Cognizant Technology Solutions is actively implementing a system known as 'token metering' to track each associate's personal artificial intelligence fluency. The IT services firm is increasingly applying this approach at a project or individual level to provide early insights into usage patterns, model management, and optimisation of inference costs.
What Is Token Metering?
Token metering refers to the measurement of tokens consumed during interactions with AI models. Tokens represent units of text processed by large language models, such as words or subwords. By tracking token usage, Cognizant can monitor how associates engage with AI tools, identify high-usage areas, and manage computational resources efficiently.
Benefits for the Organisation
The implementation offers several advantages:
- Early Insights: Real-time data on usage patterns helps Cognizant understand which teams or projects are leveraging AI most effectively.
- Model Management: Tracking token consumption aids in selecting the right models for specific tasks, balancing performance with cost.
- Cost Optimisation: By analysing inference costs per token, the company can optimise spending on AI infrastructure.
Impact on Associates
For employees, token metering provides a measure of their AI fluency, encouraging them to adopt AI tools more strategically. This initiative aligns with Cognizant's broader digital transformation strategy, aiming to upskill its workforce in AI technologies.
The move comes as IT services companies increasingly seek to integrate AI into their operations while maintaining cost control. Cognizant's approach reflects a growing trend of granular tracking in enterprise AI adoption.
