A company can show you an AI tool that sorts a messy inbox in seconds. Getting a customer to budget for it again next year is the more interesting trick. But how often does that second part happen?

Scource: Getty Images
Find the part customers will pay for
Businesses have years of files spread across different systems. A tool that helps an employee find the right document can save time. To do more than search, software also has to respect who can see the file and who’s allowed to act on it. That’s a bigger decision than adding another search box.
When I spoke with Muhi Majzoub at OpenText (TSX: OTEX) World in October, he described an “AI ladder” that lets customers start with a use they’re comfortable with and go further when they’re ready.
Majzoub talked about giving customers tools to build agents that can work together on a process. The steps are a way to start small without asking a customer to overhaul everything at once.
I’d ask how many customers are willing to put a tool into a process people use every day. A pilot can show a product works on one set of files. A renewal tells me the customer saw enough benefit to keep paying for it after the novelty wore off.
For shareholders, I’d follow the money from one step to the next. If a customer starts with search and later pays for automation, the company can earn more from a relationship it already has. If customers only use features included in existing contracts, the software may be helpful while the earnings barely move.
OTEX
OpenText sells software for managing information, documents and business processes. That puts it near the files and permissions customers already rely on. There’s a reason to consider adding AI there, especially when customers want help working with records they can’t casually send to another system.
In its fiscal first quarter, OpenText reported cloud revenue of US$505 million, up 4.3%. Total revenue fell 4% to about US$1.24 billion. The cloud number shows part of the business is growing. It doesn’t tell me what customers paid for AI separately, and the rest of the business still has to earn its keep.
I’d also watch the US$139 million in company-defined quarterly free cash flow. A paid upgrade can add sales and still cost money to implement and support. Over several quarters, I’d like to see more cash, alongside better revenue trends, before I give the new products much credit in my valuation.
Considerations
OpenText closed at $34.86 at writing. At an illustrative price of 12 times annual free cash flow, the shares would need about $2.91 in sustainable annual free cash flow per share. I’d convert the company’s U.S. figures into Canadian dollars and use a full year, since one quarter can flatter or understate the business.
The next evidence I’d like is a figure I can track. Management could disclose how many customers move from testing a feature to paying for broader use, how often AI is sold alongside a renewal, or whether those customers spend more on over time. I’d then compare that with cash after supporting those accounts.
I’d put more weight on a change in those numbers than a list of new features. A wider release can take time to sell, and some customers will go at their own pace. I’m looking for evidence that the ones taking another step are becoming better customers for the business.
Competition and a slower move away from older products can limit what OpenText earns. If I add it inside a Tax-Free Savings Account (TFSA), I’d size it alongside my other Canadian stocks so the AI bet has time to earn its place.
Bottom line
I’d hold off on paying an AI premium for OpenText. The opportunity I’d like to buy is a customer using it to do more work, renewing at a higher value and leaving the company with more cash when the year is done.