- BIST10012,151-0.94%
- USD49.220.00%
- EUR55.21+0.28%
- GOLD4,195.90+0.13%
- INTEREST37.00%-1.00
- BITCOIN82,230-1.72%
- October 13, 202618°
- BIST10012,151-0.94%
- USD49.220.00%
- EUR55.21+0.28%
- GOLD4,195.90+0.13%
- INTEREST37.00%-1.00
- BITCOIN82,230-1.72%
- October 13, 202618°
Companies Shift Focus from Token Counts to Outcomes in AI Usage
Measuring AI expenses by token count is losing ground in companies. Executives speaking at a summit in Paris emphasized that the real metrics should be efficiency, customer service, and business value.
Over the past year in Silicon Valley, AI usage has almost turned into a scoreboard. However, executives speaking at the Mistral AI summit in Paris explained that the real issue in AI is not how many tokens are consumed, but how much this reflects on business outcomes.
Charles Holive, Chief AI Officer at BNP Paribas CIB, said his team tries to avoid vanity metrics. Holive emphasized that figures like billions of tokens per day alone have no meaning; what should be tracked is the outcome.
Holive explained that instead of asking employees how many tokens they used, they are more interested in what tasks they accomplished that they couldn’t before and how quickly they did so. This approach highlights the productivity generated rather than the intensity of AI usage.
Antoine Pichot, Innovation, Digital, and Data Director at La Banque Postale, said they evaluate AI based on employee productivity, customer service, and cost-benefit balance. Pichot’s assessment shows that banks view AI not just as a technology investment but as a business tool that must be measured.
Amit Kapur, Chief AI and Transformation Officer at Tata Consultancy Services, also stated that his focus is not on token consumption but on whether business performance has improved. Sujay Bhattacharya, an NTT DATA executive who helps enterprise clients adopt tools from multiple AI providers, noted that customers now look beyond token counts to total cost and business value.
Changing Approach in the "Tokenmaxxing" Era
These views at the summit coincide with a period when some U.S. companies are moving away from the "tokenmaxxing" mindset. This approach assumed that more AI usage and higher token consumption automatically meant greater productivity.
Amazon shut down its internal AI usage leaderboard after employees began completing tasks with AI solely to accumulate points and climb the rankings. Uber’s Chief Operating Officer Andrew Macdonald also questioned whether higher AI spending truly results in more useful products.
The shift by OpenAI, Anthropic, and GitHub toward usage-based pricing for enterprises has increased pressure on companies. This model requires customers’ employees to demonstrate whether increased AI use genuinely creates value.
At BNP Paribas CIB, token consumption has not been completely abandoned. Holive said they continue to monitor token usage to control costs and track adoption levels.
Still, the common message from executives in Paris was clear: Token counts can indicate how much AI is used, but they alone do not reveal whether it truly generates return on investment.
Stay Updated with The Turkiye Times
Subscribe to be informed about special news, developments and announcements. We will keep you informed.
By signing up, you agree to our Privacy Policy
