You Are Renting Your Intelligence. Here Is the Bill

There is a line on our own website saying that most businesses rent their intelligence through scattered tools and never notice. It is worth being concrete about what that actually means, because read quickly it sounds like a slogan when it is really an inventory.

Your business knows things. It knows why the pricing is shaped the way it is, which clients are worth chasing and which are not, what your customers object to before they buy, and what your work is supposed to sound like. Almost none of that lives in one place, and most of it lives somewhere you do not control.

Four slices, four different landlords

Sit down and map where your knowledge actually resides. It fragments along the lines of whoever sold you the software.

Where it sits	The slice of knowing it holds	Why it does not come back
The CRM	Which deals closed, at what value, against whom	It records outcomes, never the reasoning that produced them
The helpdesk	Every objection your customers have ever raised	Thousands of individual tickets, no one has ever summarised the pattern
The agency or freelancer	Why the positioning is what it is, and what was rejected	It leaves with the account team, and the next agency asks from scratch
The AI assistant	The prompts, the context and every correction your team made	It accumulates in individual chat histories nobody can search

That is four slices sitting with four different suppliers in four different formats, and none of them holds the reasoning or can see the other three.

This is not a story about bad vendors. Each of those tools is doing exactly what it was sold to do. The problem is that nobody is doing the other job, the one where somebody writes down what the business decided and why, so that the answer survives the person, the tool and the contract.

Where your knowledge actually lives

The government's UK Business Data Survey puts numbers on this, for the businesses that handle digitised data. Large (64%), medium (61%), and small (55%) businesses were more likely to use a public cloud or third-party via software or a web solution, compared to sole traders (42%).

Read that again with ownership in mind. The larger and more established the business, the more of its operating knowledge sits in somebody else's system. That is not a failure of judgement, it is what buying good software looks like in 2026. But it does mean the default position of a well-run mid-market firm is that its knowledge is distributed across suppliers, and that nobody has ever consolidated it.

It also means moving is hard, and that difficulty is structural rather than imagined. The Competition and Markets Authority spent nearly two years investigating cloud services and closed the case in July 2025 with a finding of an adverse effect on competition. The investigation identified limits to customer choice as a result of data egress fees and barriers to interoperability restricting switching and multi-cloud. When the competition regulator concludes that switching is genuinely obstructed, a business that feels stuck is reading the market correctly.

The bill arrives as re-derivation

Here is the part that makes this worth a board conversation rather than an IT one.

You are not really paying for this in licence fees. Licence fees are visible, they get negotiated annually, and somebody in finance is already on top of them. The expensive part never appears on an invoice at all, because it is paid in hours spent working something out that the business already knew.

It looks like this. A new starter asks a question that three colleagues have each answered before, and none of those answers were written down. A new agency is briefed on positioning that was settled two years ago, and settles on something slightly different, because the rationale for the original was never recorded. Somebody rebuilds a report that already exists. A proposal gets written from a blank page when four similar ones are sitting in an email archive. Teams re-make decisions they have already made, and the second version is rarely as good as the first, because the reasoning that made the first one right was lost along the way.

None of that shows up as a cost. It shows up as everything taking slightly longer than it should, forever.

The other side of the ledger is what happens when the knowledge is written down properly first. We built a weekly content engine for Excellerate Services across three regions, where producing one strong, research-backed post had taken roughly 12 hours a week and now takes around 2 hours of oversight. The gain did not come from a better writing tool. It came from the brand, the audience and the standards existing as something a system could be faithful to, instead of living in the heads of the people reviewing the drafts. If you want a figure for your own business, the capacity calculator will do the arithmetic in a couple of minutes.

What the AI layer changes

Everything above has been true for a decade. What changes it is what is being installed into those same rented tools right now.

The CMA's March 2026 announcement of an investigation into Microsoft's business software ecosystem put the timing plainly. The embedding of advanced AI, including assistants and emerging agentic technologies, into familiar workplace tools means this is a pivotal moment for the sector. The regulator's concern is competition. The concern for a founder is narrower and more immediate.

An AI assistant inside a tool you rent gets better at your business by learning from your team. Every correction they make, every bit of context they paste in to get a usable answer, every prompt they refine over six months, is your standards being written down by accident. It is genuinely valuable, and in most businesses it exists nowhere except in individual chat histories that nobody can search and nobody owns.

