The Offerings Transform Too
Why the second wave is the one nobody is pricing
I have been having a version of the same argument for months now, about what a services business inside a software ecosystem is actually worth today. Somewhere in the middle of one of them I said something out loud that I had not written down anywhere.
We have all been talking about AI transforming how these firms deliver. Everything in the last three pieces was some version of that argument. How the work gets done, what it costs to produce, who does it, how you price it once the hours collapse.
That is the first wave, and it is the one everybody is looking at.
The offerings are about to transform too, and a good deal faster than the delivery model. Almost nobody is pricing that.
The first wave was delivery. The second is the catalogue
Here is the difference, and it is not subtle once you see it.
Wave one asks what it costs you to produce the thing. Same offering, dramatically cheaper to make, still priced as though it were expensive. That is the pricing problem, and it is solvable by anyone willing to rebuild their commercial model.
Wave two asks whether the thing still needs making. That one is not solvable by better pricing, because there is nothing left to price.
Think about what a services firm inside a large software ecosystem actually sells. Implementation. Configuration. Integration. Data migration. Custom development against the platform’s gaps. Managed services to keep all of it running.
Every one of those offerings exists for exactly one reason. The platform is hard.
Now watch what the platform vendors have been shipping. Every major one has announced capability this year that does some meaningful share of what its own partner channel bills for. Not as a threat, and not as a strategy against their partners. Simply as the obvious product move, because making the platform easier to adopt is what a platform company is supposed to do.
A partner’s catalogue is a list of things that used to be hard. When the platform makes them easy, the catalogue becomes a list of things nobody needs to buy.
That is not a margin problem. It is a revenue base evaporating from underneath a firm that is busy congratulating itself on becoming more efficient at producing it.
Roadmap concentration
The compounding factor is specialisation, and this is the part that turns a risk into a real threat.
The entire positioning of a good mid-sized services firm is depth. You are the shop that genuinely knows one ecosystem better than anybody. That depth is why clients call you, why the platform vendor sends you referrals, why your win rate is what it is. It has been the right strategy for twenty years and it built a lot of very good businesses.
It also means one company’s product roadmap can reprice your entire revenue base, and you have no vote, no warning, and no recourse.
I have started calling this roadmap concentration, because nobody seems to have a name for it and it behaves nothing like the concentration risks we all already track.
Customer concentration is on page four of every information memorandum ever written. Top five clients as a percentage of revenue, every buyer checks it, every seller has an answer ready.
Roadmap concentration is on no checklist I have ever been handed. What share of your revenue is work your platform vendor could make unnecessary inside eighteen months? I have never once been given that number, and I have never once seen a seller volunteer it.
Every deck reports customer concentration. None of them report roadmap concentration. It is the larger risk in a lot of these businesses right now.
The diversified firms hedge this almost by accident, because no single vendor decision can take out more than a slice of them. The specialists cannot hedge it at all, and the specialists are the businesses everyone has spent a decade telling to specialise further.
The obvious answer, and why it is usually the wrong one
Part two ended on the open question. Nobody has worked out how to bill for what is left once the hours go away.
I now see one answer more than all the others combined. Take the intellectual property you have accumulated over fifteen years of doing this work, turn it into a product, and sell it on a subscription. Become part software company.
The instinct is exactly right. The execution is usually wrong, and it goes wrong in a way that is very hard to reverse.
The moment you price something as software, you have not sold a thing. You have sold a promise to keep that thing working. Forever. Current with every platform release, compatible with whatever the vendor ships next quarter, supported when it breaks, on a roadmap the client now believes they have bought a say in.
Software is not a pricing model. It is a permanent maintenance obligation attached to a small recurring number.
Intellectual property you hand over is a completely different animal. Here is the accelerator. It is yours. It reflects the platform as it stands today. That prices as leverage inside a services engagement, it makes your delivery faster and your margin better, and it carries no obligation into next year.
The distinction between those two is not legal or semantic. It is the entire difference between a durable business and a slow bleed, and it comes down to one question. Who owns the obligation after the invoice is paid?
Scale decides whether you can survive getting it wrong. A very large firm can run a software business next to a services business, because it can afford two operating models, two cost structures and two completely different kinds of people. The examples that work are all enormous, and inside them the two are run as genuinely separate units.
I have never seen a small services firm pull it off. One of the two becomes a drag on the other, and it is almost always the software, because it carries the larger obligation against the smaller revenue.
You either become a software company or you become a services company that uses software very well. Below a certain size, trying to be both is how you end up doing neither.
The trap underneath the trap
It is worth being fair about why owners reach for software, because the reason is a good one and it is rarely stated out loud.
A project-based services firm rebuilds its revenue from zero every single year. You walk into January owing yourself the entire number, with a backlog that covers some of it and a sales pipeline that has to produce the rest. Do that for fifteen years and the appeal of a subscription line is not really about margin.
It is about sleep.
That instinct deserves an answer rather than a lecture, and the answer is recurring services revenue rather than recurring software revenue. Retained capability. Managed outcomes. An ongoing relationship priced on the value delivered rather than a licence that drags a support commitment behind it.
You get the predictability you actually wanted. You do not get the maintenance liability you did not know you were buying.
