Podcast circuit is actively debating AI-native SaaS pricing swinging from per-seat to usage/outcome-based and back

The podcast circuit is wrestling with a pricing problem that matters more than most founders realize. AI-native SaaS companies are bouncing between per-seat models, usage tiers, and outcome-based pricing, and nobody has figured out what sticks. One recent episode of The Series A Show dropped a striking claim: companies going all-in on AI are now spending $7,500 per employee per month on tools. That number is unverified, but even if it’s directionally true, it points to a real mess.

The old per-seat model made sense when software was static. You bought licenses, you knew the cost. But AI tools scale differently. Usage can spike unpredictably. A single power user can burn through tokens that cost more than ten casual seats. So vendors are experimenting with consumption pricing, which sounds rational until your bill triples in a month because someone ran a few extra queries. Then they pivot to outcome-based pricing, which is hard to measure and harder to enforce.

The churn this creates is a headache for buyers. You can’t budget when the pricing model changes every quarter. You can’t compare vendors when one charges per seat, another per API call, and a third wants a cut of your revenue. It’s not strategy, it’s whack-a-mole.

Flat retainers cut through that noise. When you know the cost upfront and it doesn’t move, you can plan. You can test. You can scale without watching the meter. That predictability is exactly what we build into the Omni Command Centre, where the monthly cost stays fixed whether you’re running ten workflows or a hundred. It’s not exotic, it’s just sane pricing for companies that need to move fast without surprises.

The pricing debate will keep going. But if you’re tired of the thrash, the answer is simpler than the podcasts make it sound.