🌊 How SaaS Inverts Into Agents

Six frameworks for the post seat-based software world.

Feb 26, 2026

👋 I’m Ivan. I study how top 1% startups raise and grow.

This week’s sponsor is AI CRM Attio!

Attio is the AI CRM for modern go-to-market teams. Attio connects to your email, calendar, calls, product data, billing data and more — so your CRM is always complete, always enriched, always in sync.

Prep for any call in seconds with full context across your business
Instantly prospect and route leads with research agents
Build powerful AI automations for your most complex workflows

Try Attio Pro for FREE

Hello there!

I’m moderating a panel at 4YFN Barcelona next week called “How Agents Are Prompting the Business Playbook” with Alvaro Higes (CEO, Luzia, 65M users), Daniel (Chief AI Officer, WPP), and Elisenda (CEO, Cala AI).

To prep I went deep on every framework I could find about what’s actually happening to SaaS (and it’s been quite a week to do so) to answer one question:

If AI agents eat the business logic of software applications…
what happens next?

Today’s tl;dr:

The stack is inverting: Microsoft’s framework.
We’ve seen this movie before: OpenAI Chairman’s analogies.
Builders aren’t scared: Naval Ravikant’s motorcycle for the mind.
The flywheel was designed this way: the ultimate flywheel.
Keynes predicted this in 1930: economic possibilities for our grandchildren.
Judgment is the scarce resource: 5 skills for anti-fragility.

1. The stack is inverting

A year ago or so Satya was already talking about this shift.

Before LLMs, you had a human interact with the UI of an app, which ran business logic, which ran on top of a database.

But today, as Naval coined recently, “AI is eating UX.” A new surface area of interaction is emerging: agents. A human-like presence in a box that reads context, chooses actions, and calls tools for us (what LLMs enabled).

Which raised questions about where power is shifting, and why we’re seeing big tech splashing cash and new startups popping up everywhere to surf this wave - because there is a platform shift (and therefore, opportunity).

In this context, and as we know from our friends Benjamin Graham’s Mr. Market definition (irrational, often contradictory) and John Maynard Keyne’s “animal spirits”, the stock market got fearful. All of a sudden.

Here’s my best attempt at representing Satya’s framework for going from SaaS → Agents, which has been helpful to keep in mind lately:

2. We’ve seen this movie before

Came across this pod from Bret Taylor (Sierra founder, chairman at OpenAI) with tons of insight and a few super sticky analogies.

I noticed great leaders are often amazing at this, remember Jim Barksdale’s aphorisms.

The one that stuck with me most: Systems of record are like the sun in a solar system. They hold gravitational pull and everything orbits around them (think Salesforce).

The question is: does that gravity shift to the agents in this new “era”?

3. Builders aren’t scared

This opportunity window (SaaS → Agents) is also being turbo-charged with all these emerging tools, making high-agency entrepreneurs-to-be much more leveraged.

I found some gems in Naval’s recent podcast on this:

4. The flywheel was designed this way

Matt Shumer’s essay “Something Big Is Happening” hit 80 million views recently, because it touched a social fiber that is really tender: agents are already here.

He openly talks about how recent this change has been (Opus 4.6 on Claude released earlier this month, and GPT-5.3 Codex idem), and I couldn’t agree more with him that you really don’t know what you are talking about until you feel the difference using these models and how they’ve dramatically improved in a matter of weeks.

To the point where he openly says:

I am no longer needed for the actual technical work of my job

His most important insight is that AI labs made a deliberate choice to make AI great at code first because building AI requires code. And if AI writes code, it helps build the next version of itself. Smarter version writes better code. Better code builds smarter AI. This is basically the flywheel of flywheels.

Which explains the acceleration we’re all feeling:

5. Keynes predicted this in 1930

I re-read this essay from one of my favorite Economists.

In 1930, with 20% unemployment and the world falling apart, Keynes wrote “Economic Possibilities for our Grandchildren” and made two predictions for 2030:

Living standards would rise 4-8x (correct, US is up ~6x)
and we’d work 15-hour weeks (spectacularly wrong lol).

But his real insight wasn’t about hours or GDP but rather what happens to humans when the economic problem is solved (and we may be getting there faster than many would like to admit, especially when it comes to “knowledge work”).

“We have been trained too long to strive and not to enjoy.”

He coined “technological unemployment” and called it “a temporary phase of maladjustment.” The timeline comparison is what gets me:

Agricultural revolution: thousands of years.
Industrial: over a century.
AI: possibly less than a decade (?)

At the same time, we got the folks at McKinsey saying 57% of US work hours are technically automatable. And many other “anecdotes” that are still lagging in labour statistics, but that in my honest opinion are going to get there pretty fast (we’ll see).

Meanwhile, at the “frontier” Salesforce cut support 9K→5K, Klarna downsized 40%, I published data on revenue per employee placing some of the fastest growing AI companies by demand (not dollars raised or fomo, by corporate spend), show how much revenue this new wave can generate with how few people.

And yet, there is no discernible disruption on broader employment (for now).

All I’m saying is that we should probably start thinking much more deeply about the question Keynes’ asked in 1930:

when intelligence is abundant, what are humans for?

6. Judgment is the scarce resource

AI is making teams dramatically faster but not necessarily wiser.

Which means they’re (increasingly?) not bottlenecked by intelligence anymore but bottlenecked by judgment.

Naval said it differently but also thought it was interesting:

AI doesn’t want anything, has no survival instinct (that we know of?), and no real goals. It remixes what exists really well but can’t come up with something truly new.

The thing that makes entrepreneurs different is extreme agency and AI doesn’t have it.

So if intelligence is getting cheap and abundant, the skills that compound are the ones it can’t replace.

Five stand out, and they’re all forms of the same thing: