When the Tools Stop Enforcing the Org Chart.

AI didn't just collapse the cost of building — it collapsed the role definitions our operating models were stacked on. The hard question shifts from "can we build this?" to "should we, for whom, and to what end?"

David Armano published a piece on March 29 called "We're All Builders Now" that I haven't been able to put down. He borrows Marc Andreessen's "Mexican standoff" image — a frozen tension between software engineers, product managers, and designers, all of whom can now do substantial parts of each other's jobs. Armano spent five evenings with Claude Code building a daily briefing product called The TBDaily. Two evenings to V1. By the third evening it had an AI-cloned voiceover. It costs him about fifty dollars a month to run.

That kind of timeline used to require four functions, six weeks, and three meetings about scope. The standoff isn't a turf war — it's the quiet realization in every team that the lines on the org chart are no longer being enforced by the tools.

Org charts in 2026 are descriptions of what was, not what is.

The Standoff Inside Every Cross-Functional Team.

Andreessen's frame is startup-flavored, and Armano is right to flag that it'll manifest differently inside the enterprise. But the underlying physics is the same. AI has flattened the prerequisite layer — the parts of the job that used to require a specialist — and what's left is judgment, taste, and intent. The "E-shaped worker" Andreessen describes (replacing the T-shaped one) isn't a generalist. It's someone who can now operate across multiple functions because the tools no longer gatekeep them.

I've seen four versions of this movie in twenty years across agencies and enterprise delivery. Mobile-first dissolved the boundary between art director and IA. Cloud collapsed ops into engineering. Service design swallowed half of what marketing used to own. Each shift taught the same lesson: the tools enforce the discipline. When the tools change, the roles change — whether or not anyone updates the job description.

What's different this time is the speed, and the direction. Previous shifts redistributed work toward new specialists who learned the new platform. This shift is doing the opposite. It's pushing judgment to the front of every conversation, and the operating models most companies are running were not built for that.

The Org Chart Was Always a Promise, Not a Fact.

Here's the part nobody likes to say out loud: most org charts were always fictional. They described how leadership wanted the work to flow, not how it actually flowed. The real org chart lived in Slack DMs, in the four people who actually got things done, in the unwritten rule about who you ran a deck past before sending it up.

What AI is doing is exposing the gap between the diagram and the reality. When a designer can ship a working prototype in an evening, the engineering review process that assumed a six-week build cycle starts to look like theater. When a PM can generate a decent first-draft PRD in twenty minutes, the rituals built around scarce PRD-writing capacity start to look like overhead. When an engineer can produce passable interface mocks, the design review that was the only quality gate starts to look optional.

Armano's reframe is the cleanest articulation of where this lands: "the builder barrier is no longer 'can I build this'? It's what should I build, why should I build this, for whom, and to what end?"

That's the part most operating models aren't ready for.

What Design Ops Was Really Holding.

I want to push back gently on a comfortable narrative I keep hearing in the design and program management Slacks I'm in — that design ops, program management, and product management are now redundant because everyone can build. That misreads what those functions were ever actually doing.

Design ops was never primarily about coordinating designers. It was about holding the standard — the design system, the review cadence, the intake process, the "no, that's not how we ship things here." Program management was never primarily about Gantt charts. It was about holding the intent — the reason a thing was being built, the constraints that mattered, the trade-offs that had been agreed and would otherwise get re-litigated daily by whoever showed up at the meeting.

When the tools stop enforcing role boundaries, those holding functions become more important, not less. If anyone can ship a feature in an evening, you don't need fewer people to maintain coherence — you need more, in different shapes. The risk isn't that AI eats design ops. The risk is that the standard collapses while everyone is celebrating throughput, and by the time leadership notices, the brand, the product surface, and the customer experience have all drifted into a kind of plausible mediocrity nobody decided to ship.

Brian Solis Has a Word for the Output.

Armano's piece introduces a word from Brian Solis, ServiceNow's head of global innovation, that I'm going to steal: slopware. Software, content, decks, prototypes, marketing emails — produced quickly with AI, indistinguishable from craft at first glance, hollow on contact. Every team has shipped some of it in the last six months. Most teams have shipped a lot of it.

Slopware happens when the cost of building drops faster than the discipline of deciding what to build. It's a leading indicator that your operating model hasn't caught up. And it's not solved by adding "AI guardrails" to the creative brief or stamping a new policy at the bottom of the JIRA template. It's solved by changing the unit of work the team is rewarded for — from artifacts to decisions.

That's the hinge of the whole thing.

The New Job Is Curating Intent.

If everyone can build, the program manager's real job is no longer sequencing — it's curating the intent across a team of people who can now all generate output faster than the organization can absorb it. Design leadership's real job is no longer reviewing pixels — it's defending the should against a flood of coulds. Product management's job is no longer translation between business and engineering — it's making sure the team is solving a problem worth solving, in a frame that matches the customer's actual life.

Armano puts the human side of this cleanly: "Every one of our teammates working with AI is all undergoing the same thing right now — amplifying what they do best, without outsourcing or offloading too much cognition. It's a delicate dance, we're all still learning."

A delicate dance is right. And the choreography is operating discipline. Without it, you don't get the E-shaped workforce. You get five overlapping people building the wrong thing in parallel, faster than ever, and calling it velocity.

A Working Model: Build Less, Decide More.

If I were redesigning a design ops or program management function this quarter, here's where I'd start.

Make the unit of work a decision, not an artifact. Track them. Review them. Audit them. Most teams I've worked with can't tell you the last ten meaningful decisions they made or who made them. That gap is where slopware breeds.

Move review upstream. The point of design review used to be to catch craft issues before they shipped. Craft is now mostly fine — AI handles the floor. The thing that needs review is the intent: are we solving the right problem, for the right person, in the right frame? If your review still spends 80% of its time on the artifact, you're reviewing the wrong thing.

Defend a small number of standards aggressively. Not many — three to five. The brand voice. The decision framework. The customer-facing language. The data hygiene rules. The "these are non-negotiable" parts. Everything else gets cheaper, faster, and more disposable. A short list of standards, fiercely defended, is the only thing that prevents slopware at scale.

Change what you reward. Stop celebrating shipped features. Start celebrating decisions that compounded — the ones still paying off six months later. This is the hardest one to operationalize, because it runs against every dashboard most companies have. But it's the only metric that maps to the world Armano is describing.

When everyone's a builder, the things we build should become more intentional, not less. That's Armano's line, and it's the right one. The role of the people who used to "coordinate the work" is no longer to coordinate it — it's to make sure the work is worth doing in the first place.

The org chart is fictional now. The intent is the operating system.

Further Reading:"We're All Builders Now" — David Armano, March 2026 · "Designed or Defaulted" — David Armano, April 2026 · "Six Shifts For 2026 (And The Next Six Years)" — David Armano · Brian Solis · Launch by NTT DATA