Serendipitously, I had two meetings back to back last week that felt bizarre together - like they were each from different centuries.
The first was with an SE leader at a consultancy that's known for bespoke enterprise work. Our conversation was about repeatable downmarket offerings - they're feeling the market shift, and they're shifting strategy accordingly.
The second was with a consultant on the proserv team at a SaaS company. His client - a billion-dollar enterprise - was so wary of AI that they wouldn't allow recorded meetings, never mind AI notetakers. So he and his team were doing all of it by hand - just like the old days, slowly and expensively.
On its face, these seem like opposite reactions: proactively using AI for everything and making things cheaper, vs building a wall to hide from technology and making things more expensive.
But because both strategies are focused on discovery, they are actually not that different. Discovery is where project risk lives. Done richly, it means fewer surprises and more alignment; done lightly, it leaves potential landmines. It's also sensitive - it's where the client's real data and real problems get picked open.
This seems obvious, but to say it out loud: the second org is not making anything safer by banning AI. Human consultants are still reading every document, digging into the nooks and crannies of their workflows and business processes and data. It's exposure they're used to, but it's not safer. And - it takes much longer, is much more expensive, and is not as thorough.
On the other hand, the SI trying to make discovery cheaper is getting multiple benefits. The properties that make discovery automatable - structure, consistency, auditability - are the same ones that make it rich. AI-supported discovery is faster, cheaper, and more thorough at the same time.
This isn't true for every part of every project. But for discovery, cheaper, safer, and better all go hand in hand.