One Company’s Mistake is another Company’s Audacity

The pervasive proliferation of AI being embedded into products that never asked for it and used by consumers who largely did not request the addition is something that companies continue to be oblivious to, yet keep continuing to deploy as though the demand were self-evident. Phones, search engines, spreadsheet software, fridges, washing machines; the list keeps growing, the underlying assumption behind each deployment being that an intelligent layer makes any product better, regardless of whether the user wanted one. Whether consumers actually wanted this is a question most boardrooms have shown little interest in answering, partly because the investment thesis depends on not asking it too loudly.

This is not to underplay the importance or the demand for AI, there is plenty of that, just follow the money; but focusing on AI as the product rather than as a feature embedded within one is a gamble with a poor track record. Just ask Meta who learned this the hard way when it bet its entire identity on the meta-verse: re-branding, restructuring and spending billions on a vision that consumers largely ignored. The pivot away cost thousands of jobs and years of strategic momentum that the company is still, in some ways, recovering from.

That same divide plays out right now with Microsoft and Apple, two companies facing the same cultural moment around AI and responding to it in ways that could not be more different. Microsoft went all-in on Copilot and Apple launched a $599 laptop. One company went on all in on AI as the product, the other implemented it as a feature.

Show me the AI

Microsoft did not just add AI to its products in the way that a company might add a new feature to an existing offering. It made AI the product, restructuring its suite of software around Copilot in a bet so total that the company’s identity became difficult to separate from it. Copilot now lives in Windows, Word, Excel, PowerPoint, Edge, and even the physical keyboards. The Office suite was re-branded around it. Tens of billions committed to AI infrastructure and every major communications channel including earnings calls, product launches, and press releases, all oriented toward the same message.

The market responded with considerably less enthusiasm than the investment implied it would. Of roughly 450 million Microsoft 365 users with access to Copilot Chat, only 15 million pay for it, a conversion rate of 3.3% that, in any other product category, would prompt serious questions about whether the pitch was landing.[1] Among users who stopped using Copilot entirely, 44.2% cited distrust of its answers as the main reason for leaving, which is a particularly damaging figure for a product being sold on its reliability.[2] The paid subscriber base contracted sharply too, with US paid subscriber market share falling from 18.8% in July 2025 to 11.5% in January 2026, a 39% drop across six months.[3]

Satya Nadella admitted internally that Copilot’s integrations with Gmail and Outlook “for the most part don’t really work” and are “not smart,” before assuming direct hands-on control of AI product development, a move that does not suggest a company confident in the direction its flagship bet is heading.[4] Users, for their part, have been less diplomatic. When Microsoft claimed on social media that it had “heard” users wanted Copilot Mode at work, one user replied: “No, you heard wrong. Literally no one asked for all this AI. In fact, everyone wants to know how to remove it.”[5] Microsoft eventually halted forced Copilot installs in March 2026, after sustained backlash from consumers and commercial customers who had not been consulted before the roll-out began.[6]

The most revealing detail, though, came not from user reviews or subscriber data but from Microsoft’s own legal team. Buried in Copilot’s Terms of Use, updated quietly in October 2025, is this line: “Copilot is for entertainment purposes only. It can make mistakes, and it may not work as intended. Don’t rely on Copilot for important advice.”[7] A company spending billions to position AI as the future of productivity was, in the same breath, telling its users not to rely on it, a contradiction that the marketing department and the legal department appear to have never reconciled.

The Lawyers, The Psychics, And The Fine Print

That disclaimer was, in its own way, good lawyer-ing. It was not an accident or an oversight but a legal calculation made by lawyers who understood exactly the kind of exposure the product created. When a product is an AI system making claims, generating content, and influencing decisions at scale, the gap between what it promises and what it reliably delivers becomes a legal liability, and Microsoft’s legal team responded by doing what legal teams do best: update terms and conditions to bury a clause for when the product cannot fully back its own marketing, transferring all liability onto the user before anything went wrong.

The Terms of Use extend well beyond the entertainment disclaimer. The same document states that Microsoft makes no warranty or representation about Copilot’s outputs, that users are solely responsible for anything they publish or share from it, and that users must indemnify Microsoft against any claims arising from their use of the tool.[8] Legal analysts have described the clause as an attempt to limit liability when the product fails, and one publication noted the phrasing matches disclaimers used by psychics “to avoid getting sued,” which is a comparison that presumably nobody in Redmond planned for when the product road-map was approved.[9]

This is the bind Microsoft built for itself over several years of aggressive marketing. It spent that time positioning Copilot as an essential productivity tool, a workplace revolution, something Nadella himself claimed to use in running a multi-trillion dollar company, while its legal department was simultaneously confirming that none of those claims carried any warranty whatsoever.[10] Enterprises that built work-flows around Copilot found themselves holding a tool their own vendor would not stand behind, operating in a legal gray area that Microsoft’s terms of service had carefully constructed to ensure the exposure sat with the customer.

Apple with the MacBook Neo, on the other hand, faced none of this because hardware is concrete, verifiable, and guaranteed in ways that AI output simply cannot be. A 13-inch Liquid Retina display either delivers 500 nits of brightness or it does not, and battery life either holds for 16 hours or it does not. These are claims Apple is legally comfortable making because the hardware can actually back them up, without any hallucination risk to hedge against or legal small print that undercuts the marketing. The contrast is structural. Microsoft bet on a technology it could not fully control and then had to legally protect itself from the consequences of that bet, while Apple bet on hardware it engineered from the ground up, where every spec is testable and every claim is defensible. The battle, as a result, become one between liability exposing software against veritable hardware.

