| Last Week in HR AI: Issue #9 | Week of August 24, 2026 |
OpenAI data shows almost no connection between AI usage and revenue per employee.
Amdocs cut agency spend 90%. Goldman puts call center employment 39% below trend while Oracle's SEC filing names AI outright.
|
1 CNBC2 thenextweb.com3 UNLEASH4 Reworked
Ikona's Take on the past week (August 24, 2026) in HR + AI
|
IO
|
Ian O'Keefe
Co-founder & CEO, Ikona Analytics
|
Two things happened this week that should be read together. OpenAI published a 69-page report on enterprise ChatGPT adoption, and buried in it is the finding that there is no statistically significant relationship between how intensely a company uses AI and its revenue per employee. Meanwhile Goldman put numbers on the labor side: U.S. call center employment sitting 39% below trend, entry-level hiring slowing most sharply across 800-plus occupations, and Oracle telling the SEC in plain language that AI is why 21,000 people are gone. Disruption is measurable. Returns, at the aggregate level, are not.
The through-line I see is this: AI is repricing the labor market and the HCM vendor market faster than it is producing proven P&L returns, and the small group of companies actually capturing value got there by doing the historically boring-but-critical foundational data work first. Amdocs is the cleanest example in this issue. Their turnover analysis found that employees rejected for internal roles left four times faster than their peers. That is not an AI insight. That is a well-instrumented People Analytics function asking a good question. The AI came after, in the form of a skills-based mobility model, and the results (40% internal mobility, a 90% reduction in agency recruiting spend) are the kind a CHRO can bring to the Board and a CFO will actually credit. Jennifer Guitard in Nova Scotia is describing the same sequence from the other end: 2.4 million pages of employee files digitized, payroll modernization still years out, and a flat refusal to talk about AI before the data foundation is real.
So, the uncomfortable question for your HR AI agenda: if someone asked you to name the single business outcome your AI spend has moved this year, could you point to a number, or would you point to usage stats? Usage is not value. And while you are answering that, the Silver Lake and Workday talks are a reminder that the platforms underneath you are being re-architected whether or not your roadmap is ready. I would genuinely like to hear where you land on this. Reply and tell me what you are seeing inside your own function, or reach out if you want to dig in.
| The six to read | The stories that matter for the Office of HR — with our take on each. |
Fortune reporter Emily Forlini read OpenAI's 69-page enterprise ChatGPT adoption study and surfaced a finding the report itself soft-pedals: no statistically significant relationship between AI usage intensity and revenue per employee prior to adoption. The data also shows executives, the people setting AI strategy, use the tools least intensely, and that enterprise usage flattened in Q4 2025 before picking back up. Two of the academic co-authors were paid OpenAI contractors. Separately, OpenAI replaced its chief revenue officer after less than a year in the role.
Goldman Sachs research identifies specific developed-economy sectors where AI is already suppressing hiring, with call centers hit hardest: U.S. employment in the category sits 39% below trend. Software publishing, consulting, and advertising show similar drag. Across more than 800 occupations, entry-level roles have the steepest AI-linked slowdown in headcount growth. Broad AI adoption now runs 15% to 20% across major economies, led by France, the U.S. the Netherlands, and the U.K.
Also covered by: payscale.com
Oracle's headcount dropped from 162,000 to 141,000 over the past year, and the company named AI adoption as a cause directly in its SEC filing rather than using standard restructuring language. Reductions were concentrated in Oracle Health and legacy SaaS units while cloud and AI teams expanded. Restructuring costs came in at $1.84 billion and capital expenditure rose to $55.7 billion, driving negative free cash flow. Cloud revenue grew 39 percent over the same period.
Also covered by: ibtimes.sg
Reuters reported that Silver Lake is in talks to take Workday private, pushing shares up roughly 18% to a $51 billion valuation. Analysts read the move alongside the Dayforce buyout as a repricing of legacy HCM platforms for an AI-driven future rather than evidence that HR software is in decline. Workday's 2025 acquisitions of Paradox, Sana, Flowise, and Pipedream, combined with Aneel Bhusri's return as CEO and other leadership changes, point toward a significant AI re-architecture. The co-founders control 68% of votes, so no transaction proceeds without them.
Human Resources Director profiles Jennifer Guitard, Executive Lead for People and Payroll in Nova Scotia, who is running a multi-year digitization of 2.4 million pages of employee files alongside a payroll modernization program still years from completion, all during a hiring freeze. She argues that clean data and workforce digital literacy are prerequisites for any AI ambition. She positions HR as data steward rather than data owner, and says HR should lead transformation rather than support it.
UNLEASH profiles Amdocs, a 25,000-employee telecom services firm, as one of the small share of companies showing measurable P&L impact from AI. The company's VP of People Insights describes a turnover analysis that found employees rejected for internal roles left four times faster than peers, which drove a move to a skills-based internal mobility model built on Eightfold's Career Navigator. Reported outcomes include 40% internal mobility, a 90% reduction in agency recruiting spend, and 10% gains in skill proficiency. The company is now extending the approach toward agentic HR service delivery.
Also covered by: talentnews.substack.com
Most HR functions have plenty of system data and almost none of the tacit knowledge that explains how work actually gets done. That missing context layer is why generative AI projects stall in HR, and claiming ownership of it is the strategic move available to analytics leaders right now.
Guitard's line that HR should be the data steward rather than the data owner is the most useful sentence in this issue, because it names the posture that separates the 5% getting P&L returns from everyone else running pilots. This is a hard moment to lead through: the evidence is mixed, the vendors are in motion, and the pressure to show AI progress is real. The leaders who keep making deliberate, sequenced moves on data quality, data governance, and overall data foundations will be the ones with something to point to a year from now, while everyone else is still reporting adoption rates. Keep going.
If your team is wrestling with workforce intelligence, AI readiness, or a transformation that has to land, we'd love to help. Reach out and tell us what you're working on.
Get in touch →Last Week in HR AI
Practitioner perspectives and diagnostic frameworks, delivered the week they publish. Be the first to know when new research drops.
No spam, ever. Unsubscribe anytime.