Boring AI Is the AI That Pays

Boring AI Is the AI That Pays

Everybody spent two years being told AI would change everything, and this month the mood flipped to a smaller, sharper question: did it actually pay for anything? The reckoning is real and overdue, and the number underneath it is not flattering. Only about one in four companies has gotten AI into real production at scale; nearly half are still running pilots. But a small group is quietly getting real money back, and their returns have been checked by an independent firm, not the vendor that sold the software. What those companies share is almost disappointing: none of them "did AI." They each found one specific, expensive-in-hours chore and handed exactly that to the machine. Stephen Forte on the ROI reckoning, three audited examples across manufacturing, consumer goods, and frontline services, the pattern that separates the winners from the pilot pile, and the single question that tells you which group you are in.

AI Just Went From Answering to Doing

Last week Anthropic did something that looked like a menu cleanup and was actually a strategy reveal. It merged Claude Chat, the back-and-forth you already know, with Cowork, Claude's agent that goes off and does a whole task across your files and tools, into a single home, and moved the agent to the cloud so it keeps working after you close your laptop and can even run on a schedule with no device on at all. Their own words: "handing Claude a task starts the same way a conversation does." Underneath the low-key rollout is the biggest change in how knowledge workers touch AI since ChatGPT arrived: the shift from consulting a smart assistant to assigning work to a tireless one. And Anthropic's data on 1.2 million sessions gives away what it is really for, over 90 percent of it is not software engineering, it is the administrative grind that surrounds every job. Stephen Forte on the workflow shift your team is about to feel, the four-way land grab it touched off with OpenAI, Microsoft, and Google, and the three moves to make before an always-on agent lands on your systems.

You Don't Know What AI You're Running

This week looked like a fireworks show of AI launches — OpenAI's GPT-5.6, new real-time voice models, Microsoft leaning on its own in-house models. The more important story ran underneath all of it: the AI inside your company has quietly become a black box you can neither see into nor fully trust. Microsoft has begun replacing OpenAI and Anthropic with its own cheaper MAI models inside Excel and Outlook — its AI chief said the goal is to "eliminate that cost." The security firm Wiz found six major AI coding assistants showed users a fake file path in their safety confirmation while writing to sensitive files. And an independent developer discovered Anthropic had run an undisclosed location tracker inside Claude Code for months. Stephen Forte on why you are now accountable for an AI you cannot inspect — and the three clauses to put in every AI contract before your next renewal: model-transparency and change-notification, an independent audit-logging layer, and a named owner for what is actually running in your stack.

Your AI Bottleneck Was Never the Model

The strange truth of AI in 2026 is that the technology keeps clearing bars we thought were years away — Alberta's provincial government just used Claude to scan 466 million lines of code in 20 hours, work that would have taken six and a half years by hand — while the business results stay stubbornly flat. MIT finds 95% of enterprise AI pilots deliver no measurable impact; an NBER survey of more than 6,000 executives across four countries finds roughly 90% saw no productivity gain over three years. This week the most sophisticated vendors on earth told you, in dollars, where the real bottleneck is: Microsoft committed $2.5 billion and 6,000 of its own engineers to embed inside customer companies and deploy AI for them — following Amazon's $1 billion, and Anthropic's and OpenAI's own embedded teams. Stephen Forte on why your AI bottleneck was never the model, and the three moves to make before you fund one more pilot.

AI's Insiders Just Started Hedging

Every boom has a tell, and it is never in the press releases. This week the AI boom's insiders started hedging their own story: Meta announced it will rent out its "excess" AI compute while chipmakers sold off, Oracle's SEC risk factors laid bare the strain of its $300B OpenAI/Stargate commitment, and Mark Zuckerberg told his own employees that AI-agent progress "hasn't really accelerated" as expected. Yet the same week, Abu Dhabi's MGX closed a $49B AI fund and Anthropic signed a 20-year, ~$19B data-center lease. Stephen Forte on what it means when sellers plan for surplus while buyers still pay scarcity prices — and the three moves to make before signing any multi-year AI contract: shorten and reopen, read your vendors' risk factors like a credit file, and re-run build-versus-rent every quarter.