The leaders investing seriously in AI have worked out that this is the constraint. In IBM's study of 2,000 chief executives, 68% of surveyed CEOs identify integrated enterprise-wide data architecture as critical for cross-functional collaboration. They are not describing a database project. They are describing the absence of exactly this: one place where what the business knows is joined up, rather than four systems each holding a quarter of it.

Which is why agent pilots stall for reasons that have nothing to do with the agent. The model is fine. It just has no access to the reasoning, because the reasoning was never written down anywhere it could reach.

What owning it actually looks like

Owning your intelligence is not about hosting your own software or refusing to use the cloud. That would be a silly argument and an expensive one.

It means the layer underneath the tools belongs to you: the decisions and why you made them, the standards your work is held to, the objections you have learned to answer. That layer needs to be current and traceable back to a dated source, and it needs to sit somewhere you control, so the tools draw on it instead of each keeping a private copy. That is the difference between a system that remembers and a set of applications that each remember a bit, and it is what owning your intelligence means in practice.

The good news is that this is a smaller job than it sounds, because it is a writing job rather than a migration. The reasoning does not need extracting from four systems. Most of it was never in those systems to begin with. It needs writing down once, by the people who hold it, while they still work for you.

Practical takeaways

1. Do the inventory first. Write down which system holds which slice of what your business knows. Most leadership teams have never seen it laid out, and the list is usually the argument.
2. Separate the records from the reasoning. You can export the records from almost anything. The reasoning behind them was probably never captured at all, which is the actual gap.
3. Treat the correction history as an asset. What your team teaches an AI tool is your standards, written down by accident. Keep it somewhere you own rather than in individual chat histories.
4. Count the re-derivation, not the licence fees. The hours your team spends recovering knowledge it already had are the real bill, and nobody is currently measuring them.
5. Keep renting the software. The argument is not to stop using good tools. It is to stop letting them be the only place your reasoning lives.

The businesses that pull ahead over the next few years will not be the ones with the best stack, because the stack is available to everyone at the same price. They will be the ones who wrote down what they know while their competitors were still comparing platforms. If you are weighing up whether to do that with your own team or alongside someone, the trade-offs land differently depending on what your team already does.

Frequently asked questions

What does it mean to own your business intelligence?

It means the reasoning behind how your business works exists in a store you control, rather than only inside the tools you pay a monthly fee for. Owning the software licence is not the same thing, and neither is owning the records. You own your intelligence when someone can ask why your pricing is shaped the way it is, why you walked away from a particular client, or what you promise customers, and get a current answer with a dated document behind it, without opening a single vendor's system. Most businesses have never had that in one place, which is why the question rarely comes up until something forces it.

Is my CRM data not already mine?

The records are yours and you can almost certainly export them. What is not captured anywhere is the reasoning that produced them. A CRM holds that a deal closed at a particular figure; it does not hold why you held the price, which competitor you were against, or what the client actually needed to hear before they signed. That reasoning existed in somebody's head and in a few messages, and it was never written down as a decision. So the export gives you rows, the rows tell you what happened, and the part that would let you or an AI system make the same call again is missing.

What happens to what my AI tools learn about my business?

In most cases it stays where it was produced. The prompts your team refined over months, the corrections they made when the output was wrong, and the context they pasted in to get a decent answer all accumulate inside individual accounts and individual tools. That accumulated correction is genuinely valuable, because it is your standards written down by accident. If it only exists in a chat history or a vendor's workspace, then changing tools means starting the teaching again, and the business has no record of what it taught.

What does it cost to rent your intelligence?

The cost does not appear on any invoice, which is exactly why it runs for years unchallenged. It is paid in re-derivation: the hours spent working out something the business already knew, briefing a new supplier on positioning that was settled two years ago, rebuilding a report someone made before, or answering a new starter's question that three other people have already answered. Count the hours your team spends recovering knowledge rather than creating it. That is the bill, and in most mid-market businesses it is considerably larger than the software spend it sits alongside.

Does using cloud tools mean we do not own our intelligence?

No, and the argument is not that you should stop using them. Renting software is sensible and almost everybody does it. The mistake is letting the reasoning that makes your business distinctive accumulate only inside those tools, in whatever shape each one happens to keep it. Keep using the CRM, the helpdesk and the AI assistant. Just make sure the decisions, the standards and the corrections also exist in a store you own, current and traceable, so that swapping any one tool costs you a migration rather than a memory.