There is one data point I would pin to the wall of every HR office. In late 2025, Gartner asked CHROs how ready they are to embed AI into the way they work, and 88% said they are not, at least not fully. In the same study, only 18% admitted they had extracted significant or transformative business value from AI. On top of that, just 10% of senior leadership believe their CHRO has the AI knowledge the company will need in 2026 and 2027.
Now compare that to adoption speed. SHRM's State of AI in HR 2026 shows AI usage in HR functions jumped from 26% to over 40% in a year, with recruiting as the most common use case. The tools are everywhere, the value is not, and that is not a contradiction, it is a diagnosis: most HR teams skipped strategy and went straight to tooling.
The problem is proliferation. The answer is prioritization.
Dave Ulrich, whom the profession rightly calls the father of modern HR, summed this up in two words: proliferation versus prioritization. When every team pilots its own tool, you end up with a pile of disconnected pilots that look great in a demo and never reach the P&L. His point is simple and uncomfortable: the question is no longer which AI tool to buy, but where AI creates the most value for people inside and outside the company, and what you will therefore not do.
"Replace the mirror with a window. Stop only looking inward at engagement, and start looking outward at the customers and owners the business serves."
Ulrich estimates the profession is only 20 to 25% along its learning curve on using AI to create value, which means we are far earlier than we like to admit. His framework for 2026 to 2030, which he calls the tipping point for HR, sits on four shifts. First, HR is not about HR, it is about value for all stakeholders. Second, that value comes through human capability, which is the sum of talent, organization and leadership, not talent alone. Third, HR itself must upgrade its skills for that to happen. And only fourth comes AI, as the information layer that makes the first three possible. Notice where AI sits on that list: last, as an enabler, not the purpose.
Gartner's order: five steps you do not skip
The Gartner deck I have been analyzing breaks HR's AI strategy into five steps in a specific order, not because a diagram looks nice, but because each step feeds the next.
- Analyze AI trends and their impact on HR. Before you buy anything, run a gap analysis and see where you stand versus the market and where AI can realistically move the needle. The output is a short priorities memo, not a tool catalogue.
- Vision and ambition. This is the step everyone skips. Your vision must explain why AI matters for your HR function specifically, and it must be concrete enough that a recruiter and a payroll specialist both understand it. Calibrate it with the board before moving on.
- Maturity and readiness assessment. An honest look at where you are today, where you want to be, and which gaps are priority. This step determines the rest of the plan; without it you build on gut feel.
- Strategy document. Goals, roles, risks and an adoption plan, aligned with business and IT strategy. This is where the buy-versus-build decision lives, and where you ask the board for budget.
- Continuous evolution. A strategy that is not refreshed becomes dead paper. Define what triggers a review — a technology shift, a legal change, a business pivot — and commit to at least a yearly refresh.
One point I would underline: Gartner runs this process across six leaders, not the CHRO alone, but also the heads of recruiting, learning and development, HR technology and analytics. An AI strategy written in one head does not survive contact with reality.
Agentic AI: the next wave, and the next trap
If 2024 was the year of generative AI, 2026 is the year of agentic AI. The difference matters: generative AI writes you a job description, agentic AI plans and executes multi-step tasks on its own under human supervision. Per Gartner, 82% of HR leaders plan to introduce agentic AI within the next twelve months.
And now the uncomfortable part from the same source. Gartner predicts that more than 40% of agentic AI projects will be cancelled by 2027, which is the hype wave many are just wading into. ADP data shows a clear split by company size: agentic AI is used by roughly 48% of large companies, a quarter of mid-market, and only 4% of small businesses.
There are two reasons to slow down here. First, candidate trust is fragile: Greenhouse research shows that only about a quarter of candidates believe AI evaluates them fairly. If recruiting leads AI adoption and candidate trust drops, you are risking employer brand, not just efficiency. Second, regulation: from 2 August 2026, the EU AI Act's obligations for AI systems used in hiring kick in, with fines up to 15 million euros or 3% of global turnover. For anyone operating or planning to operate in the EU, this stops being a later problem.
What this means for companies without a Gartner licence
This whole framework is written for corporate CHROs with a budget, but the logic works the same for a forty-person company. In fact, it works better, because smaller teams can align strategy and execution faster.
"Do not pick a tool, pick a problem. Then measure what it costs you today, and only then ask whether AI pays it back."
- Pick the one HR process where you lose the most time or money — usually sourcing and screening in recruiting or onboarding admin — measure the current cost, then ask whether AI pays it back.
- Write the vision on one page, not thirty. Say clearly what you want from AI in HR over the next two years, why, and what you will not touch.
- Do not buy readiness, build it. The 88% who do not feel ready are not a tool problem, they are a skill problem. Before an agent runs candidate conversations, someone on the team has to understand what it does and where it fails.
- Keep a human at the critical points. For hiring decisions, performance ratings and exits, AI can prepare and propose, never adjudicate. This is not only ethics, it is also what protects you from the regulation that is coming.
Ulrich is right that AI does not change HR's purpose, it amplifies it. That purpose is still to put the right people in the right roles so they create value for the business and the people around it. AI is the best assistant our profession has ever been given, but an assistant works to a strategy, and the strategy is still ours to write. The 18% who are extracting value do not do it because they have a better tool. They do it because they wrote down where they are going first, and only then chose what to go with.
Sources
- Gartner, AI Roadmap for HR (2026 deck): CHRO readiness, realized value, AI budget and the five-step AI strategy framework.
- Dave Ulrich, Where HR Goes Next, HR.com, December 2025.
- Dave Ulrich, We are on the cusp of a human capability revolution, ServiceNow Workflow.
- Ulrich, Ray, Todd, Next Step in GenAI for Human Resources: Proliferation vs. Prioritization, Workforce Solutions Review.
- SHRM, State of AI in HR 2026.
- Pin, AI Adoption in Recruiting 2026 Industry Report.
- ADP, Top 6 HR Trends to Follow in 2026.
- AI in Hiring Statistics 2026 — candidate trust and EU AI Act.
