At one point in the spring of 2026, it seemed like artificial intelligence was brought up in the first three minutes of every other earnings call. After obliquely hinting at an AI strategy, Allbirds, a wool sneaker company that built its reputation on carbon footprints and minimalist design, saw a nearly six percent increase in just one afternoon. Sitting on GPU warehouses, Bitcoin miners abruptly changed their name to “compute providers.” A karaoke company also changed course. To be honest, it was a bit absurd. And investors continued to purchase it, at least for a few weeks.
It was not a novel pattern. Tesla overcame years of uncertainty. Pets.com didn’t. This time, there seems to be a difference between the promise on the slide deck and the actual conditions inside the buildings where the work is being done. When you walk into a typical Fortune 500 office, you’ll see a copilot tool that has been deployed, a Slack channel filled with screenshots of prompts, and a senior vice president who is unable to adequately explain what changed on the P&L. Ninety-nine percent of these businesses claim to have implemented AI. One percent say their rollout is mature. It’s uncomfortable math.
The amount of corporate AI discourse that still sounds like a stage play shot from a single fixed angle is difficult to ignore. In software design, skeuomorphism refers to the practice of making new things resemble old ones. Paper-planner-like digital calendars. apps for emails with tiny envelope icons. The current wave of AI deployment almost perfectly fits the pattern. With a chatbot positioned somewhere close to the customer service line, businesses are using extraordinary tools to accomplish exactly what they were already doing, just a little bit more quickly.
Businesses that have truly succeeded typically exhibit a more subdued quality. Splashy pivots are not being performed by them. In late 2025 and early 2026, Anthropic and OpenAI both placed massive bets on what insiders refer to as the “installation gap,” collaborating with infrastructure investors and private equity firms to integrate AI directly into operational workflows as opposed to selling access to a model and walking away. An MSP innovation program’s director, Andrew Zehnder, summed up something that many operators have been discussing for months. He wrote that even though there is little actual workflow integration, AI adoption has become a KPI that employees report favorably. People take a quick look at ChatGPT, send a few emails, and then return to their previous activities. Green checkmarks are seen by leadership. Nothing shifts.
In the meantime, rather than adding AI on top of flawed procedures, the companies that quietly compound gains are typically the ones redesigning the work itself. A few software-focused businesses, a few logistics companies, a smaller group of insurers, and biotech research organizations have begun to report real margin movement. Not the kind that is out of breath. Just the consistent, somewhat dull type that appears in operating income after two quarters. Speaking with people in these establishments gives me the impression that the model they purchased isn’t what makes a difference. It’s whether anyone took the time to map the work itself before automating certain aspects of it.
Brian Hopkins, a Forrester analyst, put it quite simply. According to him, the tech companies that developed this technology fabricated a story about how quickly everything would change. However, people don’t change all that quickly. As this develops, it’s possible that the true story of 2026 isn’t the unimpressive makeovers or the dramatic turns. It’s the smaller, less visually appealing group of businesses that quietly began implementing AI instead of attempting to execute it. It’s still unclear if they will be rewarded for their patience. For now, though, the contrast is becoming more pronounced.
