AI: Application matters as much as technology
Seventeen percent of clients would actually prefer to deal with an Agentic AI than a consultant, our research tells us. It’s the kind of statistic that sends shivers down the back of the professional services industry. But, while AI will undoubtedly change the face of consulting, the way it’s applied will create more value than the technology itself.
Eric Abrahamson is a professor at Columbia University. Although he’s now mostly writing about AI and interactive chatbots, some of his earlier research around the diffusion of innovation may also be useful now. Back in the 1990s, Abrahamson looked at management bandwagons; fast-forward to today, we’ve been looking at what his ideas might tell us about the adoption of AI by client organisations.
Management bandwagons before AI
Traditional consulting begins with a new approach—an idea, methodology, or technology that client organisations don’t have. Some new approaches simply don’t work, of course. But those clients that take a leap of faith and become early adopters of solutions that can deliver potential benefits will have an advantage over their slower-moving competitors—and this will ultimately create economic value for them. However, that value comes at a price. Because the approach is new, the cost of implementation is high: It has to be invented from scratch and depends on an organisation’s unique circumstances; mistakes are made; work needs to be re-done. Because it’s new, it’s also high-risk: Theoretical benefits may not be realised in practice.
This picture changes over time. As the small number of early adopters trumpet their success, other organisations start to look at how they can benefit from this now-not-quite-so-new approach. These latest adopters replicate the approach of those that adopted early on rather than developing their own, because they’re in a hurry and copying is faster and lower risk than reinvention. But, because they spend less time thinking about how the approach should work in their organisation, the value delivered is lower. More time passes: Costs fall, and standardisation increases. The value delivered declines, slowly at first and then more precipitously as organisations copy each other, rather than going back to the first principles.
Consultants play a critical role in this process. They may see the value in working with the small number of early adopters but the cost of delivery for them, too, can be astronomical and the risk of reputational damage is high. It’s better for them to standardise and de-risk what they do before they take this approach to other organisations. Indeed, most consulting work takes place in the subsequent waves of adoption. Consultants don’t generally innovate, they codify. And that’s not a bad thing: It introduces new approaches to swathes of organisations that aren’t prepared to invest at the leading edge of innovation but can still benefit from it. The irony—indeed, the multi-billion-dollar problem—is that codified approaches create less value for client organisations precisely because they’ve not been customised to their specific needs. In the death throes of a management approach offering, it’s not possible to standardise further but the costs are rising, and value is negligible. Clients stop buying it and consulting firms stop selling it.
The changing relationship between value, customisation and cost in traditional consulting
AI has to be implemented, not standardised
Forty-five percent of clients say that AI is one of the areas of technology they’ll be investing in the most over the next 12 months, and 99% of them expect to use consulting support to do this. All this is driven by the hope that AI will cut costs, enable organisations to respond faster, and help their employees work smarter. Only 9% of clients would currently describe themselves leading the way in AI adoption, while two thirds would put themselves in the hesitant or wait-and-see category.
AI promises instant gratification—supply chain consulting out of a box! Robo-strategists!—all of which is very attractive in an environment where organisations need more external support than ever before but struggle to afford it. The problem this creates is that it risks putting AI-delivered services towards the end of the traditional bandwagon trajectory: Standardised solutions that don’t cost much but don’t add much value either.
To avoid that scenario we have to go all the way back to the beginning.
Clients are buying an outcome, an end, not a means. For an outcome to deliver value to the wider organisation, it has to be tailored to the very specific needs of that organisation, the kind of needs that can only be uncovered, articulated, and championed by people—and typically by people with deep experience in the business issues being impacted. That 17% of clients would prefer to deal with an Agentic AI instead of a human being represents a failure on the side of clients to recognise that’s the case—and suggests that consulting firms are not sufficiently making the case for using people’s knowledge to customise the technology. Not thinking about the business issues and rushing too quickly to create highly standardised solutions risks destroying the potential to create value for both clients and consultants.
What can firms do next?
Clients are not looking for a blanket approach, but for solutions that are tailored to their specific needs and goals. To find out more about clients’ attitudes to AI and their areas of concern, take a look at our latest Emerging Trends report—it’s full of forward-looking insight to allow you to anticipate change and uncover new growth opportunities in a rapidly developing field. If you would like to speak to one of our experts for insights that are more tailored to your firm, get in touch.