Anthropic launches a European consulting practice

Why this announcement wasn’t really groundbreaking… 

Technology companies launching consulting businesses is not a new phenomenon.  

If we scroll back to the 1990s, the development of ERP systems saw the emerging market leaders (SAP and Oracle) launch their own consulting firms. Their aim was to ensure that their software was properly implemented at a time when skills were scarce and to develop independent, deep, and business-led relationships with the end-users. The result was a certain amount of discomfort, as the ERP and consulting firms sorted out whether they were collaborators or competitors, but this settled down to the amicable relationships that persist today, in which everyone knows what their role is.  

This wasn’t an isolated phenomenon. The emergence of ecosystems in technology delivery made it possible to think differently about partnerships, with many firms having co-investment partnerships with multiple tech companies. 

If—and it’s a big if, which I’ll come back to in the next section—Anthropic’s move is motivated by the same things that drove SAP and Oracle into consulting, it would tell us something about how Anthropic—and potentially the other frontier AI firms—assume the market for AI services will develop. It implies that they think there will be significant demand for large-scale AI-led projects and that it will be important to the company’s future that it controls both the quality of what’s delivered and access to clients. Like systems integration, AI implementation could be a market worth many hundreds of billions of dollars, so claiming your stake now will be essential to future success. 

And this may be a dangerous assumption, for four main reasons:  

• Tech firms moving into consulting has rarely worked: Consultants have historically been snobbish about the tech-focused counterparts, a perception that was reinforced by technology’s relatively low price point in the professional services hierarchy. Snobbishness often translated into internal culture clashes, which in turn created divisions that made synergies difficult to realise.  

• Small is beautiful, at least for now: All the senior clients we’re speaking to at the moment remain convinced that AI will have a huge impact on their organisations in the future, but an increasing number are starting to believe that that, at least in the short term, AI deployments will be built data-brick by data-brick. They’re concerned that AI implementation, like transformation, will morph into the type of giant, never-ending IT projects that they dread. Instead, they’re looking for AI-enabled fixes to specific, deep-rooted, but potentially pivotal pain-points. There will be big oak trees out there in the future, but they’ll grow out of small acorns. 

• Scale takes time: There’s already a ready-made solution. The major consulting firms have spent the last decade building ecosystems that have become a staple of how they deliver services. Even if Anthropic were to scale its consulting business inorganically, that takes time (and is likely to trigger greater cultural discontinuity). The major firms are already building industrial levels of training around AI, which Anthropic and the other frontier AI firms are tapping into. 

• The commoditisation of forward-deployed engineers: FDEs are software engineers who are embedded by the tech firm they work for in client organisations with the aim of identifying and implementing AI solutions. This strategy makes sense, especially in the context of a market preference for small-scale starter projects rather than massive commitments. The work is controllable by clients (unlike large systems integration projects) and creates an opportunity to upskill the hosting organisation’s own staff. You could argue that Anthropic is essentially offering highly expert staff augmentation, albeit at eye-wateringly expensive rates. Embedding such people gives it an important opportunity to help shape the future use of AI in practice. But most FDEs won’t have the business knowledge to recognise the business opportunities for AI or access to the best practice an established consulting firm will have. They’ll depend instead on client staff to point them in the right direction and will rapidly become indistinguishable to any other type of tech contractor.  

… and why it might still be important 

However, all four reasons to question the significance of Anthropic’s decision can be undercut by one single factor: The idea that AI is a materially new type of technology. More specifically it hangs on the extent to which AI is different to other forms of information technology, not so much in terms of its inner workings but in its role in, and relationship to, business.  

Previous technology waves, from the advent of personal computing and the cloud, have promised business technology that’s easy for non-technical people to use. Of course, there will be applications of AI that are incredibly complex and which will be the preserve of engineers from companies such as Anthropic, but if non-technical people can start writing prompts that change not just how they work as individuals, but how their organisation works overall, then this will radically reconfigure the relationship between clients and consultants. Consultants have made money on the gap between technology and business, and if that disappears, it could spell trouble.