Sabio AI chief says the token-maxing era is over

Sabio Group’s chief AI officer says businesses are entering a more disciplined phase of AI spending as subsidies fade and companies confront the real cost of usage. The shift could push firms to focus on training, task-level automation and clearer ROI instead of chasing raw consumption.

Why it matters: - The end of AI pricing subsidies is forcing companies to reconsider how much they use AI and whether the returns justify the spend. - The shift could reduce wasteful experimentation and push enterprises toward narrower, higher-value use cases. - Sabio Group says customer service is one of the clearest places to measure AI payback in time saved or interactions removed.

What happened: - Stu Dorman, chief AI officer at Sabio Group, said the industry has moved past the “token-maxing” phase on City AM’s Business As Usual podcast. - Dorman said firms were once encouraged to use as much AI as possible, but that era is now over. - He said the change was driven by the withdrawal of price subsidies in late May and early June, not by the recent drop in tech shares. - Dorman described tokens as the usage units that determine what companies pay to run AI models.

The details: - Dorman compared token usage to a pay-as-you-go mobile phone plan, where heavy use quickly burns through credit. - He said many AI vendors had previously subsidized usage, which hid the real cost of consumption. - Once subsidies fell away, businesses could see which use cases justified the spend and which models fit each task. - Dorman said some firms even rewarded employees for maximizing AI usage, which led to high spending and little visible value. - He said the better path is enablement and training, so employees can identify specific tasks in their jobs that AI can support. - Dorman said, “AI doesn’t automate jobs; it automates tasks.” - He pointed to customer service and contact centers as Sabio’s core market and a place where AI return on investment can be measured more clearly.

Between the lines: - The comments reflect a wider reality check across enterprise AI, where easy access and subsidized usage have given way to harder budget scrutiny. - Dorman’s view suggests the next phase of AI adoption will reward precision, not volume. - He also linked the market pullback to inflated expectations, citing predictions that all jobs would disappear or work would become optional. - He said fresh research showing enterprise adoption lagging the hype has added to the pressure. - Dorman said new Chinese AI models are undercutting American rivals on price. - He noted that AI investors have faced weeks of falling share prices, including a semiconductor fund down almost 10% over the past month.

What’s next: - Dorman expects AI adoption to accelerate further, but not at the pace major technology firms want. - He cited a Gartner forecast that AI will become a net creator of jobs by 2029. - Dorman said the technology will drive productivity and job creation over time. - Sabio is positioning its AI-first customer experience approach around practical deployment rather than broad consumption. - More information is available on Sabio's website.

The bottom line: - AI spending is shifting from “use more” to “prove value,” and companies that cannot show returns may face tougher scrutiny ahead.