Esker exceeded its AI budget by four times this year. CFO Scott McDermott said usage surged and providers shifted to usage-based pricing. Tokens are the basic units of data processed by AI. Esker now treats AI spending as part of each employee’s cost. It uses actual usage data to estimate consumption by function. Finance calculates a run-rate AI cost per employee monthly. The approach aims to make AI spending more predictable. Finance and R&D teams drive the highest AI costs. Esker is still measuring whether higher AI spending yields productivity or revenue gains. A company survey found 72% of finance leaders overspent on AI. 65% of CFOs had trouble tying AI usage to business outcomes. Agentic AI can increase token use per task by up to 100 times.
What changed
AI costs were treated as a separate line item on the profit and loss statement.
What it unlocks
Tracking AI token costs as a predictable employee expense alongside salary and bonuses.
- AI budget exceeded by 4x
- 72% of finance leaders overspent on AI
- 65% of CFOs struggle to link AI usage to outcomes
- agentic AI can multiply token use by 100x
Sources