Opinion | Big Tech Fooled America Once. The Second Time’s Not Going So Well.
By Gammatek ISPL Last updated: September 2026 | 14 min read
Author note: Gammatek ISPL covers enterprise technology and software markets, with a focus on separating substantive product claims from marketing language. This piece draws on publicly available product announcements, enterprise software marketing materials, and industry pricing/positioning data current as of September 2026.
Why This Matters
If you've bought, sold, or evaluated business software in the last two years, you've watched the same word get stapled onto nearly every product category that exists: "AI-powered." Accounting software. Backup software. CMMS platforms. Contract management tools. HR systems. Some of these additions are genuinely useful. A large number are a logo change and a chatbot bolted onto the same product that existed in 2023. If you're the person deciding what to buy, what to build, or what to believe about where this technology actually stands, the gap between the pitch and the product isn't an abstract industry debate — it's the thing determining whether your next software purchase, your next hire, or your next strategic bet pays off or wastes a year.
A Familiar Shape
Big Tech has run variations of this playbook before, and it's worth being specific about the pattern rather than just gesturing at "hype" as a vague accusation.
Social media platforms spent a decade selling the idea that they were connecting the world and strengthening democracy, while their internal research (later surfaced through leaks and lawsuits) showed executives were aware of measurable harms to mental health and civic discourse well before public messaging caught up to that reality. The metaverse was positioned, briefly, as the successor to the internet itself — an inevitability requiring massive capital reallocation — before quietly being deprioritized once consumer adoption never materialized at the promised scale. Web3 and crypto's "future of finance" framing ran a similar arc: enormous marketing spend, genuine technical innovation buried inside a mountain of speculative excess, and a public correction once the gap between claims and delivered value became too large to paper over.
None of this means AI is following an identical trajectory to any of those — the underlying technology is genuinely more capable and more broadly useful than the metaverse ever was, and that distinction matters. But the marketing mechanismwrapped around it — bold civilizational claims, breathless competitive positioning, urgency framing designed to discourage scrutiny — is running the same playbook that's now failed publicly at least twice in the last decade.
Why "The Second Time" Is Going Differently
Here's where the current moment actually diverges from the pattern, and it's worth naming precisely why.
Enterprise buyers got burned, and enterprise buyers have long memories. A finance leader who greenlit a "blockchain transformation" initiative in 2021 that quietly got shelved in 2023 is a more skeptical audience for "AI transformation" pitches in 2026 than they were the first time around. That skepticism shows up in procurement cycles — enterprise software buyers are increasingly asking for documented ROI data before approving AI-labeled tool purchases, rather than approving based on vendor claims and competitive fear-of-missing-out alone.
The claims are more falsifiable this time, and falsification has already started. Unlike "the metaverse will be the successor to the internet" — a claim that could coast on vagueness for years before being tested — many current AI claims are specific and near-term enough to check against reality quickly: productivity gains, job displacement percentages, model capability benchmarks. When Yale's Budget Lab and the Brookings Institution independently found no meaningful shift in occupational mix through early 2026, despite two years of "AI jobs apocalypse" messaging, that wasn't a vague vibe shift — it was a checkable claim that didn't hold up on schedule, and it got reported as such.
The "AI-powered" label has been diluted to the point of losing meaning, and buyers have noticed. This is the part most directly relevant if you evaluate software for a living, so it's worth sitting with.
The "AI-Powered" Label Problem, Category by Category
Walk through nearly any enterprise software category today, and you'll find the same pattern repeating: legacy platforms that have existed for a decade or more, now marketed with "AI-powered" front and center, often describing features that are incremental at best.
Consider the accounting software space. QuickBooks Enterprise, long the default for mid-size businesses, now markets AI-assisted reporting and forecasting features — some genuinely reduce manual reconciliation work, others amount to a chatbot layered over the same reporting engine that existed in prior versions. The difference matters enormously if you're the finance team deciding whether "advanced reporting in QuickBooks Enterprise" justifies an upgrade cycle, or whether it's the same functionality with new marketing copy.
