Diving overview:
- Uber Technologies’ engineering team has overtaken finance as the company’s leading adopter of artificial intelligence.said a finance executive on Thursday, amid reports that ride-hailing companies are struggling with high costs from AI tools used to automate software coding.
- According to Tiho Nedkov, director of finance at Uber, more than 90% of Uber’s finance staff now use AI tools on a daily basis. This highlights how quickly teams are integrating technology into their daily workflows, even as engineering departments have risen to the top of the list in adoption.
- “Finance actually led this. [within] I used Uber for several years…at least until agent coding came along.” Nedkov The statement was made at Gartner’s annual conference for corporate finance leaders. National Harbor, Maryland.
Dive Insight:
The comments come amid reports of tension within Uber over AI spending, amid concerns about coding tools and token-based usage patterns that have led to higher-than-expected internal costs. This development highlights a broader trend known as “token maximization,” an industry term for the mass use of AI tokens.
Concerns within Uber over AI spending surfaced last month in a report in The Information, in which chief technology officer Praveen Nepali Naga revealed that the company had exhausted its 2026 AI coding budget within the first four months of this year as the company rapidly deployed tools like Claudecode across its engineering teams.
This issue was later echoed by chief operating officer Andrew McDonald in a Rapid Response podcast interview, saying it is difficult to show that increased usage of AI tokens is delivering a return on investment.
“That relationship has not yet been established,” he said, adding that while usage statistics are rising rapidly, it is difficult to draw a direct line between those metrics and improvements in consumer service.
According to Forbes, Uber’s cloud code deployment has accelerated rapidly across its engineering teams, with adoption rates rising from 32% in February to 84% in March. By spring, about 95% of Uber engineers were using AI tools on a monthly basis, and about 70% of code committed came from those systems, the report said. This resulted in average monthly costs of about $150 to $250 per engineer, or as much as $2,000 for heavy users.
Nedkov said AI adoption in finance remains strong despite engineering’s overall leadership.
Uber’s finance organization is driving AI-driven automation across various workflows, with technology now processing more than 96% of invoices with more than 95% accuracy and reducing human review time, according to a slide presentation used when Nedkov met with Gartner’s Mallory Berg Bulman. Contract reviews for revenue recognition were fully automated, and multilingual AI workflows reduced regulatory notification processing by approximately 70%.
He also pointed to efforts to build “data agents” used for financial intelligence that enable conversational analysis of corporate data, and “process sources of truth” that map financial workflows through AI-driven systems. Nedkov also said there is growing interest in exploring new agent AI use cases.
In an interview after speaking at the Gartner conference, Nedkov said Uber has “fundamentally democratized” access to AI tools, allowing employees to build and deploy applications without significant barriers. He said this approach had enabled “a lot of innovation and productivity” while also driving significant increases in production. recognized “Sources of truth” can pose challenges.
“But if you have guardrails, you can control it,” he said. “There’s kind of a playground where people are developing these things.”
Most of Uber’s finance professionals use AI, but the team is still in the “early stages” of AI implementation, he said.
“There is no such thing as token maximization in finance,” Nedkov said. “But you have to be mindful of that. You have to make sure there are limits in place.”
Nedkov declined to comment on reports that Uber’s 2026 budget for AI coding tools has already been exhausted. The company did not respond to requests for comment.
