HR Tech 2026 wrapped up last week at Mandalay Bay in Las Vegas. More than 7,000 HR and business leaders spent three days on this year's theme, From Change to Transformation.
If you walked the Expo Hall, you didn't need a program to spot the headline. AI was everywhere. But the conversation has moved on from last year. Fewer people were asking whether to use AI. More were asking what it should be allowed to do, who checks its work, and what it needs to know about the organization to be of any use at all.
Here are the five things we heard most, from the show floor, our own conversations and the announcements that came out during the week.
A year ago, most HR AI was a chat window. You asked a question and got an answer. This year, the launches were about agents that do the work: screening and scheduling candidates, checking payroll before it runs, flagging compliance issues and handling routine HR requests.
Workday introduced agents that run in the background across recruiting, payroll and finance, and bring a person in only when judgment is needed. SAP SuccessFactors built its second-half release around agentic AI. The show itself listed AI agents as one of the main categories among the new products debuting in the Expo Hall.
The takeaway: the question for HR is no longer "should we use AI?" It's "which work do we hand to an agent, and which stays with people?"
In our live demos, we showed what this looks like in practice: an agent flagging top performers who are at risk of resigning, and drafting retention plans for the CHRO to approve, with approved changes writing back to the system of record automatically.
Last year the keyword was "responsible AI." This year more vendors are describing how oversight actually works:
The takeaway: governance needs to be part of your platform, not just a policy document. When you evaluate an agent, ask to see the approvals, the logs and the permissions.
It's also how we built AI at ChartHop. Ask ChartHop (our AI assistant) and our AI Pro agents work inside the same permissions as the person using them, so an agent can't see or change anything that person couldn't. Changes need approval, and every action is logged.
Several vendor announcements brought HR data and actions into tools people already use every day. iCIMS put hiring decisions inside Microsoft Teams. UKG brought its capabilities into Claude. ADP added a way for employees to ask about their payroll through their employer's own AI agent.
That's good news for adoption. Nobody wants to log into another system. But it also means HR data is now reachable by general-purpose AI tools, not just HR software.
The takeaway: the more places your HR data can travel, the more it matters who's allowed to see what. Permissions are becoming the most important AI feature nobody talks about.
This is the thinking behind ChartHop's Access Guard: it gives role-based, field-level access to data, and the same permission model carries into Ask ChartHop and AI Pro. Agents work within the limits of the person using them, not a broader system account.
The most useful idea we heard all week was a simple one. Your HR data records what happened. What AI needs is the context to understand why, and what it means for the business.
On the show floor and in our booth conversations, we heard discussions about giving AI a fuller picture of the organization, and linking HR programs to business results. The common thread: an agent that only sees the HRIS can tell you the headcount went up. It can't tell you whether that hiring is paying off, because the revenue, the budget and the plan live somewhere else.
This came up in almost every conversation at our booth. Most HR leaders we met already have the data. It's just spread and siloed across the HRIS, the ATS, finance and the CRM. ChartHop connects the data from those systems, including Workday, SAP, ADP, Greenhouse and Salesforce, so people data and business data sit side by side. That's what makes it possible to answer questions like how manager effectiveness affects revenue.
The takeaway: AI is only as good as the picture it works from. Connecting people data with business data is what turns AI from a fast search tool into something you can make decisions with.
Fake candidates, deepfake interviews and AI-written applications showed up again and again on the show floor. A whole set of new products focused on verifying identity and catching fraud, and iCIMS announced a strategy to help recruiting teams spot fake applicants. Some new assessments now test how well candidates work with AI, not just whether they used it.
The takeaway: AI made applying easier for everyone, including people who shouldn't be in your pipeline. Expect identity checks to become a standard step in hiring.
Put the five together and a pattern shows up. Agents are ready to do real work. The industry is getting serious about guardrails. And everyone is running into the same limit: an agent can only be as useful as what it can see, and only as safe as what it's allowed to touch.
A few questions worth asking before your next AI decision:
We came to HR Tech with a simple message, and the week made us more sure of it. For years, HR has been told the goal is a seat at the table. The seat was never the problem. The problem is a table built on scattered information, where HR brings one set of numbers, Finance brings another, and decisions get made without the full story.
AI agents make that problem bigger, not smaller. An agent working from limited data just makes the same mistakes faster. What changes the outcome is connecting your people data and business data into a live model of the organization, with clear rules about who and what can see and change it. That's the work ChartHop does, alongside the systems you already run.
In practice, that means Ask ChartHop answers questions about your organization in plain language, AI Pro agents take on the follow-up work with a person approving each step, and People and Finance can model a reorg or hiring plan together, seeing cost, team sizes and capacity before anything is announced.
Thanks to everyone who stopped by Booth #6134, joined our live demo or took home a ChartHop bunny. If you didn't get the chance, we're happy to show you what we demoed: AI, agents, analytics and planning, built on your HR and business data, with the guardrails your team needs.
Next stop: Gartner HR Symposium/Xpo in Orlando. Find us at Booth 220.
ChartHop is the organizational intelligence company. We connect your people and business data into a living model of your organization, so you can see what's changing, plan what's next and act on it. More than 350 companies use ChartHop to run their organizations with the full picture.
Heading to HR Tech 2026 in Las Vegas? Our take on the show theme, the sessions worth your time, and where to find ChartHop's booth and live demo.
Heading to Gartner HR Symposium/Xpo in Orlando? Our picks for sessions on workforce planning, org design and AI in HR, plus where to find ChartHop at Booth 220.
AI agents that do real work, governance that got specific, and why context matters more than data. Our five takeaways from HR Tech 2026 in Las Vegas.
Heading to Gartner HR Symposium/Xpo in Orlando? Our picks for sessions on workforce planning, org design and AI in HR, plus where to find ChartHop at Booth 220.
Heading to HR Tech 2026 in Las Vegas? Our take on the show theme, the sessions worth your time, and where to find ChartHop's booth and live demo.
