Wharton finds AI pays off more when you count growth instead of hours saved, so you can put the stopwatch down
By Brett Gavaghan, Founder of Adaptd ·
The 2026 Wharton-GBK AI study found strong returns were nearly three times as common in companies that measure AI against growth as in those tracking efficiency alone.
The short version
- The 2026 Wharton-GBK study found 80% of enterprise leaders see positive returns from AI, but only 28% call them strongly positive.
- Strong returns were reported by 35% of companies measuring AI against growth, compared with 13% of those measuring efficiency alone.
- Pick one growth number for each AI tool you pay for, such as quotes won or customer satisfaction, and track it.
Most AI success stories start with "it saves me two hours a week". Nice, but a new Wharton study suggests that's the wrong thing to be timing.
What did the Wharton AI study find?
The Wharton-GBK AI study found that companies measuring AI against growth are nearly three times as likely to report strong returns. Wharton Human-AI Research and GBK Collective released the fourth edition, The Adaptive Advantage, on 13 October 2026.
Of companies that track growth, such as net-new revenue, sales conversion, output volume or customer satisfaction, 35% report strongly positive returns. Among those measuring efficiency alone, it's 13%. Overall, 80% of leaders see positive returns, but only 28% say they're strong.
Who did they survey, and does it apply to small businesses?
The study surveyed 1,416 leaders and employees at large US companies, so it isn't a small business study. The results are self-reported, and they show links, not proof that one causes the other.
One finding travels well, though. 78% of leaders said their staff get adequate AI training, but only 56% of entry- and mid-level employees agreed. If you're the one who set the AI tools up, it's worth asking whoever uses them every day whether they'd agree with you.
How should your business measure AI?
A business should measure each AI tool against one result customers would notice, such as quotes sent, quotes won, reply times or repeat orders. Hours saved still count, but only if those hours go somewhere useful.
My take: "saves two hours a week" is how AI gets bought, and it's a fine start. But if those two hours disappear into the inbox, the tool hasn't earned its subscription yet. Pick one number per tool, write down where it sits now, and check it again in 90 days.
And ask your team how the training's going. You might find the bit that needs work isn't the AI.
Sources
Wondering what this means for your business?
That's exactly what the discovery call is for.