The subscription economy is built on a lie: that software value is proportional to the number of humans logging in.
That math is breaking.
I reviewed a contract last week where a vendor tried to charge $300/month for an "AI seat." It was just a wrapper around GPT-4. They know the UI is dying. They are scrambling to tax the API.
We used to buy software to give humans better buttons. Now we build software to push the buttons for us.
The Rent vs. Own Trap
Seat-based pricing punishes efficiency. If one agent does the work of 10 junior analysts, the vendor loses 9 seats of revenue. They won't let that happen.
So they pivot. They introduce "outcome pricing" or "credits."
This is the trap. You stop renting the tool and start renting the labor.
Consider a mid-market logistics firm I spoke with recently. They pay $40k/year for a Transportation Management System (TMS). Their 15 dispatchers spend six hours a day copying load data from emails into the TMS forms.
The vendor pitched an "AI Automation Add-on" for another $20k/year. We looked at the spec. It was a fragile wrapper that locked them into the vendor's clumsy workflow.
Instead, we built a local agent for a one-time cost of $5k. It reads the emails and hits the TMS API directly.
The dispatchers now spend one hour a day verifying data. The SaaS vendor gets their base fee for the database, but they don't get the "AI tax."
We own the worker. We rent the database.
What to Keep (The Database)
Don't build your own CRM. Don't build your own ERP.
These are systems of record. You pay Salesforce or NetSuite to keep your data safe, compliant, and accessible. That is worth the rent.
But the workflow—the actual moving of data—belongs to you.
If a tool owns your core records, keep it. But change how you buy it.
The Renewal Fight
When your account rep calls this quarter, they will pitch "AI features."
Ignore the features. Look at the data rights.
Can you export everything via API? Is there a rate limit? If they throttle your API access, they are trying to force you to use their AI interface.
Don't sign.
The only metric that matters is whether you can build your own automation on top of their data layer. If the answer is no, the software is dead weight.
Sources
- IDC, 2025: https://www.idc.com/resource-center/blog/is-saas-dead-rethinking-the-future-of-software-in-the-age-of-ai/
- Intercom, 2025: https://www.intercom.com/blog/customer-service-transformation-report-2025/
- Atlassian Rovo billing, 2025: https://support.atlassian.com/rovo/docs/rovo-dev-rollout/
- TechCrunch, 2026: https://techcrunch.com/2026/02/24/openai-coo-says-we-have-not-yet-really-seen-ai-penetrate-enterprise-business-processes/
- Notion AI data practices: https://www.notion.com/pricing
I reply to all emails if you want to chat: