The AI-Native Company

Heading into Black Hat / DEF CON this week I think the biggest idea in tech right now isn’t the attacker vs. defender question as models get smarter.

Nor do I think it’s the question of if/how we should control AI models, who should make the rules, and how to ensure we don’t have an AI oligarchy.

Those questions are extraordinarily important, but I think the most important idea in tech right now is how AI changes the way businesses operate.

At the surface level it sounds something like, “AI will do more of the work in the enterprise,” or something like that.

“Jobs will be lost because AI can do a lot of knowledge work.” Etc.

This is true, of course, but there will also likely be a whole lot more companies, and more jobs for people at the same time. Maybe not the same people, and that’s worth talking about as well.

But I think the biggest idea isn’t just AI doing lots of work inside companies.

It’s the complete restructuring of companies themselves.

• The definition of the company’s ideal state, its goals, its metrics, its challenges, its strategies, its projects, its work, its people, etc. into primary constitutional documents
• The collection of internal knowledge into context domains.
• The turning of internal policies into official SOPs.
• The creation of thousands of agentic workflows that execute the company’s work according to those SOPs.
• And the publication of company activity to internal company daemons that make the company largely transparent across people, agents, departments, and the company overall.

Essentially, businesses become articulated, purpose-driven, and transparent. And the people at the company become the architects, stewards, and orchestrators.

Every company will work like this. Highly-competent generalist humans powered by their AI assistants and harnesses will be able to go from idea to testing and implementation by themselves, and within hours or days instead of a whole team taking months or years.

The closest analogy is probably the transition from paper and pencil to spreadsheets and databases. But far more extreme.

Ultimately we’re moving to a place where businesses will be able to define their ideal state using these primitives of goals, metrics, SOPs, agent workflows, etc., and the company’s AI will steward the transition from current state to that ideal state. All overseen by, and managed by, the human leadership of the company.

The question won’t be how we can find someone to do X or Y work with AI; it’ll be how we can turn that work into an agent workflow that does the work (and verifies it was done properly) according to the official company SOP.

This change will be extremely disruptive to companies that insist on working the old way, and most of those companies will not survive.

And as humans, we’re going to have to learn how to function in this world where the idea people, and the engineers, and the product people, and the go-to-market people start to merge.

Disruptive.

Inevitable.

Terrifying.

Exhilarating.