The Gap Between the Deal and the Done

There is a moment, in every business that sells the work of its people, that nobody owns.

It happens just after the celebration. The deal has closed. Someone has rung the bell, or posted the win in the channel, or simply exhaled for the first time in a month. The proposal that took six weeks to shape is now a signed contract. And in that moment, quietly, a handover begins that no system is watching.

The person who won the work knew things. They knew why the price landed where it did, which parts of the scope were firm and which were a polite fiction, which stakeholder would go quiet at the first sign of trouble, and what promise, made in a room at four o'clock on a Thursday, had actually sealed it. And now they are already gone, on to the next pursuit, because that is what the incentive structure rewards and what the pipeline demands.

The people who must now deliver the work inherit a contract. They do not inherit the knowing. So they begin, sensibly, by reconstructing. They hold meetings to work out what the client "really wants." They make assumptions about scope that the salesperson could have corrected in a sentence. They staff the engagement with whoever is free, rather than whoever was imagined. And somewhere in that reconstruction, a small distance opens between what was sold and what will be delivered. It is a few degrees off course. But a few degrees, held over the length of an engagement, is how a profitable piece of work becomes a marginal one, and how a marginal one becomes the account that everybody dreads.

Every leader of a people business has stood in this gap. Most have stood in it so many times that they no longer see it as a gap at all. They see it as the weather. Delivery always seems to cost more than the sale assumed. The best account manager always seems to be the one holding three failing engagements together with personal heroics that will not survive her going on holiday. This is simply what it is like to run a firm made of people. Is it not?

It is not. It is what it is like to run a firm made of people using tools built for firms made of products.

That distinction is the subject of this book.

The question underneath every other question

Strip away the dashboards and the quarterly rituals and the language of strategy, and every commercial leader is really asking one question, over and over, in different costumes.

Are we going to be alright?

Are we going to hit the number. Are these clients going to renew. Is this engagement going to make money or quietly lose it. Are we going to be alright.

It is a simple question and it deserves a real answer. Not a feeling. Not a forecast built on the optimism of the people whose bonuses depend on it. Not a dashboard that shows you nine things and settles none of them. A real answer, grounded in what is actually true about the business, that a leader could stake a decision on and defend to a board.

Here is the uncomfortable fact at the centre of this book. Almost no company can produce that answer. Not because their people are not clever, and not because they lack data. They cannot produce the answer because the answer does not live in the data. It lives in the decisions, and the decisions were never kept.

Think about what your company records with obsessive care. It records every transaction, down to the penny. It records every customer, every contact, every stage of every deal. It records every employee, every candidate, every timesheet, every ticket, every task, every message. If you want to know what happened, the machinery of the modern enterprise will tell you in exhausting detail.

Now ask it a different question. Ask it why you decided to take that client on despite the reservations in the review meeting. Ask it who decided to move your strongest delivery lead off the account that later imploded, and what they knew when they decided it. Ask it what you learned, as an organisation, from the engagement that lost forty points of margin, so that you would not do it again.

The machinery goes silent. Because the decision, the actual act of judgement that shaped the outcome, was never a thing the software was built to hold. It happened in a meeting, or a corridor, or one person's head, and then it evaporated. What remains is the residue: the contract that resulted, the invoice that followed, the timesheet that logged the hours. The record of what happened. Never the record of why it was chosen.

This is the strange poverty at the heart of the data-rich company. It remembers everything and understands nothing. An organisation that cannot see its own decisions cannot learn from them, improve them, or govern them, and cannot, when it matters most, prove them to anyone else.

Two words this book will make you care about

Two pieces of language, because the rest of the book depends on them.

The first is people business. I use it to mean any company whose primary product is the judgement, effort, time, and relationships of its people, rather than a manufactured good or a piece of self-serve software. Agencies of every kind. Consultancies. Professional services firms: legal, accounting, engineering, advisory. Staffing and recruitment companies. Business process outsourcers. Contact centres. Managed service providers. Any firm where, if you asked what the company actually sells, the honest answer is "our people, and the quality of what they decide and do."

People businesses are enormous in aggregate and strangely invisible in the story we tell about the economy, which is dominated by product companies and software companies. They have been handed software built for those other kinds of business and told to make do. This book is written first for them, though the argument reaches wider, because in a people business the problem I am describing is not subtle. When your product is judgement, the failure to keep your judgement is not an inconvenience. It is a hole in the bottom of the boat.

