The world will spend close to $2.5 trillion on digital transformation this year, on its way to roughly $4 trillion by 2027, according to IDC. I run a company that sells transformation work, so a number that large should make me happy. It does not, because I know what most of that money buys, and a lot of it buys the wrong half of the problem.
The product the model was actually selling
For most of the last two decades, digital transformation consulting sold one thing above all others, and that thing was a strategy. A deck, a target operating model, a roadmap with workstreams and a steering committee to watch over them. That product made sense in a world where the genuinely hard part was knowing what to do, because when the destination is unclear, paying a firm to map it is a rational purchase and the good firms earned their fee. I have sat on both sides of that table, as the buyer earlier in my career and now as the person whose company gets hired, and the shape of the engagement rarely changed. The firm produced the thinking, and you produced the results later, alone, long after the invoice had cleared.
That division of labour is the part nobody said out loud. The report was the deliverable, and execution was quietly reclassified as the client’s problem rather than the consultant’s. It worked as a business model because the strategy was the scarce, expensive, defensible thing, and everyone agreed to treat the year of grinding implementation that followed as a separate matter. For a long time that agreement held, because building the plan really was where most of the difficulty and most of the value sat.
What the market actually shows
Here is where I part company with my own industry, though not in the way you might expect. The number everyone quotes is that 70 percent of digital transformations fail, and the number is roughly right while the word is wrong. BCG’s research finds that only about 30 percent of transformations fully succeed, meaning they land their targets on time, on budget and in scope. Falling short of that bar is not the same as failing outright, but it does mean most of that multi-trillion-dollar spend underdelivers against what was promised. The part worth arguing about is the cause, not the size of the number.
When you read what actually goes wrong in these programmes, it is almost never the strategy. The deck was usually fine. What broke was everything after it, the governance, the change fatigue, the operating model on the slide that assumed a workforce with spare capacity nobody had. Knowing what to do was never the binding constraint, and the old model had sold its whole value against solving the part that was not actually the problem.
I have watched a mid-market operator go through this the expensive way. They had paid a name-brand firm for a transformation strategy, a genuinely competent document, and eighteen months later almost none of it had moved. The plan itself was fine; every initiative in it assumed the operations team had spare time it did not have, so that team spent eighteen months doing what it had always done, by hand, because nothing in the strategy touched the manual work underneath. The document sat in a shared drive, quietly correct, while the business kept paying for the gap between a good plan and an operation that could not absorb it.
What AI did to the value of a strategy deck
Then the ground moved under all of it. McKinsey’s 2024 survey put regular generative-AI use at 65 percent of organisations, nearly double the figure from ten months earlier. I am not interested in that adoption number for its own sake. I care about exactly one consequence of it: what it did to the price of the thing my industry charged the most for.
A competent transformation roadmap can now be drafted in an afternoon by a tool that several of your managers already have open. The frameworks, the maturity models, the phased plans, the benchmarked target operating models, all of the artifacts that used to justify a six-figure strategy engagement are now the cheapest part of the whole endeavour. Strategy has been commoditised, and I say that as someone whose industry was built on selling it. The scarce thing now is not knowing what to do but getting it done inside an operation that is already running flat out, where the people who would have to execute the plan are the same people spending their days holding the current process together by hand.
That last point is the one the old model has no answer for. We named the pattern the Manual Wall: the growth ceiling a business hits when the next increment of output needs a disproportionate increment of headcount, because the work is still stitched together by people rather than systems. A strategy deck does not touch that wall. It names a destination and hands the wall back to you, still standing, now with a roadmap taped to it. AI made the deck cheap and left the wall exactly where it was, which is why the value has moved decisively from the planning to the doing.
What got cheap, and what did not
It is worth being precise about what AI actually commoditised, because the distinction is the whole argument. What got cheap is the analytical layer: the market scan, the maturity assessment, the benchmarked operating model, the phased roadmap, the business case with its tidy net-present value. Those were the artifacts a strategy engagement produced, and a capable manager with a good model can now generate a defensible version of each in a fraction of the time and cost. That is genuine progress, and pretending otherwise is how firms end up charging for work the client could now do themselves in an afternoon.
What did not get cheap is anything that requires contact with your actual operation. AI cannot walk your claims desk and notice that three people exist mainly to move data between two systems nobody ever connected. It cannot absorb change on behalf of a team already at capacity, carry the political weight of retiring a process someone built their role around, or stay in the building for the eight months it takes a new workflow to become simply how things are done. The execution layer is stubbornly, expensively human, and it is exactly the layer the old model handed back to the client as their problem.