The same pattern shows up in enterprise backup and recovery software, where "AI-powered threat detection" has become a near-universal claim across corporate backup software vendors — sometimes describing genuinely improved anomaly detection, sometimes describing a rules-based system that was already there, rebranded. Enterprise contract management software has followed an identical arc: "AI-powered contract review" ranges from meaningfully useful clause-flagging to a marketing refresh on optical character recognition that's existed for years. Even CMMS (computerized maintenance management system) platforms — a category about as far from AI hype as software gets — now routinely advertise "AI-driven predictive maintenance," with wildly inconsistent levels of substance behind the label depending on the vendor.
This isn't a claim that AI adds nothing to these categories. In several of them, it genuinely does — predictive maintenance built on real sensor data and machine learning models can meaningfully reduce downtime, and AI-assisted contract review can catch clauses a rushed human reviewer misses. The point is narrower and more actionable: the label itself has stopped being a reliable signal. "AI-powered" now tells you almost nothing about whether you're looking at a genuine capability upgrade or a rebrand, which means the burden has shifted back onto buyers to evaluate the actual feature, not the marketing category it's filed under.
An Implementation Consideration for Software Buyers
If you're evaluating any "AI-powered" enterprise tool right now — whether it's enterprise HR software, enterprise scheduling software, or a CMMS platform — a few questions cut through the marketing layer faster than reading the product page:
Ask what the feature did before the AI label was added, and what specifically changed. Vendors that can answer this precisely, with a real technical explanation, are usually describing genuine capability. Vendors that answer with confidence-building language but no specifics usually aren't.
Ask for a reference customer using the specific AI feature, not just the product generally. A vendor with real traction on a genuine capability will have this ready. A vendor rebranding an existing feature often won't.
Check whether the pricing changed when the AI label was added. A feature that's substantively new and expensive to build tends to carry a real price change. A feature that's purely a marketing refresh often doesn't, because the underlying cost to the vendor didn't change either.
Run a time-boxed pilot before a full rollout, regardless of how confident the sales pitch is — this is the single most reliable way to separate marketing claims from delivered value, and it costs far less than discovering the gap after a company-wide rollout.
None of this requires cynicism about AI as a technology. It requires the same due diligence buyers eventually applied to "cloud-based," "blockchain-enabled," and "big data" in their respective hype cycles — labels that, in hindsight, described a wide spectrum from genuinely transformative to purely cosmetic, sorted only by looking past the label itself.
What Happens When the Label Stops Working
There's a second-order effect worth naming: as buyers get better at seeing through inflated AI claims, vendors with genuine capability start actively distancing themselves from the generic "AI-powered" framing, because the label has become a liability rather than an asset in sophisticated buying conversations. You can already see early versions of this — vendors leading with specific, checkable claims ("reduces reconciliation time by X hours per month, verified across Y customer deployments") rather than the category-wide "AI-powered" banner that increasingly reads as a yellow flag to experienced buyers.
That's arguably a healthy correction, not a failure of the technology. The first wave of any hype cycle rewards whoever claims the label loudest. The second wave — the one we're now entering — starts rewarding whoever can back the claim with a number.
The Honest Read
Big Tech's AI narrative isn't collapsing the way the metaverse's did, and it shouldn't — the technology has real, demonstrated utility that those earlier waves largely lacked at scale. But the marketing mechanism wrapped around it is wearing thin faster than it did the first two times, because the audience — enterprise buyers especially — has now watched this specific movie twice and knows which parts of the trailer to distrust. The label "AI-powered" used to do real work convincing people something was worth their attention. Increasingly, it's the thing making people ask more questions before they believe it. https://www.gammateksolutions.com/post/amazon-prime-members-just-got-a-huge-new-perk-here-s-how-to-use-it https://www.gammateksolutions.com/post/amazon-prime-members-just-got-a-huge-new-perk-here-s-how-to-use-it
Comments