What this changes on my side of the table
Bring all of that to a buyer, because that is where I sit. It has changed how we underwrite Services-Tech at Bambu Capital, and in a fairly short space of time.
Two questions I did not ask two years ago. What share of this revenue is work the platform vendor could make unnecessary? And if this firm is transforming, is that showing up as a margin story or as an expense story, and how would anybody know yet?
The second one is genuinely hard, and it is the reason a lot of these deals are difficult to price honestly right now.
A firm in the middle of a real transformation looks almost identical to a firm failing at one. Both show compressed margin. Both show investment landing well ahead of return. Both have disrupted delivery and unsettled people. I made that point about the industry in part two, and it is true. On an actual transaction it stops being an observation and becomes a valuation problem, because you are paying for the first one and you might be buying the second.
There is a strong argument for waiting. Six months of evidence tells you whether the transformation converts into margin or just keeps consuming it.
There is an equally strong argument against waiting, which is that by the time it is provable, it is priced. The whole return in this category comes from backing it before the evidence arrives.
That tension does not resolve, and anybody who tells you they have solved it is selling something. You price the uncertainty honestly, or you stay out of the category.
How the acquisition builds its own competitor
There is one more thing I want to put down, because it closes a loop I opened in part two and I now think it matters more than the pricing question.
Part two argued that the founders who already sold are coming back to build again, and that this makes the current wave of small firms bigger than the last two. Here is the machinery that manufactures them. I have watched it run more than once, from both sides of the table, and the striking part is that it runs inside the transaction itself.
A founder sells a majority. The payday is meaningful. It changes his family’s position permanently without ending his career. He is comfortable and he is nowhere near retired, which is exactly the outcome every buyer says they want.
He now owns a minority of a business heading into the hardest transformation of his working life, alongside new owners he has never worked with and an incentive structure he did not design.
Then something goes wrong, because something always goes wrong in the first year. A large client leaves. In a project business a top client can be a serious share of the year, and replacing that revenue is brutal work in a market where every deal has slowed down.
Now run the arithmetic he runs. Grind through a transformation he does not control, for a minority stake, under owners he is still learning. Or wait out the non-compete, take the money that is already in his account, and build the AI-native version of his own business from scratch with no legacy delivery model, no earnouts and nobody to convince.
He knows precisely which parts of the old model were load bearing and which were theatre. He has capital. He needs permission from nobody.
The acquisition does not merely fail to prevent the next competitor. It funds him, it motivates him, and it teaches him exactly what to avoid.
Which is why I treat rollover as the first question rather than a cap table detail to be settled at the end. It is the risk model. If the founder’s outcome is not genuinely tied to yours, you have not bought a business, you have bought a headcount and started a clock.
What I would tell an owner
If you run one of these firms, five things, and the first one is the one people avoid.
Audit your catalogue against your platform’s roadmap, honestly. Not the offerings you are proud of. The offerings that bill. Anything that exists because the platform is hard has a shelf life, and you should be the one who works out how long it is.
Assume the vendor is not on your side of this. They are not against you either. They are going to make their product easier to adopt, because that is the job, and your business model is not an input to that decision.
Move toward recurring services revenue rather than software revenue. You want the predictability. You do not want the obligation.
If you do productise, be precise about which thing you are selling. A product carries a promise into next year. Handed-over IP does not. Get that wrong and you will not find out for two years, by which point it is structural.
Understand when your leverage peaks. If your value is that you have worked out how to do the new thing, that value is highest before it is provable to everyone. Certainty is what you sell, and you get paid least for it once it is common knowledge.
Where this leaves the argument
Four pieces now. How a firm governs the machine, what happens to the economics of an industry that sold hours, who is actually going to do the work, and this one.
The thread running under all of them is the same, and I did not see it clearly until I wrote the fourth.
The capability arrives well before the commercial model does. Every fortune I have watched get made in thirty years lived in that gap, and every business I have watched get destroyed died in it too. Same gap, opposite outcomes, decided almost entirely by whether the people running the firm were willing to change the model while the old one was still paying them.
Wave one repriced how the work gets done. That one is survivable, and part two and part three are about surviving it.
Wave two decides whether the work exists at all. A firm can do everything right on the first wave, rebuild its pricing, rebuild its pyramid, get genuinely good at delivering with AI, and still be finished by the second one. It will have become extremely efficient at producing something nobody needs to buy.
That is the risk I would want priced, and it is the one I almost never see on a page.
Wave one repriced the work.
Wave two decides whether there is work.
Only one of them is on the checklist.
A six part series on what AI is doing to professional services.
One. AI Decides How. Humans Decide What, Why, and If. Governing AI inside a single firm, and where the human veto has to sit.
Two. Nobody Is Buying Hours Anymore. What happens to an industry that priced and sold the one thing the machine turned out to be best at.
Three. Guess Who Becomes AI Native. Whether this wipes out a generation of young professionals, and what becomes of the pyramid.
Four. The Offerings Transform Too. You are here. The second wave, where the platform makes your catalogue unnecessary, and why productising is the wrong answer.
Five. Services-Tech. Naming the category, and the line between the services businesses that get rebuilt and the ones that do not.
Six. Nobody Is Ripping Anything Out. The AI layer over the systems you already own, the migration nobody has named, and why it is the first good news in the series.