AI firms revise pricing to capture more business spending

Software companies are evolving their pricing structures for AI tools, a change that has the potential to transform business spending.

Rather than basing charges on the number of users, some AI firms are shifting to pricing their products based on the tasks they perform, according to a report from Goldman Sachs released on Wednesday.

"Companies are starting to sell their AI workflows as offering labor or productivity units, enabling them to secure larger deals and explore new financial avenues," reported Goldman Sachs analysts after discussions with approximately 40 companies in the software and internet sectors recently.

This strategy allows firms to charge based on the value added, decoupling earnings from AI operational costs, thus preserving healthy profit margins, they noted.

The practice is becoming evident in how companies configure their offerings. Salesforce has rolled out "agentic work units," whereas Workday offers credits associated with "work units."

As the substantial costs associated with developing and operating AI are encouraging businesses to reevaluate their pricing models, software expenses for clients using these services could become less consistent.

Last month, OpenAI CEO Sam Altman mentioned that AI could eventually be marketed like a utility.

"Ultimately, our business — which I believe will be the model for all providers — will resemble selling tokens," he mentioned, referring to the units utilized by AI systems to process and evaluate data.

"We envision a future where intelligence is a utility, much like electricity or water, and consumers purchase it from us for various needs," Altman added.

Overall, the software industry is shifting from monthly "per seat" licenses towards usage-based, pay-as-you-go pricing, as reported by Business Insider's Alistair Barr in April of the previous year.