Good morning. For companies figuring out how to use AI, the conversation is changing from improving existing work to rethinking what the business can do.

I spoke with John Felton, SVP and CFO of Amazon Web Services (AWS), Amazon’s cloud computing and AI platform, about that shift. Felton speaks regularly with AWS customers, including fellow CFOs, and said their AI adoption is still uneven. Many are still early in their broader cloud and IT modernization efforts, a prerequisite for deploying AI at scale.

“I think in general, I would say it feels early across the board,” Felton told me. “It’s unbelievable the pace of change, how fast things are moving. I do think every company is in a different stage of it. I think everyone is still early.” That puts him at an interesting vantage point as AI drives a significant new investment cycle.

Felton also pointed out that customer AI discussions about two years ago focused largely on productivity and cutting costs. That is changing.

“There really is a lot more focus on: how do I use AI for new revenue products, for new customer experiences?” he said.

That perspective carries particular weight given the scale of the business Felton oversees. AWS is now a $169 billion annualized business, and Felton has been with Amazon for more than 20 years. He previously led Amazon’s worldwide operations before becoming AWS finance chief in 2024, reporting to Amazon CFO Brian Olsavsky.

The bigger shift, Felton said, is from adding AI to an existing process to rethinking the process. What impresses him is seeing companies move toward an “AI first mindset,” using AI to rethink what the business should be in the first place.

“That is where a bunch of customers, including ourselves, really see a lot more value,” Felton said.

AI in finance

Inside AWS’s finance organization, Felton is encouraging employees to participate in that transformation rather than wait for someone else to tell them how to use AI.

His instruction is straightforward: “Use AI every day.”

Rather than dictate which tasks employees should tackle or which tools they should use, Felton wants them to identify opportunities themselves. They understand their work better than he does, he says. He tracks whether employees are following through, although he did not detail those metrics.

One example came from AWS’s pricing team. Employees had relied on a complicated Excel model spanning 18 tabs to evaluate customer deals. The team converted it into a chatbot interface that lets users explore scenarios through natural-language questions.

They can ask what happens if a price falls 20%, how different payment terms affect a deal, or where the break-even point lies. Employees who prefer the spreadsheet can still export the model to Excel. The point isn’t simply that AI made an existing tool easier to use. It changed how employees interact with the underlying analysis.

Another team built an AI agent to compare payment terms in customer contracts with those recorded in AWS’s payment system. Previously, employees checked a sample of contracts. The agent allows the team to check all of them.

“Let’s not just do a sample size, let’s do all of it,” Felton said, describing the team’s approach.

The two examples illustrate the broader shift he sees in companies’ AI strategies: using the technology not only to make existing work faster, but to expand what the organization can do.

Felton also uses Amazon Quick, an AI assistant for work, himself, including to query supporting materials prepared for board meetings and locate answers in the underlying files. He sees that kind of everyday experimentation as important for CFOs figuring out where AI fits.

Sheryl Estrada
Sheryl.Estrada@fortune.com

Leaderboard

David Vaillant was appointed managing director and CFO of the World Bank Group. Vaillant spent two decades at BNP Paribas, most recently as global head of finance, strategy and participations for its asset management business. Previously, he was managing director and head of banking for Europe, the Middle East and Africa at BNP Paribas Corporate and Institutional Banking.

Jonathan I. Lieber was appointed CFO of Bambusa Therapeutics, a clinical-stage biotechnology company. Lieber joins from Rallybio, where he had served as CFO since February 2023. Previously, he was CFO of Applied Genetic Technologies Corporation and a managing director at Danforth Advisors. He also held CFO roles at Histogenics, Metamark Genetics, Repligen, Xcellerex, and Altus Pharmaceuticals.

Big Deal

Bank of America is expanding its fraud-prevention education for small business owners as reported U.S. internet crime losses reached $20.8 billion in 2025, up 26% from a year earlier, according to FBI figures.

Business email compromise schemes accounted for more than $3 billion in reported losses, while phishing and spoofing remained the most frequently reported cybercrimes. The 2025 BofA Business Owner Report found that 30% of business owners plan to increase cybersecurity measures as part of their digital investments over the next five years.

BofA’s initiative includes an updated fraud-prevention checklist, educational resources and a webinar planned for later this fall on social engineering and business email compromise. Sharon Miller, president of Business Banking, said many scams succeed “not because a business lacks technology, but because criminals exploit trust, urgency and human behavior.”

Going deeper

AI governance is a growing priority for audit committees, but 82% of respondents describe their oversight as emerging or limited, according to the fifth edition of the Audit Committee Practices Report, a joint initiative of Deloitte’s Center for Board Effectiveness and the Center for Audit Quality.

Drawing on nearly 250 responses from audit committee chairs and members, the report found that 70% identify technology, including AI, as the skill most needed to improve committee effectiveness. Cybersecurity was the most widely cited top-three priority, at 87%, while 54% said their committees had become more focused on emerging risks over the past year. The report’s findings show that AI governance is a top priority, but readiness is still developing.