AI cost guides Westpac to use ‘sensible’ models for production staff

Applications of AI


(Bloomberg) — Westpac Bank is stepping up efforts to monitor artificial intelligence costs, with the Australian lender closely tracking AI tokens across the company and redirecting simpler tasks to cheaper models.

Most Read Articles on Bloomberg

As part of this push, the bank is tracking the consumption of AI tokens by its employees, including its software development team, which has access to cutting-edge and expensive models, according to new chief AI officer Dan Jarmin.

“We look at how much people are spending through AI tokens and are focused on routing work to the right model for cost and efficiency while delivering the best results,” said Jermyn, who joined Westpac this year from major rival Commonwealth Bank of Australia. The bank said it wants staff to “understand the intricacies of how different models work” and “make wise choices about the models and tools they use to get the most benefit.”

His comments come as executives around the world are increasingly concerned about the cost of implementing AI. The focus on expenses reflects a shift in companies deploying AI to improve productivity, but now closely analyzing usage to ensure AI is delivering value, a move that could weigh on the bottom line for providers like OpenAI and Anthropic PBC.

The review of Westpac’s AI costs comes within the bank’s UNITE initiative to centralize disparate systems from past acquisitions, and Jarmin said the effort could ultimately improve how the bank uses AI in its operations.

Australia’s second-largest bank by assets saw its total spending on technology rise 13% to A$3.1 billion ($2.2 billion) in its most recent financial year to September 30, highlighting the growing costs it faces in this area.

Westpac said in its latest annual report that more than 15,000 staff, half of its workforce, are using generative AI to support the way they work. The bank said it is investing in modernizing its systems to better utilize technology.

“Competitive advantage is not just about having access to one model, but how you leverage the efficiencies of AI across your bank to drive the best outcomes,” said Jarmin. “The speed of change with modern AI tools is dramatic, but we need to find a balance between empowering people and controlling costs.”



Source link