The first phase of enterprise artificial intelligence adoption was defined by experimentation, rapid spending, and a willingness to consume AI resources without fully measuring returns. That period may now be ending. According to Sabio Group Chief AI Officer Stu Dorman, businesses are moving beyond the “token-maxing era” as AI pricing becomes more transparent and organizations begin evaluating whether their AI investments deliver measurable business value.

Enterprise AI Moves From Unlimited Experimentation to ROI Discipline

The artificial intelligence market is entering a more mature stage where efficiency, measurable outcomes, and cost control are becoming as important as innovation.

For much of the recent AI boom, companies rushed to integrate large language models (LLMs) into workflows, often prioritizing experimentation over financial discipline. But as AI infrastructure costs become clearer, enterprises are beginning to question whether every AI interaction creates meaningful value.

Stu Dorman, Chief AI Officer at AI-powered customer experience company Sabio Group, believes the industry has moved beyond what he described as the “token-maxing era” — a period when organizations were encouraged to maximize AI usage because pricing incentives made heavy consumption appear inexpensive.

Speaking on City AM’s Business As Usual podcast, Dorman argued that AI adoption is now entering a more practical phase focused on selecting the right models, controlling costs, and identifying high-impact use cases.

“AI doesn’t automate jobs; it automates tasks,” Dorman said, highlighting a shift away from broad automation expectations toward targeted productivity improvements.

AI Token Costs Force Companies to Rethink Usage

Generative AI platforms typically charge customers based on tokens, the units used to process input and generate responses.

The model is similar to pay-as-you-go services: organizations pay according to how much AI processing they consume. During the early expansion of generative AI, aggressive pricing strategies and vendor subsidies encouraged companies to experiment widely.

However, as those incentives declined, enterprises began seeing the actual economics behind AI deployment.

The result has been a growing focus on questions such as:

• Which AI workflows create measurable value?
• Which models provide the right balance of performance and cost?
• Where should companies deploy premium AI systems versus smaller models?
• Which processes can realistically benefit from automation?

This transition mirrors broader enterprise technology cycles, where initial experimentation is eventually followed by optimization and governance.

Companies including Microsoft, Google, Amazon Web Services, and OpenAI are increasingly focusing on enterprise AI platforms that combine performance with cost management.

AI Market Expectations Face a Reality Check

The shift in AI spending comes during a period of increased investor scrutiny.

After rapid growth driven by enthusiasm around generative AI, technology markets have become more cautious about whether AI investments can justify their scale.

Dorman argued that expectations had become unrealistic, with some predictions suggesting AI would eliminate large portions of employment or fundamentally change the nature of work almost immediately.

Enterprise adoption data has added to the debate, showing that while companies are investing heavily in AI, many are still working to move projects from experimentation into production.

Competition from lower-cost AI models has also changed market dynamics. Newer models from Chinese AI developers have increased pressure on established providers by offering alternative approaches to performance and pricing.

The result is a more competitive AI landscape where efficiency may become as important as raw capability.

Enterprises Shift From AI Consumption to AI Enablement

One of the biggest challenges facing organizations is not access to AI technology but understanding how employees should use it effectively.

Dorman argued that many early AI initiatives focused too heavily on increasing usage rather than improving outcomes.

Some organizations encouraged employees to maximize AI consumption without clearly defining business objectives, leading to higher costs without proportional productivity gains.

The next phase of enterprise AI adoption will likely emphasize:

• Employee training
• Workflow redesign
• AI governance
• Task-level automation
• Measurable performance improvements

This approach aligns with broader research from organizations such as Gartner and McKinsey & Company, which have emphasized that successful AI adoption depends on organizational change as much as technology deployment.

Customer Service Becomes a Test Case for AI ROI

Dorman identified customer service and contact centers as areas where AI value can be measured more directly.

For companies operating large customer support environments, AI can influence metrics such as:

• Average handling time
• Resolution speed
• Agent productivity
• Self-service adoption
• Customer satisfaction

AI-powered contact center platforms are increasingly combining conversational AI, automation, analytics, and agent assistance to improve customer interactions.

The contact center market has become one of the clearest examples of where AI can move from experimentation to operational impact because businesses can directly track improvements.

AI Adoption Continues Despite Market Reset

Although the industry is moving away from unlimited AI experimentation, Dorman remains optimistic about the long-term impact of artificial intelligence.

He predicted that AI adoption will accelerate, but at a more measured pace than some technology companies initially expected.

The next phase of AI growth is likely to focus less on how much AI companies use and more on how effectively they apply it.

Rather than replacing entire roles, AI is increasingly being positioned as a tool for improving individual tasks, enhancing decision-making, and increasing productivity.

As organizations move into this more disciplined stage of adoption, the winners in enterprise AI may not be those that consume the most tokens — but those that understand where AI creates real business value.

Market Landscape

Enterprise AI is entering an optimization phase shaped by:

• AI cost management: Businesses are evaluating token usage and model economics.
• AI governance: Enterprises need controls around security, compliance, and spending.
• Agentic workflows: AI systems are moving toward task execution rather than simple responses.
• Industry-specific AI applications: Customer service, healthcare, finance, and operations are becoming major adoption areas.

The market is shifting from AI experimentation toward measurable productivity and operational transformation.

Top Insights

• Enterprise AI spending is moving from unlimited experimentation toward cost-controlled, value-driven adoption strategies.
• Token-based AI pricing is forcing companies to evaluate which workflows justify model consumption.
• Sabio Group highlights customer service automation as a measurable AI business value opportunity.
• Organizations are prioritizing employee enablement and task automation over broad job replacement narratives.
• AI market competition is increasingly focused on efficiency, pricing, and practical enterprise outcomes.

Power Tomorrow’s Intelligence — Build It with TechEdgeAI