Amazon's financing teams rely more on AI - and not just for the simple things

Copyright © HT Digital Streams Limit all rights reserved. Mark Maurer, The Wall Street Journal 6 min Read 18 Sept 2025, 07:29 IST IST Amazon made a hard pressure in AI, investing billions of rands in the boot Anthropic and planning to spend more than $ 100 billion on data centers. Photo: Toru Hanai/Bloomberg News Summary Agentic AI and other technologies increase productivity – and help avoid potholes such as hallucinations, top managers say. Amazon has made AI a hard pressure on AI, investing billions of rands in the beginning of anthropically and plans to spend more than $ 100 billion on data centers. Amazon.com’s financing teams use generative AI in new ways to meet complicated tax regulations more efficiently, help evaluate the financial impact of new products and to better analyze revenue. It is part of an effort by Amazon to further apply AI to tackle more complicated financing functions, which goes beyond the automation of a rote office processes that take place in many businesses. Amazon has expanded the financing areas where it has been using generative artificial intelligence over the past year, managers said. About a year ago, the company applied Genai to core functions such as fraud tracking, account reconciliation, financing forecast and contract assessments. The moves come because the e-commerce giant has made a hard pressure on AI, invested billions of rands in the anthropic startup and was planning to spend more than $ 100 billion on data centers. Its investments in technical infrastructure are driving cash capital expenditure, which amounts to $ 31.4 billion in the latest quarter, from $ 16.4 billion a year earlier. CEO Andy Jassy said in June that AI will result in a smaller workforce in the next few years. Amazon uses generative AI for financing, ranging from report writing and summary of documents to synthesizing data sources and reforming how the company generates financial insights, CFO Brian Olsavsky said in an interview. “Generative AI transforms our financing function at Amazon,” Olsavsky said. “It makes our processes faster and more efficient, which helps us identify trends and opportunities sooner and enable our teams to focus more on strategic thinking.” These use cases expanded as the technology, especially agentic AI, progressed. Tax time for one, generative AI helps Amazon’s financing team to improve its efficiency and consistency in complying with the rules of the Internal Revenue Service on tax issues ranging from transfer prices to research and development and patents. The team has traditionally put together a lot of data sources and tried to solve “incredibly complicated problems”, “said John Felton, Financial Head of Financial Head of Amazon Web Services, the company’s cloud computer unit. The company then began to provide help with the interpretation of rules and regulations and tax -related work. Much of the work is being done increasingly automatically and faster, while still being helped by man’s supervision, he said. Companies analyze their transfer prices, or internal corporate transactions across the boundaries, to make sure they are comparable to market prices. Amazon uses generative AI to analyze more than 600,000 businesses as part of the benchmark analysis. The AI ​​instrument can reduce the time to complete such an analysis with at least half, Felton said, referring to the initial testing. Revenue Dive Amazon’s financing staff uses a generative AI chatbot to review and analyze revenue and customer data. For example, financing leaders use natural language to create dashboards and inquire that shows revenue on a weekly and monthly basis over time, Felton said. They can ask the agent why the revenue rose last week, or why a customer recently changed or stopped purchasing or starting to buy or start buying. AWS employees use the tool daily to generate internal financing reports, Felton said. “We started playing with it a year ago, but today it really works,” he said. Felton said he would have to investigate such summaries a year ago, because hallucinations, which are misleading or wrong answers to quick answers, occur frequently. He is still looking at the results, but the effort required with it has “come across”. Amazon said its financing team is still experimenting with the use of AI agents to approve employee purchases of things such as digital advertising, opportunities and software licenses from outside sellers. Many companies apply more generative AI to their financing work after almost three years of experimentation. Large corporations such as Amazon have more freedom to assign resources to their AI efforts and move faster than smaller businesses, some of which are stuck with financing tasks in a pre-SAA landscape, says Mark D. McDonald, an independent consultant and former senior director at the research and consulting firm Gartner. The businesses at the forefront of AI in Finance “must focus on non-standard office that is more adapted to their business and that they almost have to build themselves because there is no solution,” McDonald said. Product finances Amazon agents also help to estimate the return on investment and possible prices of a new Amazon product that is about to. The financing and product management teams, among others, have struggled for decades with the preparation of a product document, known as PR/FAQ, sometimes over weeks. The document contains the business case and customer experience for a new product. Software teams then partly built the product on the document. Felton said it is a strict process to write the PR/FAQ. It is still strict, but AI drugs are very involved. The teams enter some directions and receive a near-finished product. AWS offers more than 240 cloud products. “I can definitely see a world where a press release with people is repeated, but very quickly, and then it is handed over to software agents to build the product, and you can go to the product very quickly, very quickly,” Felton said. “It’s not science fiction in my mind. I think we’re near the world. ‘ Felton said he was excited to communicate some witness of 15 or more AI agents with each other to give him insights and to complete key financing tasks. “There are a lot of things I do today that I didn’t think I would do,” Felton said, referring partly to the use of agents. AI is already making the Amazon workforce more productive, Felton said, adding that it is too early to say whether it will cause a net reduction in jobs and fall to comment on possible cost savings. Jassy said last year that the company had produced about $ 260 million in annual efficiency profit due to generative AI. Felton said the biggest limitation of his work is time. “I just don’t have enough hours in the day to do all the things I want to do,” he said. “A lot of what Genai does is that it gets rid of the muck, maybe the not-so-fun parts of your work.” Amazon managers said AI changed what finance talent looks like. “In addition to strong traditional technical skills, our financing leaders need to think creatively, quickly adapt and transform AI-generated information into meaningful business insights,” Olsavsky said. Eighteen percent of the CFOs eliminated financing opportunities due to the implementation of AI, with the majority of those saying that accounting and controller roles were cut, according to a survey released by executive search firm Egon Zehnder this month. Almost every financing professor at a business will need to become more analytical and strategic, more similar to business advisors, as AI justifies less manual intervention, says Ash Mehta, a senior director of Gartner. Mehta said: “The skills that have made financing staff successful in the past have not made them successful in the future,” Mehta said. Write to Mark Maurer at [email protected] fasting all the corporate news and updates on Live Mint. Download the Mint News app to get daily market updates and live business news. More topics #amazon read next story