| Last Week in HR AI: Issue #4 | Week of July 20, 2026 |
HR's newest hire: a planner for humans and AI agents
Ford rehired 300 engineers. Glean found a workday lost to babysitting AI.
|
1 cio.com2 hrtalk.in3 hrdive.com4 hrdive.com
Ikona's Take on the past week (July 20, 2026) in HR + AI
|
IO
|
Ian O'Keefe
Co-founder & CEO, Ikona Analytics
|
This week's stories all point to the same shift: the interesting HR AI question is no longer whether to adopt it, but who decides which work goes to a person and which goes to an agent, and how deliberately that decision gets made. Atlassian is hiring a Director of Capacity Planning for Human and AI to answer exactly that question for internal teams. Ford already lived the consequence of getting it wrong, rehiring over 300 engineers after AI quality control missed defects that needed human judgment. Glean's data shows what happens when the boundary is never set at all: workers spend nearly a full day a week checking AI output, and almost 70% admit submitting work they never reviewed.
Our take: Human and AI planning is becoming a core HR discipline, not an IT afterthought, and the CHROs who treat it that way will move faster than the ones waiting for a policy to settle the question for them. Atlassian's VP Alicia Lenart made a point worth sitting with: AI usage mandates tend to produce "compliance" rather than value. The organizations pulling ahead are the ones building the muscle to decide, workflow by workflow, where human judgment still earns its keep, and giving line managers real tools to make that call. Workday's move into autonomous agents for IT and travel tasks only raises the stakes; the infrastructure for agent-driven work is arriving faster than most HR functions have built muscle to direct it.
We'd love to hear where you are drawing that line in your own organization. Reply and tell us, or reach out if you want to work through it together.
| The six to read | The stories that matter for the Office of HR — with our take on each. |
Atlassian is creating a new Director of Capacity Planning for Human and AI role tasked with helping leaders decide which work should go to humans and which to AI agents. VP Alicia Lenart, who co-leads Atlassian's internal AI adoption, says the role will build tools and agent libraries but leaves the actual decisions about team composition to line managers. The company is pairing the role with data-foundation work and workflow redesign, and it is deliberately avoiding blanket AI usage mandates, which Lenart says trigger compliance behavior rather than genuine adoption.
Ford has rehired more than 300 experienced engineers after its AI-based quality control systems missed design and manufacturing defects that required human judgment to catch. A Ford VP said AI outcomes are only as good as the data and expertise behind them. The returning engineers now do design reviews and train the AI models rather than routine inspection work. Ford credits the correction with helping it rank highest among mainstream brands in the 2026 J.D. Power quality study and with cutting warranty costs, even as it continues expanding AI camera use on the production line.
A Glean survey of 6,000 digital workers across the US, UK, and Australia found that employees spend nearly a full workday every week checking AI outputs for errors. Almost 70% admitted to submitting AI-generated work they had not reviewed or could not defend if questioned. The survey found heavy AI use correlates with disengagement and active job-hunting. Within HR specifically, adoption sits at 90%, and roughly a third of respondents said AI now shapes hiring decisions.
Also covered by: AI may conceal growing ‘learning debt’ for fast-changing roles | HR Dive, Employers may soon find it harder to retain workers | HR Dive
Workday launched two new AI agents built on its recently acquired Sana platform, one for IT service tasks such as onboarding and access provisioning and another for corporate travel and expense management. Both agents operate inside Workday's existing security and governance framework. Industry analysis argues Workday's position as a system of record solves contextual data permissions that causes many enterprise AI agents to fail, and frames the real competitive question as whether Workday displaces adjacent point solutions like ServiceNow or standalone travel platforms.
A lawsuit filed against Meta alleges that its process for cutting roughly 8,000 jobs in May 2026 used internal AI tools, including its "Metamate" system and activity-tracking algorithms, to score and select employees, some of whom were on protected leave at the time. Meta denies that AI made the decisions. Plaintiffs cite violations of the FMLA and WARN Act and are seeking an independent audit of the scoring process. An accompanying analysis argues that speed from AI is not the same as fairness and lays out governance steps enterprises need before using AI in workforce decisions.
Hyundai's autoworker union in Ulsan, South Korea launched a partial strike after the company unveiled its Atlas humanoid robot on the production line, demanding that workers consent before robots are introduced to their work. The union is also negotiating a shift from hourly wages to fixed salaries and an extension of the retirement age to 65. Humanoid robot deployments are accelerating across the industry, including at Tesla, GM, BMW, Xiaomi, and Mitsubishi, and Korea already leads the world in industrial robot density.
Also covered by: Hyundai Motor Union Stages 4-Hour Strike, Clouding Pre-Vacation Deal - Seoul Economic Daily
The post argues that most people analytics functions have rich system-of-record data but lack the tacit knowledge layer describing how work actually gets done day to day, and that this invisible gap is why generative AI projects stall inside HR. It positions owning that context layer as the next strategic claim for analytics leaders.
Ford's own words capture the week best: AI outcomes depend on the data and expertise behind them, not the model alone. That is capacity planning in a single sentence, and it is the work every HR leader is now being asked to do in real time, without a settled playbook. The organizations making that call deliberately, rather than defaulting to a mandate or a shrug, are the ones who will look back on this stretch as the moment they got ahead. Keep making the deliberate move. It adds up faster than it feels like it does.
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.