| Highlights: |
|
In February 1995, astronomer Clifford Stoll wrote an essay for Newsweek dismissing the internet as overhyped. No online database would replace the daily newspaper, he argued, and no computer network would change how business or government worked.
By 2012, Newsweek had stopped printing a weekly edition, and most of us now read the news on a device that fits in a pocket.
Stoll was in good company. Ken Olsen, founder of Digital Equipment Corporation, said in 1977 that there was no reason anyone would want a computer in their home. Smart people dismissed each of the last three breakout business technologies at inception, and each followed the same arc anyway: new and novel, then some people use it, then it’s expected, then you can’t work without it.
We map that arc for clients in a simple diagram with four S-curves: the personal computer, the internet, smartphones, and now AI. The pattern is the point.
The personal computer climbed through the 1980s and 1990s, from hobbyist machine to the default piece of office furniture. The internet followed, compressing the same journey into roughly a decade. Smartphones compressed it further: in 2011, about 35% of U.S. adults owned one; by 2021 it was 85%, according to Pew Research Center. Each successive curve was steeper than the last.
Notice what happened at each stage. Nobody gets credit today for having email, a website, or a smartphone. They became table stakes, which means they became invisible. The technologies that provoked the most skepticism at inception are now the ones we literally cannot work without.
Here’s the part the early skeptics missed, and the part that matters most for planning: these technologies didn’t just get adopted, they spun off entirely new business processes. The internet didn’t simply put brochures online; it produced email, and email reorganized the working day for every knowledge worker on earth.
Smartphones didn’t simply untether the desk phone; they produced texting, mobile approvals, and the always-connected workplace that Slack and Teams later formalized.
Each technology arrived as a tool and ended up as an environment. The processes it enabled, not the device itself, became the lasting change. A CFO at one of our roundtables put it well: “I don’t remember deciding to adopt email. One day it was just how we worked.”
AI at work today sits roughly where smartphones sat in 2009: well past novelty, not yet expected. Gallup’s workplace tracking shows AI use among U.S. workers climbing from 40% to 45% in a single quarter, with frequent use (a few times a week or more) jumping from 19% to 23%. Those are big moves for one quarter. For comparison, earlier curves measured their progress in years.
But the slope of this curve is different, for three reasons. First, there’s no hardware barrier. The PC required buying machines, the internet required infrastructure buildout, and smartphones required a device upgrade cycle. AI rides on all three, fully built. That’s a big part of why ChatGPT reached an estimated 100 million users in two months, the fastest consumer adoption of any application in history at the time.
Second, AI isn’t only standalone software. It’s also a feature being added to applications businesses already own: the office suite, the CRM, the ERP, the meeting platform. This is genuinely new. Your company will adopt AI whether or not it ever makes an explicit decision to do so, because your software vendors have made the decision for you. The previous three curves never had that dynamic.
Third, the interface is plain language. Using a PC required training; using the early internet required patience. Using AI requires typing a question. The skill floor has never been lower for a technology this powerful, which removes the friction that slowed every prior curve.
Now the caveat, because a steeper curve doesn’t mean an easier one. Usage is spreading faster than skill. Tools travel on the new, steep curve; practices still travel on the old, human one. The gap between those two curves is where both the risk and the opportunity sit.
We saw this before. The companies that won the internet era weren’t the first to have websites; they were the first to rebuild processes around the web. The same will be true here. An organization where 45% of employees have tried AI but no workflow has changed isn’t ahead of the curve. It’s standing at the base of it with a ticket in hand.
Retire the “if” conversation. Three precedents and the steepest adoption data ever recorded say this is a when question. Leadership time spent debating whether AI will matter is time taken from preparing for when it’s expected, and “expected” is closer than it was for any previous curve.
Inventory where AI is already arriving. List your core applications and check each vendor’s AI roadmap. Much of your adoption is happening inside software you already pay for. Knowing what’s coming lets you govern it instead of discovering it.
Pick a process, not a tool. Every prior curve rewarded companies that redesigned a workflow, not ones that bought a gadget. Choose one process with real volume, such as proposals, customer responses, or month-end reporting, and rebuild it with AI in the loop. Measure the before and after.
Build fluency deliberately. Fluency is an organizational capability, not a count of licenses. That means training people on what the tools are good and bad at, setting usage expectations, and creating safe places to practice. This is the work our AI readiness assessments consistently find missing, even in companies with high tool adoption.
Treat it as change management. The hard part of the last three curves was never the technology; it was people, process, and governance. AI is the same change problem on a shorter clock. The disciplines that carried companies through ERP implementations and M&A integrations are exactly the ones this curve demands.
Nobody gets credit today for having email, the internet, or a smartphone. AI is on exactly the same path, just on a steeper slope, and the companies that look smart in hindsight will be the ones that started climbing while the curve still looked optional.
© 2026 SVA Consulting