Really good article, MC. Too bad most senior partners are too 'entertained' by their internal status games to genuinely engage. But that's where the opportunity lies for the rising tide, right?
> Share the “v0.1” of a mental model or a project struggle
Right now I've come to realize the incredible potential of Adversarial Multi-Agentic AI for strategic thinking, and I'm diving very deep into the technical side of it, so I can build a POC myself to showcase the value. After all, AI amplifies everything - it can amplify good judgment that's uncomfortable, or consensual bias towards disaster! Let's see which way modern day leaders will choose...
In practice, imagine you set up multiple AI agents with distinct adversarial roles. Say, a Bull (builds the case), a Bear (attacks it), and a Synthesizer (finds what both missed). You feed them a real strategic question and let them debate across rounds, with no social friction softening the edges.
The output is a stress-tested map of where your own reasoning is strong and where it’s fragile. Think of it as a war game that requires no extensive teams, and runs in minutes instead of weeks!
It would avoid the issue of internal status games with executives, and also of external incentives with captured non-provocative consultants. Providing something a bit under what a Roger Martin or Richard Rumelt would, but with much better scalability!
Thanks for this Maurizio - probably the most personally valuable article I have read on substack. Really helped crystallise some ideas I have been fumbling with for at least a year. Looking at it through the lens of a solo consultant these days.
One thought, the big consultancies have a great pool of individuals - partners and others - to adopt this kind of model. Some will be better at it than others. In this case, the new boy - Palantir - appears to be far from this go to market strategy so perhaps this will be their weakness?
Palantir are doing pretty well, and Karp (their CEO) has even written a book recently. I also have personally interacted with their Head of Architecture on social (Chad W). I think they may be doing it in pockets.
Also, thanks for the kind words.
This is an idea I've been toying with for a year or so. I have discussed it internally as well within my own firm, and I'm glad to report I have "inspired" a number of executives to try this route. At risk of sounding "arrogant" (but that's not the intent) I believe this approach is a few years ahead of its times. In 5-10 years, as new generations of executives form, it'll be the norm.
Really good article, MC. Too bad most senior partners are too 'entertained' by their internal status games to genuinely engage. But that's where the opportunity lies for the rising tide, right?
> Share the “v0.1” of a mental model or a project struggle
Right now I've come to realize the incredible potential of Adversarial Multi-Agentic AI for strategic thinking, and I'm diving very deep into the technical side of it, so I can build a POC myself to showcase the value. After all, AI amplifies everything - it can amplify good judgment that's uncomfortable, or consensual bias towards disaster! Let's see which way modern day leaders will choose...
Tell me more about this "Adversarial Multi-Agentic AI for strategic thinking"? How would you practically use it?
In practice, imagine you set up multiple AI agents with distinct adversarial roles. Say, a Bull (builds the case), a Bear (attacks it), and a Synthesizer (finds what both missed). You feed them a real strategic question and let them debate across rounds, with no social friction softening the edges.
The output is a stress-tested map of where your own reasoning is strong and where it’s fragile. Think of it as a war game that requires no extensive teams, and runs in minutes instead of weeks!
It would avoid the issue of internal status games with executives, and also of external incentives with captured non-provocative consultants. Providing something a bit under what a Roger Martin or Richard Rumelt would, but with much better scalability!
Man, this is cool!!!! You are cooking here... I like it. Good luck with bringing this to life!
Thanks for this Maurizio - probably the most personally valuable article I have read on substack. Really helped crystallise some ideas I have been fumbling with for at least a year. Looking at it through the lens of a solo consultant these days.
One thought, the big consultancies have a great pool of individuals - partners and others - to adopt this kind of model. Some will be better at it than others. In this case, the new boy - Palantir - appears to be far from this go to market strategy so perhaps this will be their weakness?
Palantir are doing pretty well, and Karp (their CEO) has even written a book recently. I also have personally interacted with their Head of Architecture on social (Chad W). I think they may be doing it in pockets.
Also, thanks for the kind words.
This is an idea I've been toying with for a year or so. I have discussed it internally as well within my own firm, and I'm glad to report I have "inspired" a number of executives to try this route. At risk of sounding "arrogant" (but that's not the intent) I believe this approach is a few years ahead of its times. In 5-10 years, as new generations of executives form, it'll be the norm.