The obsession with AI productivity comes at a cost. Companies are overlooking the growth opportunities in front of them.
Two ways AI adds value to a business.
The first is efficiency: fewer hours, fewer people, and fewer steps to get the same result. Customer-service reps handle more tickets. Developers ship more code. Costs go down, margins go up, and the business becomes more valuable. If AI touches half a company’s cost base and trims those costs by 10%, total expenses fall about 5%. For a typical wealth-management firm, that raises the business’s value about 10%.
The second is growth: more revenue. Investors price a business on what it will earn, so growth is worth far more than margin. A firm growing 5% a year is worth about 50% more than the same firm growing 3%. Two points of growth add half to the business’s value. Four points more than double it.
The growth blindspot.
Shlomo Benartzi, Randall Long, and Stefano Puntoni, writing in Harvard Business Review this year, asked a roundtable of senior financial-services executives to picture two wealth-management firms three years from now, identical except that one had used AI well. How much more would that firm be worth? The average answer was 2.35 times, a 135% increase.
Then they asked where the executives were putting their AI budgets. Almost all of them said efficiency. Several had never considered AI as a growth driver.
AI efficiencies top out around 10%. The executives expect 135%. AI efficiencies alone can’t get them anywhere near the valuation they’re counting on. The only thing that can is growth, which few executives are focused on. The authors call it the growth blindspot.
The efficiency gains come with a bill nobody can predict.
AI is billed by usage, and the bill is hard to predict. Only 11% of companies can forecast their AI spend within 10% of the eventual bill. Four in ten have taken a surprise AI cost to their board, and one in four has delayed or canceled an AI project over cost. The Wall Street Journal reported this month that AI spending is becoming almost impossible to budget. We think hours saved on one side and a surprise token bill on the other can net out to nothing. We wrote about why in the first part of this series.
Budgeting for the job.
The study and the Journal both recommend the same two things: match the model to the task, and watch what each request costs. Both steps help. Neither changes the unit, and we think the unit is the problem. A budget needs a different number: what a finished, defensible piece of work is worth, and what it took to get one.
The fix is to stop asking a language model to do the whole job. A language model is good at language and a poor tool for finding an answer in your records, because every step of the search is a guess. Use it for writing, and something else for finding and reasoning. Then you can see where an answer came from and judge whether it was any good.
How AI drives growth.
If you look at growth as simply revenue the business didn’t have before, AI can generate it in four ways.
- Prospecting. AI qualifies prospects and matches each one with the right person on your side, shortening sales cycles and raising win rates.
- Existing clients. Every account gets the attention and personalization only the biggest ones used to get, and each client buys more.
- The work itself. Better work keeps clients and brings referrals.
- New markets. Expertise only the largest clients could afford gets delivered to the tier below them at a price they’ll pay. The authors see it coming in financial advice, healthcare, legal services, and education.
For a law firm or an engineering consultancy, that’s more matters and bids won, more of each client’s work, and referrals. Faster work on its own doesn’t win business, and for a firm that bills by the hour, it takes a week off the invoice.
Why growth needs a hybrid approach.
Each of those levers puts AI’s output in front of a client or prospect. A large language model on its own gives you the most likely answer and can’t show you where it came from. It’s right often enough to rely on and wrong just often enough to cost you the client. That’s fine behind the scenes, where a person checks the draft before it goes anywhere. It isn’t enough for the work that brings in revenue.
Franklin Discovery™ by entigenlogic® is a hybrid built for that side. A large language model handles the language. The finding and reasoning happen on a concept map entigenlogic has built over many years, with each idea written down and connected to related ideas. A firm’s own documents are read into that map, and every answer comes back with the passages it was found in. Ask the same question twice, and the answers match. When the answer isn’t in the firm’s documents, it says so. You’re not hoping it’s right. You can check that it is.
One question to add to the list.
The authors close with six questions for leaders, the first being whether the firm is focusing too much on cost savings and too little on growth. Add one: Which revenue work would you hand to AI if you could stand behind every line it produced?
See how Franklin Discovery works at franklindiscovery.ai.