few AI use cases that genuinely pay, found and ranked, the distractions set aside.
There is a particular pressure in the air, and every leadership team feels it: adopt AI now, or be left behind. The pressure is real, but the panic it produces is expensive. It leads businesses to buy tools because they are new rather than because they are needed, to bolt a chatbot onto a problem that didn't have one, and to mistake activity for progress. The companies that win with AI are not the ones that adopt the most of it — they are the ones that adopt the right of it. Knowing the difference is the entire job of strategy.
The truth that the hype obscures is that AI is genuinely transformative for some tasks and a costly distraction for others, and the entire value of strategy is telling them apart for your specific business. It excels at things like generating and summarising language at scale, spotting patterns in large data, automating repetitive judgement-light work, and powering natural conversation. It is poor, or actively dangerous, where accuracy is critical and unverifiable, where the data isn't there, or where a human relationship is the actual product. A good strategy is mostly a map of which is which.
Underneath the use-case question sits a less glamorous one that matters more: readiness. AI runs on data and process, and most of the failures we see have nothing to do with the model and everything to do with what surrounds it — scattered or poor-quality data, processes too vague to automate, no clear owner, no way to measure whether it worked. The smartest model in the world cannot fix a problem the organisation hasn't defined. So real strategy spends as much time on data, process and ownership as on the technology itself, because that is where AI projects actually succeed or quietly fail.
In this feature
Six things a serious AI strategy settles.
Use cases, ranked
We identify where AI genuinely creates value in your business and rank those uses by return against effort — so investment goes to the few opportunities that pay, not to whatever the loudest vendor demonstrated last.
Avoiding the distractions
Half the value of strategy is the no. We name the expensive, fashionable AI projects that won't pay back for you and steer the budget away from them — because not doing the wrong thing is as valuable as doing the right one.
Build vs buy
Most AI needs are met by configuring tools that already exist; a few justify building something custom. We make that call honestly for each use case, so you neither reinvent a commodity nor force a bespoke problem into an off-the-shelf box.
Data & readiness
AI runs on data and process, and most failures live there, not in the model. We assess what you actually have — data quality, processes, ownership — so a plan is built on what's real, not on a demo that assumed perfect inputs.
The human in the loop
AI should augment judgement, not replace accountability. We design where a person reviews, approves or overrides — so you capture the speed without inheriting the risk of an unsupervised system making confident mistakes at scale.
A roadmap, not a moonshot
We sequence adoption as a series of small, provable wins rather than one giant bet — start with a contained pilot, measure it honestly, and scale only what works. Momentum and evidence, not a single all-or-nothing leap.
AI everywhere is a budget.
AI somewhere is a plan.
The defining mistake of this moment is treating "adopt AI" as a goal in itself. It isn't a goal; it's a means, and a means that is spectacular for some ends and useless for others. A business that sets out to "use AI" will always find somewhere to spend the money — but spending is not the same as winning, and a flashy pilot that impresses the board can quietly deliver nothing while a dull, narrow application transforms a single expensive process. Strategy is the discipline of spending on the second kind and skipping the first, however much less exciting that is to announce.
We start every engagement the same way: not with the technology, but with the business. Where is real time being spent? Which processes are repetitive, language-heavy or pattern-heavy? Where do bottlenecks form, and what would it be worth to clear them? Out of that comes a long list of candidate uses, which we then score honestly on two axes — the value if it works, and the effort and risk to get there. The output is not "here are forty things AI could do," which is useless; it is "here are the three that will pay, in this order, and here is why the rest can wait or never happen."
The build-versus-buy question follows close behind, and getting it wrong is expensive in both directions. The overwhelming majority of business AI needs are now met by configuring mature, off-the-shelf tools and models — building your own would be like writing your own spreadsheet software. A genuine few, where the use is core to your advantage and the available tools don't fit, justify something custom. We make that call use case by use case, because the instinct to build everything wastes money and the instinct to buy everything occasionally surrenders the one capability that would actually set you apart.
The serious version of AI strategy is sequenced and provable, never a single grand bet. We turn the shortlist into a roadmap of contained pilots, each small enough to ship quickly and measure honestly, each with a clear owner and a clear definition of success. What works is scaled; what doesn't is killed early and cheaply. This is deliberately the opposite of the all-or-nothing "transformation programme" that consumes a year and a fortune before anyone discovers whether the core idea was sound. Momentum, evidence and the freedom to stop are the whole point.
An AI adoption roadmap on screen — a value-versus-effort grid with a few clear winners, sequenced into phased pilots. The plan that turns AI hype into prioritised, provable steps. Wide cinematic 21:9 crop, dark elegant studio backdrop, solar-yellow accent, a sense of clarity and judgement.
