EY built an ‘AI router’ to stop its own AI bills from spiralling
The business move
EY has developed an “AI router,” a system designed to control rising AI expenses by directing each task to the cheapest model capable of handling it. This initiative responds to a growing issue in corporate IT where companies face ballooning AI token and usage costs as their AI workloads expand. By routing requests based on cost-effectiveness, EY aims to avoid overpaying for simpler queries that do not require the most advanced and expensive AI models.
Why it matters
AI usage in enterprises typically involves a range of models with varying capabilities and pricing. Without cost management, running all tasks on the most powerful models can lead to runaway bills. EY’s AI router addresses this by segmenting tasks and intelligently assigning them to less costly but still adequate AI services. This lowers the company’s financial exposure to AI calls and sets an example for how large firms can scale AI responsibly without sacrificing performance or inflating costs unnecessarily.
The move also pressures AI vendors to maintain flexible pricing since large clients like EY will optimize heavily for cost efficiency. It intensifies the need for clear cost transparency and promotes innovation in multi-model orchestration tools.
Who gains and who gets squeezed
Large enterprises and consultancies gain a blueprint for controlling the rapid cost growth associated with AI integrations. Builders and operators working inside these companies can apply similar routing logic to keep AI projects sustainable.
Conversely, AI vendors may face pricing pressure. Buyers deploying at scale will push back on high prices by adopting multi-model strategies that extract value without overusing premium options. Smaller companies less focused on cost may not see immediate benefit, but the rising demand for cost controls can force vendors to rethink pricing tiers overall.
What to watch next
Expect more firms to adopt multi-model routing to tame AI token costs as usage accelerates. Watch how AI vendors respond, possibly by enhancing APIs to support cost-aware routing or introducing product features that facilitate intelligent model selection.
EY’s experiment could also trigger a new wave of tooling focused on expense management for AI, including integrations with spending dashboards and automated cost-control policies. The broader AI market may see increased segmentation between premium and commodity AI offerings, driven by demand for flexible cost management.
AI Quick Briefs Editorial Desk
