The Token Economy: Analysis of Glean’s token costs and cost efficiency
Tuesday, October 6, 2026
11:00am - 11:30am PT
Speakers
Overview
The most expensive frontier model is not always the best model for the job.
As enterprises move beyond simple chat to coding, research, analysis, and workflow automation, AI token consumption and costs can rise quickly. The right architecture can help organizations do more with AI by combining model choice, intelligent auto-routing, enterprise context, and an efficient harness.
Glean’s recent Pareto frontier analysis found that its responses were preferred 78% of the time over Claude Cowork across 180+ enterprise tasks, while average cost per task was $0.58 compared with $2.98.
In this webinar, we’ll unpack the full findings from Glean’s analysis and hear from Chinmay Shepal, Senior Director of Business Systems, about Icertis’ experience reducing token consumption with Glean.
What you’ll learn
- How model choice, reasoning level, and enterprise context shape the cost-quality tradeoff
- How auto-routing and harness design improve AI efficiency and performance
- A practical framework for evaluating AI systems based on outcomes and cost