The second phrase is Decision Intelligence. I use it to mean the discipline of treating decisions as things worth keeping: capturing them, connecting them to the evidence they rested on, governing them, and learning from how they turned out. Not a product. Not a piece of software you can buy, though software can support it. A discipline, in the way that accounting is a discipline or version control is a discipline.

Decision Intelligence is not artificial intelligence, though the two are often confused. AI is a capability, a powerful new way to generate and predict. Decision Intelligence is a discipline, a way of organising how a company chooses and learns. You can have world-class AI and no Decision Intelligence at all, which is the situation most companies are quietly walking into, and it is more dangerous than it sounds. And you can practise real Decision Intelligence with nothing more exotic than rigour and a shared place to keep your reasoning.

Hold those two terms lightly for now. But you already have enough to see the shape of the argument: people businesses run on decisions, their decisions are the one thing they fail to keep, and there is a discipline whose entire purpose is to fix exactly that.

Why now, and not ten years ago

A fair objection arrives early. If keeping your decisions is so valuable, and the failure to keep them so costly, why has no one solved this before?

Two things have changed, and they have changed together, which is what makes this a moment rather than a musing.

The first is that the cost of capturing and connecting a decision has collapsed. For most of the history of business software, keeping the reasoning behind a choice, in a structured and reusable form, was simply too expensive and too laborious to be worth it. You could hold a meeting and minute it, and the minutes went into a folder nobody opened. That has changed. Cheap storage, mature data infrastructure, and machine models that can structure messy human reasoning have together made it practical to keep decisions continuously, across the organisation, rather than only in board papers, meeting notes, and individual memory.

The second thing that has changed is that the stakes have risen. The arrival of capable artificial intelligence has not made the decision problem smaller. It has made it enormous. Because the value of AI to a business turns entirely on what you feed it. An AI grounded in your real decisions and their outcomes is a genuine adviser, one that can reason from what your firm has actually learned. An AI with no access to your decisions is an eloquent stranger, generating confident, plausible answers from everything except your own hard-won experience. The companies rushing to bolt AI onto their operations without first learning to keep their decisions are building a very fast car with no memory of the road.

So the moment is this. The tools to keep decisions have finally become affordable, at exactly the point where keeping them has become essential.

The road ahead

Here is the journey, briefly.

Part I asks why the ground is moving. Three forces are reshaping what it means to run a company: trust has become the scarcest asset in commerce, machine intelligence has arrived as a co-pilot that is brilliant and unreliable in the same breath, and the old model of leadership by hierarchy and reporting is quietly failing.

Part II is the diagnosis, the part where you may find yourself nodding and wincing at once. It explains why your software estate cannot help with your most important problem, however well integrated it is, and names the reason: the modern enterprise runs on a decisionless stack, nine kinds of system that each record something while none records the decision.

Part III names the cure. It places Decision Intelligence in the honest history of enterprise software, defines it, separates it from Business Intelligence and from AI, and introduces the decision object, the small unit at the centre of the discipline.

Part IV offers a way to practise it, a philosophy developed at Optimal Nexus and built from three ideas: the people-business loop, which reframes your firm as a single cycle of selling, delivering, and learning; Business Confidence, which breaks that anxious question, are we going to be alright, into three answerable parts; and the Decision Room, which gives decisions infrastructure the way money has a ledger. It is a book of ideas, not a brochure.

Part V brings it home to the people business as a species: what these firms share, why generic software fits them so badly, and where their money quietly leaks. Part VI is the part you can act on, a way to audit your own firm and a realistic path to building a decision layer one seam at a time. It ends where the book began, with a person standing in the gap between the deal and the done, except this time they can see it, and they know what to do.

A promise and a warning

The promise is this. If you take the argument of this book seriously, you will never again look at your company's software estate, your forecast, or your Monday operating review in quite the same way. You will start to see the missing layer everywhere, the place where the decision should have been kept and was not. And once you can see it, you can begin to build it, and the building is more achievable than you fear.

The warning is this. The advantage on offer here is real but it is not free. There is no platform you can purchase that will hand you good decisions, because the discipline comes first and the tools only support it. Any vendor, including any that shares this book's philosophy, who tells you otherwise is selling you another system for the decisionless stack. What is on offer is harder and better than a purchase. It is a way of operating that compounds, so that a firm which keeps its decisions gets quietly, relentlessly better at making them, while its competitors keep forgetting and relearning the same expensive lessons forever.

That compounding is the unfair advantage. It is available to any firm willing to treat its decisions as its most valuable asset, which, in a business made of people, is exactly what they are.

Let us go and find the layer that everyone lost.