AI doesn’t eliminate roles — it strips away the execution layer and exposes the decision-making underneath. That distinction matters more than most teams realise.
The role of the content manager isn’t to write content. It’s to ensure the right content exists, is correct, and serves a strategic purpose. AI doesn’t threaten that job. It finally makes it possible.
The observation that started this
A colleague made a comment recently that I’ve been thinking about since. We were talking about how I run douli.com — one AI agent connected to WordPress, handling the work of a content writer, SEO specialist, brand manager, and developer — and he said: “You got rid of a lot of jobs with AI.”
He wasn’t wrong, exactly. But he wasn’t quite right either.
What I actually did was redesign which work happens where. The writing, the SEO formatting, the code, the audit — that’s now delegated. What I do is define the constraints, review the output, and make the calls that require judgment about my own brand and positioning. That’s not a diminished version of those roles. It’s the version that would have existed if execution weren’t so expensive.
The content manager problem
Here’s a more concrete example. Most digital teams have a content manager or content lead. That person is usually excellent at their job. They’re also usually drowning.
They’re writing new content to hit a publishing schedule. They’re briefing writers, chasing approvals, updating pages someone flagged as outdated last quarter. What they’re almost never doing is auditing the existing archive — identifying which posts need a structural rewrite, which ones are cannibalising each other in search, which ones need internal linking work, which need their SEO metadata cleaned up. Not because they don’t care. Because there aren’t enough hours.
If you give that content manager an AI workflow that handles the audit, the rewrite brief, the SEO gap analysis, and the draft restructure — her job becomes: reviewing what the AI found, deciding what to prioritise, applying judgment about tone and brand consistency, and making calls about what gets published and what gets cut.
That’s the job. That’s what a content manager with unlimited execution capacity would actually spend her time doing. That’s not a smaller role. That’s the role done at the level it was always meant to be done.
What AI actually makes visible
The pattern repeats across every function I’ve watched closely. Execution hides the decision layer.
When you’re spending 80% of your time producing outputs — writing, formatting, checking, cleaning, building, summarising — it becomes genuinely difficult to see where the value actually sits. The busyness obscures the judgment. And because the judgment is hard to separate from the work that surrounds it, it’s also hard to scope, hard to delegate, and hard to measure.
AI doesn’t create the decision layer. It just makes it visible. When the execution is delegated, what’s left is the thing that was always worth paying for.
This has a useful corollary. If you remove the execution from a role and nothing of substance remains — no decisions, no strategic judgment, no institutional knowledge that shapes outputs — that’s important information. Not necessarily a case for elimination, but certainly a case for redesign.
The product angle
As a Product Owner, I already live with this distinction. The job title implies ownership of outcomes, but a significant chunk of the role is execution: writing user stories, running refinement sessions, chasing clarifications, preparing sprint reviews, documenting decisions. Good POs find ways to protect the judgment layer — the prioritisation calls, the stakeholder alignment, the trade-off decisions. But it’s a constant negotiation with the volume of operational work.
AI shifts that negotiation. When I use a well-designed AI workflow to draft acceptance criteria, structure a product spec, or prepare a stakeholder update, I’m not doing less thinking. I’m doing more of the thinking that matters — reviewing output critically, catching what’s missing, making the calls it can’t make. The quality of the product thinking goes up when the mechanical execution goes down.
The question worth asking about any role on your team — including your own: if AI handled the execution, what would this job be? If you can answer that clearly and it sounds like valuable work, the role has a strong future. If the answer is unclear, that’s worth exploring before someone else explores it for you.
It’s a redesign problem, not a headcount problem
The “AI takes jobs” framing misses the more interesting question. The question isn’t how many roles AI displaces. It’s which organisations figure out, early enough, how to redesign roles around the judgment layer rather than the execution layer. The ones that do will have people doing more valuable work. The ones that don’t will have people competing with AI on the thing AI is better at.
The content manager who isn’t auditing her archive isn’t failing. She’s constrained. Give her the right tools and the right remit, and the job gets better, not smaller.
TL;DR: AI doesn’t eliminate roles — it strips away execution and exposes the decision-making layer underneath. The question isn’t whether AI will replace people, it’s whether organisations will redesign roles around what humans are actually for.
Delphine Ragazzi is a Product Owner with 20 years in digital analytics, CRO, and product delivery. She writes about product decisions, data, and AI at douli.com. About Delphine
