Tokenmaxxed out? The hidden costs of AI in finance

Interaction with AI tools require consumption of ‘tokens’ and the more complex the work, the more you use. Rob Steele, CFO, iplicit explains the risks of cost drift and talks to Forvis Mazars partner Robbie White about how the firm is using AI

If you want to make your ordinary daily activity sound like some sort of Homeric quest, you can try applying the suffix ‘maxxing’. We’ve had looksmaxxing and sleepmaxxing, and in finance, one of the buzzwords of 2026 is tokenmaxxing.

The practice of tokenmaxxing is encouraged by some tech companies when it comes to using artificial intelligence. Every time you interact with an AI tool, you’re consuming ‘tokens’ - and the more complex the work, the more of them you use.

Meta reportedly encouraged tokenmaxxing by ranking its employees on a leaderboard. The employee named ‘token legend’ burned through 281 billion in 30 days. The company shut down the board after its existence was made public.

S

Your free features:

  • Breaking news and expert analysis
  • Customisable daily newsletters
  • Six free CPD learning modules each year
  • Personalised CPD tracker
  • Top 75 Firms league tables
  • Regulatory changes
  • Hardman’s Tax Data

Sign up to Business & Accountancy Daily

Related Articles
Subscribe