EY's internal AI router cut token consumption by 60% in months as corporate spending fears grow. The system directs queries to cheaper models without changing the user experience.
EY has deployed an internal AI routing system that has already cut token consumption by up to 60% since its April rollout. The firm's global consulting AI leader Dan Diasio told Business Insider the router acts as an invisible intermediary, directing employee queries to the most cost-effective model rather than sending every prompt to the most powerful one.
"To the users, it makes no apparent change," Diasio said. Employees open the frontier model interface, enter their prompts, and receive responses. The router decides where the query actually goes behind the scenes. "Many times, it's faster than the frontier models, because to use the heavier models means there's more thinking that gets put in place, and it takes longer to be able to get back to the answer."
The router sits behind EY's specialized AI tools for tax and risk. It is not deployed on the general Microsoft Copilot chatbot available to all EY staff. The system has been paired with training and governance strategies, according to the firm.
The move comes as AI providers increasingly charge by token usage rather than flat-rate billing. Between February and June, OpenAI, Anthropic, and GitHub all introduced pricing models tied to tokens. EY's own AI Pulse survey, released Tuesday, found that 82% of senior leaders at companies investing in AI were concerned about token costs. The survey drew from 534 senior decision-makers at U.S. businesses surveyed between April and May.
"A lot of these big bills that companies are getting surprised by are by the top one or the top 2% of people inside the employee base that are just using the wrong tool for the job," Diasio said.
The first phase of the corporate AI push focused on getting employees to use the technology. Rising costs have shifted the emphasis to controlling that use. Disney and JPMorgan have installed dashboards to track employees' AI usage. Startups focused on AI routing are also proliferating, helping developers direct tasks to different models, monitor for overspending, and resolve outages.
EY has set token budgets based on employees' roles and departments. Those who exceed their allowance can request more through an approval process. Some 64% of senior leaders in EY's survey said their organizations now monitor AI token usage and have clear budget guardrails. Diasio expects that figure to rise quickly. "If we ask this question in six months from now, it would probably be like 90% of people," he said.
Diasio noted that other levers can help companies get more value from AI spending. Businesses often overlook the importance of building strong internal knowledge layers and designing better prompts grounded in their own data. He said firms should focus investment on areas where AI can have the deepest impact on how work is done, rather than spreading resources across every function.
Measuring AI's value requires looking beyond profit and loss, Diasio said. The broader questions businesses should ask include whether AI enables employees to get through tasks faster and how to incentivize workers to take on new initiatives outside their current job descriptions.
EY's own results suggest the router approach is gaining traction. The firm said the combination of routing, training, and governance has reduced token consumption by 60% since April.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.