I’m going to say what everyone’s thinking but nobody wants to admit in public:

Most project managers are expensive human middleware, and AI does their job better.

What PMs Actually Do

Let’s be honest about the role:

• Status updates — asking developers “what’s the status?” and reformatting the answer into a spreadsheet
• Meeting scheduling — coordinating calendars and sending Zoom links
• Timeline estimation — asking engineers how long something takes, then multiplying by 2
• Stakeholder translation — converting “we need it yesterday” into “Q2 priority”
• Risk tracking — maintaining a list of things that might go wrong (spoiler: they already know)
• Ticket hygiene — nagging people to update Jira

Now tell me: which of these requires a human making $120k/year?

What AI Does Better

I’ve been running projects with Claude as my “PM” for the last 3 months. Here’s what happened:

Status Updates

Before: PM Slacks me every morning: “Hey! How’s the API refactor going?” Now: AI reads my commits, checks branch status, cross-references the project plan, and generates a status report in 10 seconds.

No interruptions. No context switching. Just data.

Timeline Estimation

Before: PM asks me to estimate, I say “3 days”, they write “1 week” in the roadmap to be safe. Now: AI analyzes similar past tasks, accounts for dependencies, and gives me a realistic range based on actual velocity. Then I add 20% buffer myself because I’m not insane.

Meeting Scheduling

Before: 6-message Slack thread negotiating who’s free when. Now: AI checks calendars, finds the optimal slot, sends invites. Done.

Risk Tracking

Before: PM maintains a “risk register” in Confluence that nobody reads. Now: AI scans PRs, CI logs, and recent commits, flags actual risks (flaky tests, breaking changes, missing migrations), and surfaces them in priority order.

No theater. Just signal.

The Uncomfortable Part

Here’s what kills me: the good PMs know this already.

The ones who actually add value aren’t doing admin work. They’re:

• Unblocking political bullshit
• Making hard prioritization calls
• Keeping execs from derailing sprints
• Protecting the team from chaos

That’s leadership, not project management.

The rest? They’re glorified Gantt chart operators. And Gantt charts don’t need a salary.

What About “Soft Skills”?

“But AI can’t handle people problems!”

Sure. But most PM “people problems” are:

• Reminding someone to finish a ticket (automated)
• Asking for updates (automated)
• Scheduling a sync (automated)
• Translating engineer-speak into exec-speak (automated)

The actual hard people stuff — conflict resolution, morale, team dynamics — that’s not the PM’s job. That’s the manager’s job. And if your PM is doing that, they’re actually a manager with the wrong title.

The Real Test

If your PM disappeared tomorrow and you replaced them with:

• A decent task board (Linear, Jira, whatever)
• Automated status reports (GitHub Actions + Slack webhooks)
• A shared calendar
• One weekly sync

…would anything break?

If the answer is no, you don’t need a PM. You need better tooling.

Who Survives?

The PMs who survive this are the ones who:

• Actually understand the product deeply
• Make strategic calls, not just coordinate tasks
• Unblock teams by navigating org politics
• Own outcomes, not just timelines

If your entire job is asking “what’s the status?” and updating spreadsheets, yeah — you’re cooked.

AI doesn’t get tired. It doesn’t forget. It doesn’t need vacation. And it sure as hell doesn’t send passive-aggressive Slack reminders at 4:55pm on Friday.

The Transition

We’re already seeing it:

• Startups aren’t hiring PMs until Series B
• AI tools (Notion AI, Linear AI, GitHub Copilot Workspace) are doing 60% of traditional PM work
• The PMs who remain are shifting into product strategy or engineering management

In 3 years, “project manager” will sound as outdated as “secretary” does now. Not because the work disappeared — because it got automated.

The Memo

If you’re a PM reading this, here’s my advice:

1. Automate yourself out of the boring parts. Use AI to handle status updates, reporting, scheduling. Free yourself up for actual strategic work.
2. Go deep on product or go deep on people. Be a product manager or be a people manager. The middle ground is getting AI’d out of existence.
3. Own outcomes, not processes. If your value prop is “I keep things organized,” you’re a tool. Be the person who makes hard calls and ships results.

And if you’re on a team with a PM who’s just human middleware?

You don’t need to fire them. Just stop asking them to do the work AI can handle, and see what’s left.

---

Disclaimer: I know some PMs will read this and get mad. That’s fine. But deep down, you know I’m right. The best PMs are already acting on this. The rest are updating their LinkedIn.