What replaced it, and why we built the company this way
I did not arrive at this view from the outside. Digital Forms started around 2015 as a development shop, Mariusz Graczkowski and I in a small office in Warsaw, doing good work for reasonable money and slowly realising we were answering the wrong question. Clients kept asking whether we could build a particular thing, when the question we actually wanted to answer was what their business needed built at all. In 2019 a friend showed me a digital transformation deck from his old employer that argued entirely in P&L terms, what was broken, what it cost, how to fix it, and something clicked, because that was the conversation I had been trying to have for years. Turning the company toward it took four hard years, in which we learned to sell outcomes instead of deliverables and to price the return before writing a line of code, and doors closed because I could not yet explain the value clearly enough. The company that came out the other side is the one I am describing here: we went from ten people to seventy, and the friend with the deck, Przemysław Wójcik, is now my co-owner.
So we built the company deliberately against the model I have just described. We do not sell a strategy you then execute alone. We start by putting a number on the manual work, which is what our Profit Leak Diagnostic does in about four weeks, because you cannot sequence what you have never measured and most operations have never measured it. Then we put a single change into production, a first Operations Sprint that gets a working result live in weeks rather than quarters, so the saving funds the next move instead of a budget meeting. The fuller plan, the Growth Readiness Roadmap, comes once there is a result on the board rather than before, because a roadmap written against a proven number is a different document from one written against a hope.
Every initiative is sequenced in ROI order, biggest recoverable cost first, and nobody hands you a slide and wishes you luck. Where a business needs one person accountable for the operational outcome across the whole arc, that is what our External CDO as a Service is, and it is the opposite of a report, because it is an owner who stays in the operation until the change is real. None of this rejects strategy; it recognises that strategy is now the easy, cheap part, and that the honest version of digital transformation strategy is mostly a sequencing and execution discipline wearing a strategist’s coat.
The shift this forces on a firm like mine is uncomfortable and correct. It means we get paid for outcomes we can be measured on rather than for the quality of our thinking, and it means the work is harder, slower, and closer to the client’s actual mess than a strategy engagement ever was. It also means we cannot hide behind a deck when a project stalls, because we are the ones in the room when it does. I think that is the trade the market is moving toward, and the firms resisting it are defending a margin structure rather than a value proposition. The most defensible thing a transformation partner can now offer is not intelligence, which has been commoditised, but accountability, which cannot be.
The reason this matters beyond my own company is that AI is going to make the imbalance worse, not better. As the strategy layer keeps getting cheaper and faster, the operations that win will not be the ones with the best plan, because everyone will have access to a perfectly good plan. They will be the ones that can actually absorb change without drowning the people who have to deliver it, and that capacity is built, not bought in a deck.
What a mid-market CEO should actually buy now
If you are running a mid-market business and a firm arrives offering you a digital transformation strategy, the first question worth asking is who executes it, and what happens to the plan the day after they leave. If the answer is that execution is your job and their deliverable is the thinking, you are buying the half of the problem that AI has already made cheap, at a price set when it was expensive. That is not a reason to buy nothing, but a reason to buy differently.
Buy the execution capacity, measured. Ask for a number on your manual work before you commit to a programme, insist that the first result lands in weeks so you can see the mechanism work on your own operation, and refuse to pay for a roadmap that has no one attached to delivering it. The firms worth hiring now are the ones willing to be measured on whether the change actually happened, not on whether the strategy document was comprehensive. Ask the awkward commercial question directly too: what portion of the fee is contingent on the change landing? A firm confident in its execution will have an answer ready. A firm selling you a plan will change the subject to the sophistication of its methodology, which tells you what you needed to know.
The part I am certain about
I do not think digital transformation consulting is dead, and I would be a fool to write its obituary while running a firm that does the work. What is dead is the specific bargain at the centre of the old model, the one where you paid for the thinking and inherited the doing. AI killed the economics of that bargain by making the thinking cheap, and it exposed what was always true underneath: the plan was never the hard part. If your next transformation goes the way most of them have gone, the reason will not be that nobody knew what to do, but that knowing was the only thing anyone was ever really paid to deliver. The doing was always yours, and what has changed is that you no longer have to accept that split, because the market now has partners who will take it on with you and be measured on it. The strategy was never going to save you; the execution always was, and now you can finally buy the two as one thing. That is the version of this work worth paying for, and it is the only version I am interested in building.
