Most discussions about AI as an investment opportunity center on the same cast of characters: chipmakers, hyperscalers, and model developers. Their revenues are visible, their capital expenditures are enormous, and their stock charts dominate the headlines. The infrastructure thesis is well understood.
But for advisors managing client portfolios in the midst of the current AI cycle, the more difficult – and arguably more consequential – question sits one layer down: which companies buying and deploying this AI capacity will actually convert that spending into enterprise value? Research from Irrational Capital suggests that the answer to this question may depend less on the technology itself and more on the relationship between a company and its workforce.
AI does not install itself. It needs people willing to design it, train it, and trust it enough to use it honestly.
The COVID-19 pandemic offers a useful comparison. It arrived with a vengeance, forced sudden operational transformation, and separated companies that adapted from those that stumbled, often along lines that weren't visible in the prior year's financial statements. The COVID-19 pandemic was a more episodic transition versus the ongoing AI shift, but asks the same of the workforce: Do you trust the direction of your leadership and understand the vision for the future? Or do you fear it?
Irrational Capital analyzed more than 2 million unprompted employee reviews from S&P 500 companies going back to 2015. Using this data, the firm constructed its Human Capital Factor (HCF) score – a measure of the depth of trust, autonomy, and shared purpose between a company and its workforce. Companies entering the pandemic with high HCF scores outperformed their peers by approximately 1.3% on an annualized basis from March 2020 to May 2023.*
A case study from the HCF research illustrates the dynamic across sectors.
Over the period from February 2015 through February 2026, Microsoft generated an annualized total return of approximately 24.1%, compared with approximately 13.9% for Oracle – a gap of more than 10 percentage points annually, compounded over a decade. Throughout that period, Microsoft's HCF percentile score was consistently in the upper tier of its peer group, while Oracle's remained materially lower. Both companies are major AI participants. The return differential predates the current AI cycle and tracks a sustained difference in how each company managed its relationship with its workforce.
Irrational Capital's dataset now covers more than 2 million unprompted employee reviews from S&P 500 companies spanning 2015 to 2025. No survey question asked specifically about artificial intelligence. All of the AI-related data below reflects employees raising the subject themselves – organically, in open-ended responses.
Following the launch of ChatGPT in November 2022, unprompted mentions of AI in employee feedback nearly tripled across the S&P 500. Among companies most actively pursuing AI transformation, those mentions rose nearly 15-fold over three years.
To understand the qualitative character of those mentions, Irrational Capital constructed an Opportunity/Threat Ratio – comparing positive framing (innovation, growth, productivity) against negative framing (job loss, displacement, anxiety). The results varied significantly by sector. Employees in technology, finance, semiconductors, and telecom entered the AI cycle with relatively balanced sentiment, having worked alongside automation tools well before the current headline cycle. Healthcare, consumer goods, and manufacturing moved through an early-optimism phase before more cautious language set in. Utilities, real estate, and commodities started in threat-dominated territory and have only gradually shifted.
The counterintuitive finding: employees who feel secure – who trust their organization – adopt AI more enthusiastically than workers who feel pressured by it.
Where employees trust leadership and believe in a shared future with the company, Irrational Capital's analysis suggests they are more likely to train AI systems honestly, share judgment calls that distinguish adequate output from professional-quality output, and integrate new tools into how the business actually runs. Where that trust is absent, compliance tends to be weak.
That difference, compounded across thousands of small decisions, may determine whether an AI rollout produces differentiated results rather than simply adding cost.
Workforce sentiment tends to move before earnings do. In the companies Irrational Capital tracks, high-HCF firms are currently showing AI-related employee sentiment that skews more toward opportunity than threat. The trust dimensions underlying those scores are also holding more stable than at lower-HCF peers – a pattern that, in prior transition periods, preceded the financial divergence rather than followed it.
The practical implication is that AI positioning is no longer solely an infrastructure trade. The next increment of return may belong to the companies deploying the technology that enables workforce trust. This meaningful input will likely separate those likely to build enterprise value from those spending and perhaps disappointing.
For advisors evaluating AI exposure in client portfolios, workforce data anchored investment options appear to offer a signal that is finally observable. The evidence is mounting: how a company treats its people is emerging as one of the most reliable predictors of long-term returns and now there are options for access.
After more than three decades of building and leading technology, human capital and financial services related firms, David van Adelsberg co-founded Irrational Capital, an investment research and strategy development firm that developed and continuously researches an innovative, proprietary investing factor - the Human Capital Factor.
*Data shows annualized excess total return of the HCF “Top 100” Equal weighted index, compared against the S&P 500 from March 1, 2020-May 31, 2023. Data source: FactSet. Past performance is not indicative of future results.
New research tracking Form ADV disclosures finds registered investment advisors embracing artificial intelligence are growing headcount, not cutting it.
A new ACLI survey finds most middle-class Americans plan to rely on Medicare for long-term care, but that’s not covered!
The proposal to eliminate Rule 206(4)-5 would end a 16-year-old restriction that firms say has unfairly suppressed political speech.
A recently settled lawsuit over Vanguard’s first-ever acquisition named a current sales executive at Altruist, the RIA custodian that Vanguard just bought for $4.6 billion.
Advisors who lead with cost savings when pitching pooled employer 401(k) plans may be underselling — and misrepresenting — the fiduciary value.
Northern Trust’s Ken Lassner shows advisors how to convert volatility into after-tax portfolio gains
Dan Biagini of American Equity says the steady decline of pensions, longer lifespans and a reset in interest rates are rewriting how advisors build retirement income