A ranked, sequenced plan — the few high-value AI uses, in order, with the distractions named and set aside.
Five questions we ask before recommending any AI.
The companies that win with AI don't adopt the most of it. They adopt the right of it — and knowing the difference is the entire job of strategy.
AI strategy is the front door to the whole Artificial Intelligence pillar. It is where we decide, with you, which of the pillar's capabilities actually fit your business — whether the win is in AI-generated content, conversational interfaces, automation, custom models or personalisation — and in what order to pursue them. Rather than selling you a fixed product, it points the rest of the pillar at the few places it will genuinely pay. Strategy decides the where and the why; the other disciplines deliver the how.
We advise on AI the way we'd want to be advised — honestly, with as much attention to what you should not do as to what you should, and with every recommendation tied to a return rather than to a trend. The brief is a plan that pays back, not a portfolio of impressive pilots. That combination of ruthless prioritisation, honest readiness assessment and provable, sequenced delivery is exactly why strategy comes first: it is the difference between using AI and merely spending on it.
A whiteboard of AI ideas being narrowed to a shortlist — most crossed out, three circled and sequenced. The moment a hype-driven wishlist becomes a focused plan. Contemporary, shallow depth of field, dark desk, solar-yellow marker glow.
From a long wishlist to three real bets — the distractions cut, the budget aimed where it pays.
Representative scenario · not a named client engagement
A firm was about to spend big on the AI project everyone was excited about — until a strategy found where the money actually was.
The leadership team felt the pressure to "do something with AI" and had landed on a flagship idea — a customer-facing AI assistant that demoed beautifully and excited the board. It would also have cost a great deal and taken many months, against a benefit nobody had actually quantified. Before committing, they asked us to pressure-test the plan and make sure they weren't missing anything. The honest answer was uncomfortable: the exciting project was high-effort and low-return for them, and the real opportunity was somewhere far less glamorous.
We ran the full prioritisation: mapped where time and money were really being spent, generated a candidate list of AI uses, and scored each on value against effort and readiness. The flashy assistant fell down the list. Rising to the top was a dull but expensive bottleneck — a manual, repetitive document-handling process — that AI could automate cleanly, plus two smaller language-heavy tasks. We delivered a sequenced roadmap: a contained pilot on the document process first, with clear success measures, before anything else.
The exciting project was shelved; an unglamorous automation nobody had championed became the win that actually paid.
The pilot paid back quickly and visibly, which did something a grand programme never could: it built real confidence to go further, on evidence rather than hope. The team avoided sinking a large budget into an impressive project that would have changed little, and instead banked a fast, measurable win and a clear-eyed plan for the next. The strategy's most valuable output was the expensive thing it talked them out of — which is exactly what good advice is for, and exactly what a vendor selling a product can never give you.
We were ready to pour a budget into the AI project everyone loved. Revolutionize showed us, with actual analysis, that it wouldn't pay — and found the unglamorous one that did. They talked us out of the expensive mistake and into the win. You don't get that from someone selling you a platform.
What a serious AI strategy engagement actually involves.
AI strategy is an investment in spending the rest of your AI budget well, and it is scoped as advisory work rather than a build. Engagements range from a focused opportunity assessment and prioritised roadmap to an ongoing advisory relationship that guides adoption as it scales, and the scope rises with the size of the organisation, the number of processes in play, and the depth of the readiness and data assessment required.
The largest variable is the breadth of the business in scope. Assessing AI opportunities across a single team or function is a contained piece of work; mapping them across a whole organisation, with a proper readiness and data audit, is a larger one. The strategy is always costed separately from any delivery that follows, precisely so our advice on what to build — or not build — stays honest and independent of who builds it.
Every engagement includes the full discipline: opportunity mapping against the business, value-versus-effort prioritisation, a readiness and data assessment, build-versus-buy recommendations, and a sequenced, pilot-led roadmap with clear success measures — delivered as honest advice you own, whoever ends up implementing it.
We scope every engagement to your business rather than to a price list — which is why we don't publish rate cards. Every one begins with a free 30-minute scoping conversation, and we will tell you honestly whether you need a full strategy or just a steer on one decision — and, often, that the most valuable thing we can do is talk you out of an expensive project. We would rather give you the right answer than the biggest invoice.
When you're ready
Stop buying AI. Start using it.
Tell us what you're being pushed to "do with AI," or the project you're not sure about. We'll respond within 24 hours with an honest first read on whether it's likely to pay — and where, in your specific business, AI would genuinely earn its keep instead of just its headlines.
Begin the conversation →