If you're not first, you're last
The AI-native enterprise is here
AI is here to stay, and there is no depth of sand in which you can bury your head that will change that reality. Embracing AI is an existential test that most businesses will fail.
The road to irrelevance will be paved with excuses, timelines, and fear of misstep.
But the only fatal misstep is to take no step at all.
You probably recognize it in your own business, today. The “AI committee” of cross functional leaders that meets every 4 weeks to discuss the “challenges” of AI, which holds supreme the need to “map use cases” before “doing anything,” and points to a so-called “multi year roadmap” as if a multi year roadmap was ever feasible, even before the age of intelligence that’s evolving faster than you can shake a hype cycle at.
If you’re in this stage, you’re behind.
You’re further behind than you know, and the compounding gains of AI are going to make you more irrelevant than Bing’s Sydney. Doesn’t ring a bell? Yeah.
Mock the giants all you want, but they get it. And they got there building in public, leading from the front, without fear of misstep. Marc ripped Dreamforce to shreds 8 weeks before it started and pivoted the entire company to be agent-first, before anyone really knew what agents were.
That’s Marceting. [sic.]
And it takes your company 8 weeks to collect surveys on “What we should do with AI.”
Talk doesn’t win championships. You can’t yap your way to victory1. Pessimists sound smart, optimists make money. Work for, and with, optimists.
TikTok might get mad at you. Some of your employees might, too. But your job is to lead, not to be loved.
And as a leader today, it’s your job to create the conditions for AI to permeate your company. It should be an expectation. Power at today’s organizations is implicitly drawn from the size of the teams you lead. You need to make it so that in the future, power comes from the outcomes you drive.
Investors have historically moved heaven and earth for a few points of operating leverage. Markets are won over less. Today we have the tools to drive operating leverage at a level that was previously unfathomable. And you’re whiffing because you’ve empowered an AI council to make decisions by the notoriously reliable consensus.
This isn’t as simple as replacing a person with an agent, though it may be for markets where the agent abstraction already exists. Work gets done by workflows. And workflows can be done by agents.
What is a job, but responsibility over a set of workflows? As these workflows transition to agents, jobs will change. Jobs whose workflows can be sufficiently (or superiorly) done by agents will be wholesale replaced. But most jobs will simply change.
There are many things that computers do better than people. And coincidentally, they’re often the monotonous things that people dislike about their jobs. To hand these things over to agents is a gift, not a curse. Let people be people, computers be computers, and bask in the leverage that comes.
We used to do a lot of inefficient things by hand. The definition of inefficient has just changed. And the pace at which that definition changes is only going to increase.
Humans are notoriously bad at comprehending compounding change.
You think you have time, but you don’t.
If you’re not first, you’re last.
1 With notable exceptions for those employed as venture capitalists, or as the President of the United States.