• by Dave Morgan , Featured Contributor, Yesterday

I shared a fun dinner Tuesday night with longtime friends and legendary tech CEOs Matt Blumberg, Chad Dickerson and Mark Josephson. We had a particular mind-stretching moment when Matt shared a concept he'd just heard at a recent Gartner event: that AI is evolving so fast right now that most agentic AI projects are likely to be obsolete before they have a chance to reach maturity.

That is a very powerful concept to wrap your head around -- and the more I think about it, the more I agree with it. Here’s why:

Failure of corporate imagination. It’s certain that AI models will continue changing faster than most agentic project plans. They will be on a treadmill running faster than they can run. The most likely solve will be highly imaginative project scopes that will carry a lot of uncertainty and engender a lot of risk. But most large corporations do not possess deep imagination, nor do they encourage risk.

Too much focus on buzz rather than business. Let’s face it, a lot of agentic projects today are launched to “keep up with the Joneses” -- basically, lots of folks running around with hammers looking for nails. So many agentic AI projects today lack the thoughtfulness, planning and focus essential for any chances of success.

Models don’t fit complex business cases. Too many people are trying to solve every kitchen-sink issue in their company -- the more complex, the better. Someday, agentic AI will be ready for those approaches, but not yet. By the time someone can get a model to fit, the data won’t be ready, and the market will have moved on.

Problems with data quality and costs, lack of controls, and failure to drive business value. One of Gartner’s predictions for why so many agentic AI projects will fail: These projects will not solve for data quality, for inflating cost, for lack of controls and for failure to drive real value in the business.

We could look backward, replace the term “agentic AI” with “CRM” or “big data,” and see a history of failures for “bright, shiny object” tech opportunities of old for all those same reasons.

I am super-bullish on what AI will do to transform our industry. But I know that we’ll see a bunch of missteps along the way. Which is why it was a lot of fun to talk about this with such experienced entrepreneurs and CEOs, particularly since Chad and Mark are two of the very top executive coaches out there in the tech world these days.

What's your sense of things? Will “obsolescence before maturity” be the epitaph for all too many of our favorite AI projects of today?