It started with a single screenshot.
I was looking at our Claude for Organization billing dashboard. We have 48 seats. Between the base costs and the erratic, prorated invoices for new seats, we are looking at a burn of $3,600 per month. That is over $43,000 a year—and that’s if we stay exactly this size.
Beneath that total was a scrolling list of small, erratic invoices—$2.47 here, $36.03 there. It was the digital equivalent of “death by a thousand cuts.”
As the CEO of OpenLM, a company built on the mission of license transparency, I realized I was falling into the same trap I help our customers avoid. I was staring at a “Team Plan” and weighing a high-stakes choice: Do I stay monthly, or do I lock us into an annual contract?
The 20% temptation
The math seemed simple on the surface. Switching to the Annual plan would drop the cost from $25 per seat to $20. For a team of 48, that’s thousands in annual savings—a “no-brainer” for any finance department.
But as I dug deeper, the “All or Nothing” reality of SaaS billing hit home.
In the world of AI, you can’t just “mix and match” billing cycles. You can’t have ten power users on an annual plan and thirty-eight casual users on a monthly plan. You are either all-in, or you’re paying the monthly “flexibility tax.”
Additional Read: Why FinOps is the new standard for software asset management in 2026
The $40,000+ blindfold
When you factor in our anticipated growth and the need for Premium seats (which run $125/month for power users), we aren’t just talking about a few thousand dollars. We are looking at a **$40,000 to $50,000 annual commitment**.
Then there was the Enterprise question. The dashboard teased us with “Enterprise Features”: SCIM for automated onboarding, Audit Logs for security, and a 500K context window.
But Enterprise moves you to a “Pay-As-You-Go” model based on API rates. For a CEO, this is the ultimate visibility gap. If my team uses Claude heavily for a big project, does my bill suddenly double? Without data, Enterprise isn’t an upgrade; it’s a variable risk.
The “data-driven” pause
Most companies at this stage would simply “guess.” They’d look at the 20% discount, cross their fingers that all 48 people are actually using the tool, and click “Switch to Annual.”
But at OpenLM, we decided to take our own medicine.
Before committing to a massive annual bill, we are implementing our own monitoring. We’ve decided to pay the “flexibility tax” for one more month. Why? Because we need to see the “ghost users.”
Additional Read: Shelfware licenses: Identifying and reducing wasted software costs
The three types of AI users
Through the lens of OpenLM monitoring, we expect to see three distinct groups among our 48 seats:
- The power users: The developers and analysts who hit their limits daily. For them, a Premium Seat is a massive ROI.
- The standard users: The teams who use it for drafting and brainstorming. The Standard Annual Seat is perfect for them.
- The ghost users: The people who were invited during the initial hype but haven’t logged in for three weeks.
Stop guessing. Start monitoring.
The “all or nothing” billing structure of AI platforms is designed to capture the “ghost user” revenue. By switching to annual billing before you have usage data, you are essentially paying for empty chairs.
We are staying monthly for 30 more days. Once our monitoring reports come in, we won’t just be guessing which plan is “better.” We will know exactly how many seats to buy, who needs a Premium upgrade, and who can be offboarded.
Don’t fly blind with your AI budget. At OpenLM, we’ve developed the solutions to give you this exact visibility into your AI seat usage. If you are facing a similar dashboard and a $40,000 question mark, I invite you to join us and see exactly what your team is doing before you sign the contract.
Explore OpenLM Platform for AI optimization today.
In the age of AI, the most expensive seat isn’t the Premium one—it’s the one that nobody is sitting in.
Oren Gabay is the CEO of OpenLM, a global leader in software license management and usage monitoring.
