Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
Your business deserves an address
Start for free, no credit card required. Get a building of your own with your name, logo and a link to your website.
A Startup Fortune initiative
You are invited to be on the Metropolix. Your invitation expires in 60:00.
Start for free, no credit card required. Your name, logo and a link to your website on a building of your own.
AWS pricing analysts once worked a 18-tab Excel workbook by hand; now a chatbot answers questions on discounts and break-even points, while the original spreadsheet stays one export away for every deal.
The spreadsheet that decides how much a company pays for cloud computing just got a conversational front end, and Amazon’s cloud chief says that’s the real story behind AI adoption right now.
AWS pricing analysts used to evaluate customer discounts, payment terms and break-even points by opening an Excel workbook with 18 separate tabs. AWS has now wrapped that model in a natural-language chatbot, according to Fortune, which first reported the change on October 8. Employees can ask it questions like what happens to a deal if the price drops 20%, or where the break-even point sits if a customer wants to stretch payment terms over three years instead of one.
AWS CFO John Felton described the project to Fortune as a signal of where AI inside the company is actually heading. Two years ago, he said, customer conversations about AI centered on productivity and cost-cutting: write code faster, summarize documents, automate the obvious busywork. Now those same conversations are shifting toward something riskier and more interesting, using AI to redesign the workflows that decide how money moves, what customers see, and what they pay.
That shift matters because pricing is not a low-stakes corner of a cloud business. AWS posted $42.2 billion in second-quarter revenue, up 36.7% year over year, with an AI revenue run rate north of $25 billion growing at a triple-digit clip. Discounts and payment terms on deals at that scale are not rounding errors. They decide whether a contract is profitable. Letting a chatbot sit between an analyst and that calculation is a different kind of bet than letting one summarize a meeting.
Felton was careful to frame this as a conversational layer, not a replacement. The underlying Excel model still exists, and anyone who wants to check the chatbot’s math can export it back into the original 18-tab workbook. AWS isn’t asking a language model to invent pricing logic from scratch or make a final call on a discount. It’s asking the model to translate a plain-English question into the existing, auditable spreadsheet and hand back an answer a human can still trace line by line.
STMicroelectronics raised its 2026 data center revenue outlook to approximately $1 billion on June 2, nearly doubling prior guidance, sending shares to a 25-year high. Fueled by a strategic multi-billion-dollar AWS partnership and surging demand for silicon photonics and power chips, the upgrade signals that AI infrastructure spending is flowing… – STMicroelectronics data center revenue growth AI demand – semiconductor makers benefiting from AI infrastructure boom
That distinction is the whole point. It’s also the thing easiest to miss in a headline about a hyperscaler letting AI near enterprise pricing. The guardrail isn’t a policy document. It’s the fact that the old model is one export away at all times.
Felton also made a broader point to Fortune about how AWS is approaching this internally: he’s not mandating identical tools or tasks across finance teams. Instead, he’s betting that the people closest to a clunky process, the pricing analysts juggling 18 tabs under deadline, understand its friction better than anyone issuing directives from above. That’s a specific management choice, not a generic AI talking point. And it cuts against the top-down rollout plans most large companies default to.
None of this erases the accountability question a knowledgeable reader will ask first. If a chatbot recommends a discount worth millions of dollars on a multi-year contract, who signs off? And what happens when the model’s natural-language interpretation of a question doesn’t match what the analyst actually meant. AWS hasn’t published details on the approval chain sitting on top of the chatbot’s output, and Fortune’s reporting doesn’t address it either. The exportable spreadsheet is a check on the math, not necessarily a check on the decision.
Felton’s wider comments to Axios on October 16 point to why AWS is willing to take that risk. He told Axios that AWS recoups the cost of its AI server investments within three years. A nearly $500 billion order backlog as of the second quarter, he said, gives him confidence the current pace of AI capital spending is sustainable rather than speculative. Internal projects like the pricing chatbot are the quieter half of that story. A hyperscaler is using the same AI it’s selling to reshape how it runs its own back office, discount by discount.
Also read: Krea 3 Ultra edits images faster and cheaper than GPT Image 2.5 • Notion agents can now run unsupervised for 24 hours, up from five minutes • OpenAI says its own AI agents weakened the system built to flag misbehavior
The Metropolix, from Startup Fortune
→
This article is posted in AI News, check it out for more related stories.
Almost there. Sign in and your reply posts straight away.