Yahoo Finance
Yahoo Financetoday
Tech

AI Is Erasing The Corporate Job Ladder

22 min video4 key momentsWatch original
TL;DR

May Habib, CEO of Writer, says half of Fortune 2000 work will be done by AI agents within 18 months—and companies must choose between building their own digital workforce or renting it from tech giants.

Key Insights

1

Half the work in 18 monthsWithin 18 months, half of work in Fortune 2000 companies will be done or heavily augmented by AI agents—not just productivity tools for individuals, but organizational digital workforces.

2

Narrow job ladders are deadTraditional narrow job ladders—analyst relations, competitive analyst, campaign analyst—are obsolete because agents can be specialists while humans must become generalists to leverage them effectively.

3

5-10% become power usersOnly 5-10% of trained users become power users of AI agents; success requires isolating those people to build critical workflows, then distributing what they built company-wide rather than training everyone equally.

4

Layoffs are lazy not braveCEOs who simply lay off workers and lean into AI are being lazy, not brave; real bravery means telling people their seat will change fundamentally while building them up alongside agents, not replacing them preemptively.

5

Agents get smarter dailyWriter's agents learn and improve daily from organizational data—unlike base LLMs—which is crucial for siloed enterprises where AI talent is scarce and knowledge hoarding is endemic.

Deep Dive

The Digital Workforce Era Is 18 Months Away

May Habib opens with a bold prediction: in 18 months, half of all work inside Fortune 2000 companies will be done or heavily augmented by AI agents. He's not talking about ChatGPT plugins or productivity aids—he means treating AI as a true digital workforce with deep functional expertise. Writer, his 5-year-old company, helps enterprises build agentic solutions that codify business knowledge into autonomous workflows. The key shift Habib emphasizes is ownership versus rental. Companies can build their own digital workforce and own the IP moat, or they can rent capability from the two big LLM providers. That choice will define how much leverage and transformation a company actually gets. He's careful to position Writer as infrastructure for companies that want sovereignty and real organizational change—not just another tool for individuals to toy with.

The Narrow Job Ladder Is Dead

Habib doesn't mince words: traditional corporate hierarchies where you climb from analyst to senior analyst to manager are finished. The culprit is horizontal job overlap and jurisdictional redundancy. When someone builds an agent, they're naturally bumping into someone else's turf—and that kills momentum because people get protective of their roles instead of reimagining workflows. He saw this play out: a 54-step manual handoff process got reduced to 15 steps instead of one because people won't let go of their functions. Real transformation requires erasing narrow job descriptions like analyst relations, competitive analyst, and campaign analyst. What companies need instead are generalists—people who can think about leverage differently, who understand how to multiply their impact through agents rather than climb functional hierarchies. The profile of people who will lead in this era looks nothing like the people promoted over the last few decades. Few CEOs have been brave enough to call this out and rebuild their organizations around different people.

Bravery Isn't Layoffs—It's Transformation With Honesty

When Brian brings up Block's aggressive layoffs as an example of boldness, Habib pushes back hard. He says brave does not mean firing everyone and starting over—that's lazy. Real bravery is harder: looking employees in the eye and saying the company will stay roughly the same size but will deliver 10x more through AI, and the seats themselves will change fundamentally. Some people will thrive in that new world if they can leverage AI; others won't have a place. But you can't know which is which if you fire indiscriminately—you might lose the exact people who could excel in the new system. Habib distinguishes between tech companies, which grossly overhired during COVID for manual tasks that AI can now augment, and enterprises in regulated industries where there are shop floors, operations, and compliance roles that will shift but not disappear. The missing element isn't the technology; it's leaders willing to be transparent about structural change and invest in bringing people along rather than out the door.

Agents Learning From Each Other Is the Secret Sauce

One of Writer's recent breakthroughs is agents that talk to each other inside a company and learn from what colleagues are building. Habib uses his own one-on-ones as an example: agents prep him by picking up on things his leadership peers are doing—patterns he didn't even know existed—and suggest adjustments to his own meeting approach. This is intelligence multiplied. Most AI vendors haven't cracked this because it requires agents to improve daily from real organizational data, not just from base LLM training. For siloed enterprises where AI talent is scarce, this matters enormously. Only 5-10 out of 200 trained users become power users obsessively building new workflows. That small cohort becomes invaluable. Writer's model takes what those power users build and distributes it to everyone else—turning isolated brilliance into organizational leverage. This is why Habib believes Writer has succeeded where others haven't: they're not trying to democratize AI to everyone equally; they're multiplying the output of the few who truly get it.

Legacy SaaS Can Survive But Must Add an Agentic Layer

When asked if companies like Salesforce and ServiceNow can reinvent themselves, Habib says absolutely—they're not going away. But every company needs an agentic layer on top. What's happening is companies are building personalized, ephemeral software experiences on top of Salesforce or Snowflake data warehouses. His own one-on-one apps run largely on Salesforce because that's the source of truth for customer data. The interface might look like a simple chat, but underneath there's massive differentiation. The risk for legacy SaaS is that their interfaces haven't evolved to account for humans supervising agentic execution instead of doing the work themselves. Project management looks totally different when agents run the execution and humans supervise. If the systems of record don't evolve fast enough, they'll lose relevance. But they have a moat in the data they own, so they have time to adapt. Habib sees the real TAM as an entirely new market: software to manage digital labor, optimize agentic workflows, and coordinate the hybrid workforce. That market hasn't been invented yet.

Takeaways

  • Identify the 5-10% of your team who are already thinking in terms of leverage—put them to work on critical workflows first, then distribute their solutions company-wide rather than training everyone equally.
  • Audit your org chart for narrow functional ladders that no longer make sense—begin planning how to transition those roles into leverage-focused generalist positions before AI forces the issue.
  • Build your own agents on your proprietary data if you can afford it, rather than renting from LLM providers; ownership of your digital workforce defines how much IP moat and competitive advantage you retain.

Key moments

2:08Half of Fortune 2000 work in 18 months

In a year, a year and a half, half of the work that gets done today inside of the global 2000 will be done or at least largely augmented by a digital workforce.

6:30Narrow job ladders are dead

These like narrow job ladders are a thing of the past. They're dead already. Like there is no world in which these horizontal job descriptions from analyst relations to competitive analysts to campaign analysts are just going to go away.

9:00Layoffs are lazy not brave

Brave does not mean layoffs. And I actually think that is lazy, not brave. You need leaders who can look people in the eye and say this will most likely be a smaller company, but not by a lot. But I need a company here with 10x the footprint we have today.

13:00Agents that learn from each other

We have built agents that talk to each other, right, inside of a company, right, and inside of an organization. And the intelligence that is possible is just incredible. Agents are prompting me to do my work differently based on what others are doing.

Get AI-powered video digests

Follow your favorite creators and get concise summaries delivered to your dashboard. Save hours every week.

Start for free