The music industry spent a decade fighting streaming. Labels had a product that worked — packaged albums at $15 a unit — and they were not 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 asked "what can we sell now that we could not sell before?" won. The ones that asked "how do we keep selling what we already sell?" didn't.
That pattern has repeated in every technology cycle since. I have been through three of them. The story is always the same.
The Bitcoin version
I was at NYDIG working with major banks when Bitcoin started showing up in boardrooms. Most banks treated it as another asset for the balance sheet. Buy it, hold it, maybe offer clients a way to get exposure through an existing product. That was the efficiency play: take something new and fit it into your current model.
Meanwhile, companies like River and Strike built entirely new products around digital payments. They did not try to make Bitcoin fit into the old banking model. They asked what new services became possible because the technology existed, and they built those services from scratch.
The banks that bolted Bitcoin onto their existing model watched the revenue go somewhere else.
Same technology. Different question. Different outcome.
The Microsoft cloud version
Microsoft hired me to work with about a dozen of their software partners during the cloud transition.
Some partners 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 partners that won asked a different question: "What can we sell now that we could not sell before?" They redesigned their products around the subscription model. They built features that only worked because the software was always connected. They went from selling licenses to selling outcomes.
Same cloud platform. Same migration. The firms that treated it as cost reduction lost. The firms that treated it as a commercial opportunity won.
The professional services version
Professional services firms are making the efficiency play right now. They are 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 is not a projection. It is happening now.
The instinct is to protect the current model. Do the same work faster. Cut costs. Hope the billing model holds.
It will not hold. It has 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 software partners that redesigned for cloud first. It will be true for the first professional services firms that launch genuinely new offerings built around what AI makes possible.
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 is 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 treated Bitcoin as just another asset class are watching payment companies serve their former customers. The software partners 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.
I have been on the winning side of this three times. The firms that won always started with the same question: "What can we sell now that we could not sell before?"