The Last Word: AI’s next battle is not the model, but the organisation. From ‘What can the model do?’ to ‘What can the organisation absorb?’
For the past few years, attention has centred on the models: who had the smartest system, the fastest capability gains and the most impressive demonstrations. But the more important battle is now moving elsewhere—from model creation to enterprise transformation.
This is why Anthropic’s new enterprise AI services company, announced with Blackstone, Hellman & Friedman and Goldman Sachs, is more than a corporate partnership. It is a signal. The new firm is intended to help companies bring Claude into core operations, with Anthropic applied AI engineers working alongside its team to build custom solutions and support customers over time. At almost the same time, OpenAI was reported to be building its own private equity-backed deployment venture to help businesses use its AI software at scale.
Taken together, these moves suggest that the industry has discovered an awkward truth. Enterprise customers do not buy intelligence in the abstract. They buy reduced cost, faster cycle times, better decisions, lower risk and new capacity. They buy outcomes. A magnificent model that sits outside the organisation is only a very clever ornament.
The private equity angle gives this shift its commercial muscle. Blackstone’s release points to opportunities across healthcare, manufacturing, financial services, retail, real estate and infrastructure—precisely the kind of sectors where operational gains can be turned into enterprise value. If AI can lift productivity in one portfolio company, the playbook can be adapted across dozens more. A modest operational improvement can become a valuation story. A repeatable improvement across a portfolio can become an investment thesis.
This is not simply software distribution. It is the industrialisation of implementation. The frontier is moving from What can the model do? to What can the organisation absorb? That is a far more uncomfortable question, because the obstacles are rarely technical alone. They live in incentives, habits, governance, data quality, leadership alignment and the quiet resistance of people who have seen many transformation programmes arrive with a logo and leave with a dashboard.
This is where the new entrants become dangerous to the established consulting order. For decades, Deloitte, PwC, EY and KPMG—alongside Accenture, Infosys and Capgemini—have occupied the space between boardroom ambition and operational reality. They speak fluent process, compliance, migration and change. But the AI labs bring privileged proximity to the technology, engineering talent close to the models and investors hungry for measurable gains.
Delicious irony
That does not mean the Big Four should start clearing their desks. It does mean the question has changed. They are no longer competing only with each other, or with specialist boutiques promising agile transformation and better slideware. They may soon be competing with AI-native services firms that arrive with capital, model access and a direct line into the owner’s value-creation plan.
The irony is delicious, if slightly cruel. The consultancies have spent years advising companies to anticipate disruption. Now a sharper version may be landing on their own doorstep. Did the Big Four anticipate that the next challenger would not be another partnership with better methodology, but a model company with private equity at its elbow?
Still, this new order is not guaranteed to work. Technology companies are brilliant at building products. Transformation is messier. It asks whether leadership teams can redesign work, rewrite incentives, clean data, retire habits and hold their nerve when productivity gains become political. Embedding AI is not installing software. It is remaking the operating system of the enterprise while the enterprise is still running.
So the value question has moved. In the first phase of AI, power accrued to those who built the intelligence. In the next, it will flow to those who can make intelligence useful, trusted and repeatable inside real organisations. The winners may not have the best models, but the best reinvention discipline. And the Big Four should be asking whether they saw this coming.
Photo: Dreamstime.