Microsoft Corp. launched Scout, a new artificial intelligence software designed to function like an always-active executive assistant, marking the latest evolution of workplace AI efforts. According to Business Standard, Scout will appear on internal email and calendar systems as if it were just another helpful employee, allowing the assistant to handle a wider range of tasks autonomously. The tool is built on OpenClaw, a platform that turns the models behind ChatGPT and Claude into always-on agents, though Microsoft didn't disclose pricing for the new service. Scout will initially require a subscription to Microsoft's GitHub Copilot coding assistant, with customers likely charged based on usage rather than a flat subscription fee. This launch comes at a critical moment as the AI industry enters its first real price discovery phase, with GitHub's decision to shift Copilot toward consumption-based billing revealing the true cost of token consumption that has been artificially discounted for the better part of two years.
The AI boom is entering its first real price discovery phase as token consumption finally meets economic reality, with enterprises built operating models around artificially cheap intelligence now discovering that usage scales far faster than budgets. As reported by Investing.com India, the online backlash was not really about GitHub but about the sudden realization that the economics underpinning the entire AI boom may be very different from the economics investors have been assuming. The meter started running for millions of users at the exact moment bankers were preparing to package future token growth into one of the largest AI equity stories ever floated, creating a remarkable juxtaposition between demand-side reality and supply-side preparation for public market debuts. The difference is that intelligence is proving far more difficult to price than any of those businesses, with traditional software costs compounding exponentially as autonomous agents generate their own workload through hundreds of micro decisions beneath the surface.
Major companies are implementing strict controls on AI usage as computing costs mount, marking a sharp contrast from earlier encouragement campaigns. Walmart Inc. has capped staffers' use of an in-house AI agent that helps with workplace tasks, while Uber Technologies Inc. is limiting each employee's monthly spending on certain AI coding tools to $1,500 per tool. The ride-hailing company had already blown through its annual budget for Claude Code, a popular tool from Anthropic PBC. Google CEO Sundar Pichai recently noted that monthly usage of the company's AI products has increased sevenfold in the past year to 3.2 quadrillion tokens, stating that "many companies are already blowing through their annual token budgets, and it's only May." As reported by Bloomberg, agentic AI is "uniquely expensive," consuming one-thousand times more tokens than more basic tasks since agents continue to soak up computing power in the background. This shift from "ownership to consumption" represents a fundamental change in how companies approach AI budgeting, with finance departments experiencing sticker shock as costs compound rather than rise in straight lines.
Global stocks reached near record highs on Monday, with S&P 500 Futures up 0.3% and Nasdaq Futures gaining 0.3% following a record-breaking week for both benchmark indices. According to Reuters, markets in Tokyo, Seoul, and other cities traded at or close to all-time highs, largely driven by demand for AI-related products. The AI boom drove demand on Monday, while news of military strikes in the Gulf tempered optimism for a reopening of the Strait of Hormuz, pushing up oil prices. Brent crude futures rose by nearly 3.3%, to $94.12 per barrel, prompting a sale of government bonds as investors held onto the idea that Iran/US negotiations remain possible despite attacks by both sides.
The AI transformation creates clear winners and losers within wealth management roles. Junior analyst roles structured around production are the most exposed, along with middle-office functions built around reconciliation and reporting, and pure investment writers who repackage third-party research. As reported by Investing.com India, McKinsey frames the divide with admirable bluntness: firms focused on delivering outputs will see their economics jeopardised by AI, while those focused on outcomes will fare considerably better. The curator role becomes more valuable as judgment becomes the binding constraint when AI makes content infinite, while relationship orchestrators who can translate complex questions into succession planning remain deeply human and contextual work that AI cannot replace. The engineer-investor role emerges as firms develop internal AI infrastructure, combining financial understanding with software and data engineering capability. However, many companies did not merely adopt AI but reorganized themselves around it, with headcount reductions justified by productivity gain assumptions that now face scrutiny as the era of unquestioned consumption ends.
Companies are implementing various strategies to reduce AI costs amid the price surge. As reported by Wall Street Journal, some companies are switching to free, open-source AI models that anyone can download, though these are not as powerful as ChatGPT or Anthropic's Claude. Others are moving to smaller, more specialized models built for specific industries like real estate or finance, rather than giant general-purpose ones. The price difference can be dramatic, with Adrian Balfour of consultancy Envorso noting that "the big large monolithic model costs $15 per million tokens, but you can get that down to like five cents if you use the smaller mini model." Some companies are also breaking big AI tasks into smaller steps, handing each piece to the cheapest model that can handle it effectively. However, some 94% of senior executives plan to continue investing in AI even if it does not pay off in 2026, according to a BCG survey, suggesting companies may need to find a Goldilocks approach to adoption rather than completely limiting usage.
