The Shift from SaaS to GaaS:Generative Agent as a Service

Be honest. SaaS made powerful tools affordable and accessible. The software made the work easier. It didn’t do the work.

Jensen Huang, NVIDIA’s CEO, the guy whose chips basically power the entire AI revolution stood up at his GTC 2026 keynote and said something that should have made every software company uncomfortable:

Instead of handing your team a dashboard to operate, an AI agent does the work itself. A new lead comes in, and the follow-up happens — without anyone clicking through the process.

I keep hearing people talk about AI agents in this vague, futuristic way. So here’s what it actually looks like when it’s working.

A customer order that used to pass between three teams is processed the moment it lands, and the end-of-month marketing report assembles itself. People step in only when something needs a judgement call.

If you think this is still theoretical, the stock market disagrees.

Investors are recalculating what software is worth when fewer people need a seat. Atlassian reported its first-ever decline in enterprise seats, and one customer went from 50 seats to 15.

Now, before this starts sounding like every tool you’ve invested in is about to become worthless – it’s not. Your CRM stays. Your ERP stays. Your accounting software isn’t going anywhere. What changes is who uses it.

Think of it like a self-driving car: the road and the car are still there. The agent is simply the one driving.

This is the part that should genuinely matter to anyone signing software contracts.

Instead of paying per user, you pay for outcomes — and the vendor carries the risk of delivering them. Salesforce already charges per resolved customer-service conversation rather than per seat.

You don’t need to rip everything out. But you do need to start asking a different question. Not “what software should we buy?” but “what work could an agent do?” Start with tasks that are repetitive and predictable:

I want to be clear about this because it’s the elephant in every room where AI gets discussed.

Agents take the mechanical operating of software off your team’s plate, so skilled people spend their time on strategy and creative work instead.

The companies that get there first won’t just be faster. They’ll be running a different kind of operation entirely — leaner, sharper, and increasingly hard to compete with.