Human-assisted consulting agents (2/n)

The business model of service-based activities (including consulting) relies, in my view, on three elements: information asymmetry, competence, and time. In that order.

Take the example of hiring an external firm for data analysis (completely randomly👀). Consultants know that there are tools and methods that can solve a given problem : that’s the information asymmetry.

Even if the client were aware of these tools and methods, do they actually have the internal capability to use them? (An SME/mid-sized company typically doesn’t have an army of data scientists combined with management consultants reviewing COMEX issues and trying to solve them.) That’s the competence.

And even if the company had both the knowledge and the capability in-house, it doesn’t necessarily have the time to dedicate to it (because employees are usually caught up in day-to-day tasks and firefighting). That’s the time.

So despite all the apocalyptic narratives about the end of consulting (and the end of everything, really) because of AI, I remain rather optimistic about our industry. It’s true that AI will kill consulting (at least as we know it), but that’s not entirely accurate, nor is it imminent.

It’s true in the sense that GenAI gives everyone the impression that they now have access to both information and skills, quickly: anyone can prompt an LLM and get something that looks reasonably convincing. But the gap between “reasonably convincing” and genuinely impactful is not that small.

As soon as you want to actually use AI outputs (and provided you have a minimum level of professional standards) you need to spend time challenging the result, adapting it, iterating. And this is where time and competence come back into play.

Consulting will indeed be displaced by AI, but only in companies that manage to give their employees enough time to think and to experiment on non-productive tasks. And those are precisely the companies that, with or without AI, already make very limited use of external consulting anyway.