A lot of modern white-collar work was built around process, coordination, SaaS rituals, and status performance. AI did not need to be perfect to threaten it. It only needed to be better than low-signal process.
Once work becomes measurable by output instead of meetings, a lot of job descriptions start looking like costumes.
The Room Where It Happens
Big-company work became entangled with politics and ceremony.
There is a version of white-collar work that produces things. A product ships. A customer renews. A campaign changes demand. A model improves decisions. A support queue gets healthier. That work is not theater. It is difficult, often underappreciated, and very real.
But around that work, especially inside big companies, another economy grew up. It was made of alignment meetings, stakeholder maps, visibility rituals, review cycles, strategy decks, internal roadshows, dependency negotiations, and the quiet art of seeming important without being pinned to an outcome. The theater economy did not look fake from the inside. It looked like professionalism. It had vocabulary, dashboards, ceremonies, and calendar density. It had all the surface area of serious work.
The trouble is that AI is unusually good at embarrassing work whose main product is a trail of process. It summarizes the meeting. It drafts the memo. It turns the transcript into action items. It writes the first version of the analysis, the email, the brief, the plan, the backlog item, the landing page, the test query, the code scaffold, the sales deck, and the content calendar. It does not need to be the best employee in the company. It only needs to compress the ritual enough that leaders notice how much of the ritual was padding.
Sprint Theater
Engineering orgs became governments, and Product became the embassy.
Software companies did not merely build software. They built political systems for deciding what software was allowed to exist. Engineering accumulated power because engineering controlled execution. Product became the mediator between executive desire, customer pain, design preference, revenue urgency, and what engineers would actually accept as a sane request. This was not always bad. Complex systems need translation. But the translation layer became its own career ladder.
The sprint was supposed to create cadence. In many companies it became a ritual for converting uncertainty into the appearance of control. Tickets moved. Points were estimated. JIRA boards glowed with activity. Asana tasks multiplied. Standups created the feeling of momentum. Planning meetings created the feeling of seriousness. Retros created the feeling of learning. Then the same unresolved tradeoffs returned next quarter with a different document title.
AI attacks this arrangement from two sides. First, it lowers the cost of execution. The distance between a product idea and a working prototype collapses. Second, it lowers the status value of coordination that cannot show its effect. If one strong operator with AI can produce a credible spec, prototype, test plan, and user-facing copy in a day, the organization has to ask why it needed three weeks of stakeholder choreography to discover the same thing.
The Silo Tax
Disconnected arms of the company learned to optimize their own theater.
Silos are not just org-chart inconvenience. They are theater multipliers. The promotional arm tells a story the development arm cannot support. The sales arm promises a feature the product arm has not prioritized. The data team defines a metric the marketing team does not trust. The content team fills a calendar while the customer team sees the same complaints repeat. Everyone is working. Everyone is busy. The actual customer experience moves sideways.
This is how companies end up with internal markets for attention. Teams do not only produce work; they produce proof that their work deserves resources. They create dashboards, trackers, weekly updates, initiative names, and strategic narratives. The more disconnected the company becomes, the more time it spends explaining itself to itself.
AI is a threat here because it can connect surfaces that used to require human bureaucracy. It can read a call transcript, compare it to churn notes, inspect the docs, draft a product fix, produce a customer email, and summarize the revenue risk. It may still need judgment and review. But it makes the old silo tax harder to defend. If the machine can move context across walls faster than the organization can schedule a sync, the walls start to look expensive.
The SaaS Ritual Layer
We bought tools to connect work, then built work around the tools.
The SaaS era had one reliable move: build a platform that connects something. Connect the CRM to the marketing platform. Connect the marketing platform to the analytics platform. Connect the analytics platform to the warehouse. Connect the warehouse to the dashboard. Connect the dashboard to the weekly meeting where nobody agrees what the number means.
This created jobs. Not useless jobs, exactly, but jobs whose value depended on the persistence of the maze. Admins, ops specialists, enablement leads, implementation managers, lifecycle strategists, integration consultants, reporting owners, and tool-specific experts became necessary because the stack itself became a terrain. A company could spend millions on software and then hire people to coax the software into admitting what happened.
AI makes this awkward. If an agent can query systems, move data, generate the report, draft the follow-up, and explain the tradeoff, the stack stops being a kingdom. It becomes plumbing. The old software economy rewarded owning the dashboard. The new one rewards proving the decision that came from it was correct.
Content Calendar Purgatory
Marketing became very good at manufacturing activity.
No department made theater look more productive than modern marketing. The playbook was elegant in its emptiness: buy paid keywords, post on social, fill the blog, refresh the content calendar, produce the webinar, write the ungodly SEO article, pull the attribution dashboard, declare a learning, and repeat. Every motion could be justified. Every motion had a metric. Many of the metrics were just shadows on the wall.
Keyword-stuffed organic content is the purest artifact of this world. Nobody loved it. Writers hated writing it. Readers hated reading it. Search engines tolerated it until they did not. Companies published it because the model said attention could be harvested if enough pages were produced with enough semantic obedience. Then AI arrived and made mediocre content effectively infinite. The floor collapsed because the floor had always been made of volume.
Paid media had its own theater. Budget moved through channels, dashboards attributed value, and everyone argued about what deserved credit. I come from media measurement, so I have sympathy for the difficulty. Attribution is hard because human behavior is messy. But the legacy attribution world often turned uncertainty into a ritual of false precision. We assigned neat percentages to chaotic influence, then treated the spreadsheet as if it had witnessed the customer deciding.
Output Wins
AI does not have to replace genius. It replaces the excuse layer first.
The funniest and meanest truth about AI and white-collar work is that the target was never just excellence. The target was mediocrity protected by process. AI does not need to be a brilliant strategist to threaten a strategy deck that nobody uses. It does not need to be a world-class engineer to threaten a workflow where a simple integration takes six meetings and a quarter of political budgeting. It does not need to be a great writer to threaten content nobody wanted to read anyway.
This does not mean judgment disappears. It means judgment gets pulled closer to output. The valuable person becomes the one who can decide what matters, frame the problem, evaluate the result, and take responsibility for the next move. The vulnerable person is the one whose job depends on the organization being too slow, confused, or polite to ask what changed because they were there.
White-collar theater had a good run because measurement was expensive and coordination was hard. AI makes both cheaper. Once work becomes measurable by output instead of meetings, a lot of old job descriptions start looking like costumes.
JobsJudo Ice Breaker
Fresh roles, cleaner shots.
Bring better signal into the search before another good-fit opening turns stale.
Next Part