AI doesn’t kill engineers. It kills excuses.

Every tech conference these days pushes the same lazy narrative: Artificial Intelligence (AI) will replace software engineers. Consultants sell it, journalists repeat it and the c-suite believes it.

They’re wrong.

AI doesn’t reduce the need for engineers but multiplies it. It eliminates the layers of organizational theatre that once stood between problems and solutions. The coordination overhead, the translation functions and the process management roles that only existed when engineering was slow and expensive.

AI doesn’t kill jobs. It kills excuses.

For non-engineering roles that have multiplied across enterprise tech, that’s the real threat.

The automation pattern

Automation doesn’t shrink professions. It expands them.

Spreadsheets didn’t kill accountants. VisiCalc and Excel destroyed manual calculation, not accounting. They removed the constraint on financial analysis. Suddenly, every business decision needed modeling; every department demanded forecasts. The profession grew.

The same happened with desktop publishing and graphic designers, industrial robots and factory workers, ATMs and bank tellers. Automation eliminates friction. When friction vanishes, demand surges.

Software is no different. AI coding assistants don’t replace engineers; they remove building constraints. That complex feature? Ship it. That painful integration? Deliver it this quarter. That technical debt? Fix it now.

When one engineer with AI produces ten times the output, companies don’t cut their workforce by 90%. They build ten times more products. They enter new markets. They solve problems they used to dodge. The constraint shifts from capacity to imagination. And imagination demands more, not fewer, engineers.

The exposed value chain

When engineering velocity jumps by an order of magnitude, every other role in the value chain becomes visible. And many look fragile.

These roles existed to manage friction: prioritization, translation, coordination, and governance. When AI removes that friction, the roles are exposed; their purpose is obviated.

What does a Product Owner do when requirements can be tested in production within hours? What’s the point of a Business Architect when strategy itself can be coded into systems? What does a Transformation Lead add when change is continuous?

These aren’t rhetorical questions. They’re existential ones.

Role by role

• Product Owners: Most are backlog administrators. They maintain lists and translate vague requests into tickets, prioritizing by stakeholder noise.

AI breaks this model. If engineers can prototype five options faster than writing one specification, the backlog is meaningless. If user feedback arrives in hours, not sprints, roadmaps are fiction.

The survivors will be genuine owners: Defining outcomes, measuring impact, taking responsibility for results. Not list managers. Not feature factory clerks. Actual owners.

• Business Architects: Most produce artifacts—diagrams, capability models, process maps—that are ignored the moment they’re published. They document what exists instead of designing what should exist.

AI demands executable models, not pictures. It also demands business rules that compile, strategies you can test, and constraints enforced programmatically.

If you can’t express business architecture in code, you’re ornamental.

• Enterprise Architects: Most practice archaeology, not architecture. They document what evolved, invent principles after the fact, and enforce governance by committee.

In the AI era, architecture is code. Infrastructure is code. Policy is code. Compliance is code. Continuous, enforced, automated.

The Enterprise Architects who survive won’t publish a ‘target states’ deck. They’ll build and code platforms with Terraform.

• Solution Architects: Traditionally, they produce options papers. They debate trade-offs and create slide decks that take longer to prepare than building the solution itself.

The approach assumes engineering is expensive and mistakes are fatal. With AI, an engineer can prototype three working solutions in the time it takes to draft one decision paper.

The solution architect who builds wins. The one who deliberates loses.

• Transformation leads: They depend on transformation being slow and painful. By the time delivery occurs, the workshops, stakeholder maps and roadmaps they rely on are already obsolete.

If systems evolve daily, if change happens continuously, if rollback is trivial, what is there to ‘lead’? The theatre of transformation—town halls, champions, readiness assessments—becomes absurd when the system is fundamentally different.

• Scrum Masters: They facilitate ceremonies, update tracking tools, create visibility charts. They protect teams from dysfunction that shouldn’t exist in the first place.

AI automates coordination. Dashboards expose blockers in real time. Teams self-organize around outcomes. What’s left to facilitate?

The Scrum Masters who will remain relevant in the future will be those focused on eliminating systemic blockers and reducing friction, rather than simply managing processes.

The speed test

AI isn’t magic. It’s speed.

And speed is a mirror. It reflects value or the lack of it.

When engineering accelerates tenfold, every other role too must prove proportionate value or disappear. There’s no room for theatre. No confusing motion with progress. No mistaking documentation for delivery.

The question becomes unavoidable: What value do you create?

Not which meetings you attend. Not what documents you generate. Not which processes you manage.

It boils down to: What exists because of your work that wouldn’t exist without you?

The simplified organization

In the AI-accelerated enterprise, three roles matter:

• Engineers who build systems, solve problems, create value in code. Their numbers grow as AI amplifies output.
• Outcome Owners who define success, measure results, and take responsibility for business impact. Titles vary; accountability doesn’t.
• Platform Architects who encode decisions, enforce standards, and build the foundations that enable velocity.

Everyone else—the coordinators, translators, documentarians—are friction. And friction has no place when speed is infinite.

The reckoning

For years, these roles justified themselves by complexity. They thrived because engineering was slow, expensive, and hard. They survived by managing the gaps.

AI closes those gaps. It makes the dysfunction obvious. It turns theatre into redundancy.

This reckoning isn’t about engineers. They’ll be fine. It’s about everyone else: Those whose jobs consist of talking about engineering rather than doing it. Those whose contributions are ceremonies, documents, and slides. Those who depend on slowness to make themselves useful.

AI is asking one question, everywhere, without pause: What exactly do you do?