By Ion Anghel · April 2026
Every week, another product announces its shiny new AI feature. Adobe Acrobat now has an AI assistant — for an extra $4.99/month on top of your already expensive Pro subscription. Windows 11 spent the last two years cramming Copilot into Notepad, Paint, Snipping Tool, Photos, File Explorer, notifications, and the Start menu. JetBrains added an AI Assistant to PhpStorm that burns through your credits in two sessions and, until recently, made simple copy-paste operations take 15 seconds.
And yet, when you open these products, the core experience isn't noticeably better. It's often worse.
The Pattern
The pattern is always the same: take an existing product with existing problems, bolt an AI feature on top, and call it innovation. Don't fix what's broken — just add a chatbot next to it.
Adobe Acrobat is a perfect example. Users have been complaining for years about basic UX issues — page numbers that don't display, zoom controls that vanish after updates, unintuitive navigation. The response? An AI assistant that summarizes PDFs. Meanwhile, people are accidentally getting subscribed to the AI add-on and can't cancel without paying fees. Community forums are full of users calling it a scam. The core product isn't getting better — it's getting more expensive.
Windows 11 took this to an art form. Microsoft shoved Copilot into every corner of the operating system, from dedicated keyboard keys to lightweight apps like Notepad. Users reported the assistant activating unexpectedly during routine tasks, interrupting workflows, and consuming system resources even when not in use. The backlash was so severe that the term "Microslop" went viral, and Microsoft was forced into a humiliating retreat — publicly admitting they would "reduce unnecessary Copilot entry points" starting with apps like Snipping Tool, Photos, Widgets, and Notepad. In March 2026, Windows head Pavan Davuluri acknowledged "pain points" and confirmed Microsoft is hitting the brakes on aggressive feature expansion to focus on core OS stability. People familiar with Microsoft's plans say the company is moving to reduce AI bloat across Windows 11 this year.
Users didn't want AI assistance in Notepad. They wanted a text editor that works.
JetBrains tells a similar story. PhpStorm's AI Assistant plugin caused the IDE to freeze during basic operations — one user described it as "what used to be instant now feels like turn-based strategy." The credit system is opaque, with JetBrains quietly adjusting quotas downward in August 2025. Developers started looking at alternatives, not because those tools are better IDEs, but because JetBrains' AI integration was actively degrading the core product.
AI Amplifies. It Doesn't Fix.
Here's what nobody in product management wants to hear: AI amplifies whatever is already there. If your product has great UX, solid architecture, and clean data, AI features can genuinely enhance the experience. If your product has a confusing interface, legacy baggage, and frustrated users — congratulations, you now have a confusing interface with an AI chatbot in the corner.
This isn't a theoretical concern. Volkswagen launched Cariad in 2020 to build a unified AI-driven operating system for all 12 of its brands. Five years later, it had become one of the automotive industry's most expensive software failures — a 20-million-line codebase riddled with bugs, delays on major vehicle launches including the Porsche Macan Electric and Audi Q6 E-Tron, and 1,600 job cuts. They tried to build the future while the foundation was crumbling.
An ISACA report from 2025 put it bluntly: AI-enabled features are often added to products merely for show — lightweight features that offer little functional value, included to feed into market hype or meet investor expectations. Rather than driving innovation, AI is increasingly being used to meet market expectations rather than operational needs.
The Real Cost
The cost of "AI everywhere" isn't just annoyed users. It's real performance degradation. Windows 11 users report slower search, a less reliable File Explorer, and increased memory usage — all while Copilot runs in the background. JetBrains developers see their credits vanish and their IDE freeze. Adobe users pay more for a product that doesn't do the basics better.
The memory demand from AI integrations is exposing rampant software bloat — apps consuming resources far beyond necessity, as AI integrations add layers of complexity that demand more RAM, even as users question their utility.
There's also a trust cost. When every app has an AI button that doesn't meaningfully help, users start ignoring AI features altogether — including the ones that actually work. The signal gets lost in the noise.
What Should Happen Instead
Before adding an AI feature, product teams should answer one question: what specific problem does this solve for the user that we couldn't solve without AI?
If the answer is "it looks innovative" or "competitors are doing it" or "the board wants an AI strategy," that's not a product reason. That's a marketing reason.
Fix the zoom controls in Acrobat. Make File Explorer fast again. Let PhpStorm copy text without a progress bar. Then, and only then, think about where AI genuinely adds value.
AI is the future. I believe that without reservation. But the future doesn't arrive by stapling it onto the present and hoping nobody notices the duct tape.
Disclaimer: I believe AI is transformative technology and that most of the current issues will be resolved. But ignoring these issues doesn't help — acknowledging them does. The companies that will win with AI are the ones building it on solid foundations, not bolting it onto broken products.