Only 12% of Companies Quantify AI Productivity Gains, Barclays Finds
Despite widespread AI adoption claims on earnings calls, most firms remain silent on measurable returns from their investments.
Nearly half of all public companies now discuss artificial intelligence on earnings calls, but only a small fraction are willing to put numbers behind their claims of productivity improvements, according to new research from Barclays.
Strategist Venu Krishna found that just 12% of companies mentioning AI were prepared to quantify the value they're capturing from the technology. Among those willing to share metrics, tech, financial services, and healthcare firms accounted for 68% of the total.
The measurement gap
The firms that did provide numbers reported operational efficiency gains averaging 56%. But Krishna noted these assessments "have not meaningfully changed over the last three years of LLM advances, calling their precision into question."
The disconnect is stark: companies want to appear "AI-enabled" but struggle to demonstrate actual deployment and execution at scale. Krishna described economically meaningful AI implementation as "uneven at best," with anecdotal examples of margin gains easier to find than systematic evidence.
Rare examples of quantified returns
A handful of companies have offered concrete figures. Meta reported that its AI tools drove an 8.3% increase in Facebook ad clicks and a 15.7% lift in conversions. Online pet retailer Chewy is targeting $50 million in cost savings for fiscal 2027 through AI deployed in fulfillment and customer service operations. Hilton expects AI integration to drive 75 to 100 basis points of margin expansion for hotel owners over time.
Beyond these examples, Krishna said, AI productivity winners remain scarce.
Why it matters
The measurement problem arrives as infrastructure spending accelerates dramatically. PwC's Global Data Centre Outlook projects that global AI infrastructure investment will reach $31.6 trillion through 2050, with annual data center capital expenditure climbing from roughly $800 billion in 2026 to $1.8 trillion by 2050. Hyperscalers including Meta, Microsoft, and Amazon are committing billions to build out AI capacity. At some point, companies will need to demonstrate returns that justify these investments—particularly as the spending begins to weigh on cash flow. The gap between AI rhetoric and measured business impact suggests many organizations are still in early experimentation phases, despite public positioning that implies mature deployment.
The findings were first reported by Yahoo Finance Executive Editor Brian Sozzi, based on Krishna's client note published Thursday.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call