OpenAI's $4bn company: what it means for UK small businesses

OpenAI just launched a $4 billion company staffed by what they call "Forward Deployed Engineers" - humans embedded into businesses to find AI workflow gaps and rebuild around them. HoursBack applies the same diagnostic principle to UK small businesses. For £799.

I want to be careful about what I am claiming here. I am not saying OpenAI copied the HoursBack model. I am not predicting where AI is going. What I am saying is narrower and more specific: the company that built ChatGPT, backed by McKinsey, Bain and Company, TPG, Bain Capital, and Brookfield, raised $4 billion at a valuation of $10 billion on 11 May 2026 - and they chose to build a company staffed by human experts who go into businesses. Not a SaaS platform. Not a chatbot. Not a training course. That is a choice worth understanding.

What the OpenAI Deployment Company actually is

There is no app. There is no subscription to log into. The product is a person.

The OpenAI Deployment Company is built, in part, on their acquisition of Tomoro - a UK-based applied AI consulting firm, hired specifically for its Forward Deployed Engineer talent. These are people who know how to walk into an organisation, identify where the AI value is hiding, and rebuild the workflows around it. The model is not new - large consultancies have operated this way for decades. What is new is applying it specifically to AI deployment.

The reason this model exists is that the obvious alternative was not working. Three years into the age of widely available AI tools, most businesses - even the large ones - were not capturing the value. They had access to the tools. They were not getting results. The gap was not access. It was deployment.

The OpenAI Deployment Company is a $4 billion bet that the missing variable is human expertise - someone who knows what to look for, walks in, looks for it, and builds the fix.

This is a validation signal, not a prediction

I am not going to tell you where AI is heading. I do not do that, and I do not think anyone does it reliably. But I can describe what just happened.

The most prominent AI company in the world - with more access to AI capabilities than any external organisation - looked at the question of how to actually get value out of those capabilities inside real businesses, and concluded that the answer is a human expert running a structured diagnostic. Not a smarter tool. Not a better prompt. A diagnosis first, by a person, specific to your business.

There is a version of this announcement that would have proved me wrong. If the OpenAI Deployment Company had launched a new AI platform and said "the tools are now so capable that businesses just need better access", I would have updated my view. They did not say that. They said the model is a Forward Deployed Engineer. A human. In your business. Diagnosing before deploying.

That is a data point. Not a boast, a data point.

They spent $4 billion on it. I offer it for £799. I think the model works for anyone outside of enterprise-scale organisations.

The gap this creates

The OpenAI Deployment Company is working with global corporations. Enterprise clients. The businesses that can spend McKinsey money on McKinsey-scale deployment.

If you run a coaching practice in Manchester or an estate agency in Kent, there is no version of this programme available to you. The Forward Deployed Engineer model is not coming to your inbox at enterprise pricing. It is not designed for the plumber in Birmingham or the wedding planner in Bath. It is designed for companies that have entire IT departments and still cannot figure out where the AI value is.

Which means there is a gap: the same diagnostic process that enterprises are now paying a premium for - someone who knows what they are looking for, walks in, and finds it - is simply not available to most UK small businesses.

Many of them are buying tools instead. That is the mistake.

The most common AI mistake I see

I have sat with enough UK small business owners to recognise the pattern. They are subscribed to two to four AI tools (or none whatsoever). One they use regularly. One they opened twice. Two they forgot they were paying for. They spent money. They saved no time.

The cause is almost always the same: they bought before they diagnosed. They picked tools because a supplier recommended them, or because they saw them in a tutorial, or because everyone seemed to be using them. They skipped the step where someone sits down and asks: what is actually eating your time, in what order, and what is the most efficient way to reclaim it?

Running AI tools without the diagnostic is the equivalent of building a data centre without a deployment engineer. You have the hardware. You have not mapped the cables.

Most AI problems I see are not tool problems. They are sequence problems. Automate the wrong workflow first and you have just made the wrong workflow faster. The diagnosis tells you which workflow to fix first - and sometimes whether to automate it at all. Sometimes the right answer is "delete this step entirely." No AI tool will tell you that. A person who has spent an hour in your business will.

What HoursBack is

I built HoursBack to do, for UK small businesses, what the OpenAI Deployment Company is charging enterprise rates to do for global corporations.

The model is the same: a human expert, a structured diagnostic, a specific plan. The scope is different - a 60-minute session rather than a six-month embed. The price is £799. The output is a personalised report delivered within 48 hours, with the specific workflows identified, the specific tools recommended, and a 5-day implementation plan written in plain English. No jargon. No 10-hour course. No new subscription you buy and forget about.

The underlying architecture - diagnosis before deployment, human expertise over tool access - is built on the same principle as the model OpenAI just spent $4 billion on.

If you run a UK small business and want that same diagnostic, the Assessment is at hoursback.co.uk/assessment. £799. Or if you want a quick sense of where you stand first, the free AI Readiness Quiz takes two minutes.