To AI or not to AI

Apple did not hold a flashy AI keynote for the MacBook Neo, and there was no stage moment built around artificial intelligence changing the nature of computing forever, which, given how those keynotes tend to age, was probably a wise editorial decision. What Apple announced in March 2026 was quieter, and in some ways more radical than any AI announcement could have been: a $599 laptop, the cheapest Mac the company had ever made.[11]

That price is significant in context. The MacBook Air, Apple’s previous entry-level machine, starts at $1,099, which means the Neo sits $500 below it[12], a gap wide enough to reach an entirely different category of buyer. For that money, the machine delivers a 13-inch Liquid Retina display, up to 16 hours of battery life, a recycled aluminum body available in four colors, and Apple’s A18 Pro chip, the same silicon powering the iPhone 16 Pro.[13] iFixit found it to be Apple’s most repairable laptop in 14 years, which matters to a growing portion of buyers who have started factoring repairability into purchasing decisions.[14] It is fast, silent, and by every early account, genuinely capable for the vast majority of users. Reviewers running real workloads, including browsing, writing, photo editing, and light video work, found nothing to complain about.[15]

There are trade-offs, of course: including no backlit keyboard, no fast charging, and external display support that tops out at a single 4K monitor.[16] For the student, the first-time Mac buyer, or the person who has been on Windows and waiting for a reason to switch, those gaps are unlikely to be dealbreakers, and Apple designed the machine with precisely that buyer in mind, making every trade-off consciously rather than accidentally.

What makes the Neo genuinely audacious is that it contains AI too. The A18 Pro’s 16-core Neural Engine runs Apple Intelligence fully on-device, and Apple claims the machine handles AI tasks up to three times faster than comparable Windows laptops.[17] Apple did not lead with that. The pitch was the price, the design, the product, and the AI was mentioned in passing because the product was confident enough not to need it as the headline. That is a fundamentally different philosophy from Microsoft’s.

The Product Point

All of this is not to say that Microsoft is a failing company, far from it; it remains one of the most valuable companies on the planet, but right now its AI strategy is visibly struggling in ways that cannot be explained away by early-adoption curves or market timing. Consumers do not want AI forced onto products they already use and trust. They want it to work quietly, in the background, improving something they chose to buy rather than replacing the thing they chose with something they did not ask for. Apple understood that. The MacBook Neo is not an anti-AI product, it is an AI product that has the confidence to not lead with AI, and that distinction is the whole game.

The tech industry keeps arriving at this lesson and leaving before it sinks in, convinced that the next platform shift will be different, that this time the technology is so transformative that consumers will follow regardless of whether the product is actually good. Consumers are not waiting to be told what the future looks like. They are waiting for something good, at a price that makes sense, that works from the first day they use it, and the companies that build that tend to win, while the ones that get lost chasing the investor narrative tend to spend years and billions finding their way back to a position they could have held all along.

Microsoft built a Copilot key into its keyboards before most users had any reason to press it. Apple built a $599 laptop and sold out.

The score, for now, is not even close.

[1]Microsoft’s own ToS calls Copilot ‘entertainment only’ amid adoption slump. The Next Web, April 2026. https://thenextweb.com/news/microsoft-copilot-entertainment-only-disclaimer-adoption

[2]ibid

[3]ibid

[4]Microsoft CEO admits Copilot integrations “don’t really work” as adoption falters. PPC Land, December 2025. https://ppc.land/microsoft-ceo-admits-copilot-integrations-dont-really-work-as-adoption-falters

[5]”You heard wrong” — users brutally reject Microsoft’s “Copilot for work” in Edge and Windows 11. Windows Latest, November 2025. https://www.windowslatest.com/2025/11/28/you-heard-wrong-users-brutually-reject-microsofts-copilot-for-work-in-edge-and-windows-11

[6]Microsoft says it won’t auto install Microsoft 365 Copilot app on Windows 11. Windows Latest, March 2026. https://www.windowslatest.com/2026/03/18/microsoft-says-it-wont-auto-install-microsoft-365-copilot-app-on-windows-11-likely-due-to-outrage-over-microslop

[7]Microsoft’s own ToS calls Copilot ‘entertainment only’ amid adoption slump. The Next Web, April 2026. https://thenextweb.com/news/microsoft-copilot-entertainment-only-disclaimer-adoption

[8]Six More Warnings Hidden in Copilot’s Legal Fine Print. Office Watch, April 2026. https://office-watch.com/2026/microsoft-copilot-terms-of-use-warnings

[9]Microsoft’s legal terms claim Copilot for entertainment only. Technobezz, April 2026. https://www.technobezz.com/news/microsoft-s-legal-terms-claim-copilot-for-entertainment-only

[10]Microsoft: Take Copilot Seriously, Not Legally Liable. Moyens I/O, April 2026. https://uk.moyens.net/ai/microsoft-copilot-take-seriously-not-legally-liable

[11]MacBook Neo: Price, release date, specs, features and MacBook Air comparison. Macworld, March 2026. https://www.macworld.com/article/2854313/macbook-neo-design-processor-specs-release.html

[12]ibid

[13]MacBook Neo. Wikipedia. https://en.wikipedia.org/wiki/MacBook_Neo

[14]ibid

[15]MacBook Neo review: A truly great Mac at an unbelievable price. 9to5Mac, March 2026. https://9to5mac.com/2026/03/10/macbook-neo-review

[16]MacBook Neo 2026: Specs, Price & Full Guide. Laptop Outlet, March 2026. https://www.laptopoutlet.co.uk/blog/macbook-neo-2026-guide.html

[17]Say hello to MacBook Neo. Apple Newsroom, March 2026. https://www.apple.com/newsroom/2026/03/say-hello-to-macbook-neo