The new feature tracks AI usage across all departments in a company and shows spend, leverage gained, and efficiency scores by department and workflow.
Lanai, an enterprise AI accountability company, today announced Token Tuner, a new feature that helps enterprises understand where their AI spending is occurring, which workflows are driving results, and where low-cost models can reduce unnecessary token costs.
Tokenmaxxing is becoming an emerging problem in enterprise AI, with teams writing more tokens, using more models, and generating more AI activity. CFOs may receive an AI bill that is 30% higher than the previous month, but they still lack visibility into what caused that increase and what results were achieved. Token Tuner fills in the missing context for enterprises by mapping token spend to workflow, model selection, efficiency, and value created. This new feature connects each AI interaction to a measurable outcome and generates a productivity score based on how well each user adapts their token usage and model selection to the task. For example, an employee who uses Opus 4.7 to reply to emails may have a lower efficiency score than if they used a smaller model for that task.
“Tokenmaxxing is real, it’s expensive, and it’s expanding beyond just a few engineers and companies,” said Lexi Reese, co-founder and CEO of Lanai. “It’s a vanity metric that looks like a measure of efficiency or progress, but says nothing about net value. ‘Maximize Outcomes’ is the solution businesses need right now to see which workflows are actually increasing productivity, accelerating decision-making, and delivering measurable results. That’s exactly what Token Tuner does for businesses using AI at scale.”
Also read: AIThority interview with Rohit Agarwal, Founder and CEO of Portkey
Initial customer analysis shows significant differences in the value generated between workflows, with some users identifying $50,000 to $150,000 in waste per month in the first week from a large number of low-value workflows that could be run on lower-cost models with similar output quality. In the beta, one Lanai user delegated 4.2% of the total AI time across the organization and used only 0.7% of tokens. Their efficiency score was 6.0, indicating that they were matching tasks to the right model, while other companies were burning 10 times more tokens at half the efficiency.
“Enterprises are using AI across engineering, sales, marketing, finance, and operations, but not all use cases need to be treated the same,” said Mohit Mehta, chief product officer at Lanai. “Complex customer sentiment analysis workflows across Snowflake, Salesforce, and multiple MCPs can justify a premium model. Using the same expensive model for simple formatting, search, and email validation typically isn’t. Token Tuner helps leaders see the difference, so businesses can invest in workflows that create value and tune workflows that simply burn tokens.”
The main features are:
- Workflow-level value visibility: Shows which teams, workflows, and use cases are driving AI spending and whether that usage is leading to measurable business value.
- Measuring productivity and efficiency: Compare token spend and leverage gained by users, teams, and workflows to show where AI is creating the most value per dollar.
- Spending optimization recommendations. Identify runaway workflows, mismatched tasks, and use of premium models for work that can be handled by lower-cost models.
Also read: AI-powered risk intelligence: How financial institutions are anticipating systemic shocks
[To share your insights with us, please write to psen@itechseries.com]
