The music industry spent the better part of two decades fighting streaming. Labels had a product that worked, packaged albums at $15 a unit, and they weren’t going to let a new technology take it away. They sued Napster. They sued individual downloaders. They lobbied Congress.
The labels that licensed their catalogs to Spotify and Apple Music early set the terms of the new market. The ones that spent years suing ended up licensing too, but on someone else’s terms, with less negotiating power and less revenue to show for it.
I watched this happen from inside the music-tech world, backing companies that were building for the streaming model while the incumbents were still trying to protect the old one. The companies that won asked what was newly possible to sell. The ones that lost asked how to protect what they already sold.
That pattern has repeated in every technology cycle since. I’ve been through five of them. The story is always the same.
The digital payments version
I was at NYDIG working with major banks when Bitcoin started showing up in boardrooms. Boardrooms fixated on Bitcoin the asset. The deeper shift was something else: fast, cheap, programmable digital payments that didn’t need a bank to clear.
The instinct at most banks was to fit Bitcoin into the products they already sold: custody for institutional clients, ETF access for wealth advisors, structured exposure inside the existing book. That was the efficiency play, take something new and slot it into the current model.
Meanwhile, companies like Strike, Cash App, and Wise built new products around what fast, cheap digital payments now made possible. They didn’t try to make new payment rails fit the old banking model. They asked what services became possible because the technology existed, and they built those services from scratch.
The banks that wrapped Bitcoin in existing products instead of building around the new payment rails watched the payments revenue go somewhere else.
Same technology. Different question. Different outcome.
The companies that won asked what was newly possible to sell. The ones that lost asked how to protect what they already sold.
The enterprise software version
I worked at Platform Computing in the years before IBM acquired the company in 2012 to build out their cloud business, and at Varicent when sales performance software was moving from on-prem to SaaS. Both companies sat in the middle of the same transition.
Some vendors saw cloud as cheaper distribution. Take the same software, host it somewhere else, cut the infrastructure costs. That was the efficiency play. They competed on price, margins shrank, and most of them are gone.
The vendors that won asked a different question: what could they now sell? They rebuilt their products around what only an always-connected, multi-tenant platform made possible: real-time data, continuous updates, and services delivered as software instead of bought as software. They went from selling licenses to selling outcomes.
Same cloud. Same migration. The vendors that treated it as cost reduction lost. The ones that treated it as a commercial opportunity won.
The professional services version
Professional services firms are making the efficiency play right now. They're buying AI tools, training their people, automating workflows. All necessary. None of it sufficient.
The math is straightforward. If AI lets your team do in one hour what used to take five, and you bill by the hour, your revenue per engagement drops 80%. Among law firms that already use AI widely, 20% report challenges meeting billable targets. That’s not a projection. It’s happening now.
The instinct is to protect the current model. Do the same work faster. Cut costs. Hope the billing model holds.
It won't hold. It's never held. The firms clinging to hourly billing while AI compresses delivery time are doing exactly what the labels did when they tried to keep selling packaged product at $15 a unit after distribution costs had fallen to zero.
What happens next
The firms that move early set the terms. That was true for the labels that licensed to Spotify before their competitors did. It was true for the enterprise software vendors that rebuilt for SaaS first. It’ll be true for the first professional services firms that launch genuinely new offerings built around what AI makes possible.
And this time, the disruption has a funding thesis behind it. In March 2026, Sequoia Capital told its portfolio companies that professional services revenue is the next trillion-dollar opportunity. They published a map of every major vertical (legal, accounting, staffing, consulting, insurance) with the dollar value of outsourced work in each. The labels didn’t see Spotify coming. You can see this coming. The map is public.
Protecting the old model always costs more than building the new one. The music industry spent hundreds of millions on lawsuits and got nothing from it except a delayed transition and a worse negotiating position. Every month a firm spends optimizing hourly billing instead of designing new offerings is a month a competitor uses to get ahead.
And the expertise to adopt a technology is different from the expertise to sell with it. The labels knew music. The banks knew finance. The software companies knew their products. None of that told them what to sell differently. Adoption is a technology skill. Commercialization is a business skill: pricing, positioning, service design, market entry. The people who are good at one are rarely good at the other.
BDO’s managing partner put a number on it: efficiency within the existing business model has a ceiling of 25 to 40 percent. After that, there’s nothing left to extract.
The question is what you do after the ceiling. The music labels that licensed early and built new revenue models around streaming are thriving. The banks that missed the shift to fast, cheap digital payments are watching new entrants serve their former customers. The enterprise software vendors that saw cloud as “cheaper hosting” are gone.
Professional services firms choosing efficiency-only AI are on the same path. The technology is different. The commercial pattern is identical.
The firms that move first on AI won’t have built better automations. They’ll have built different offerings.