Agentic AI is moving HR technology beyond answering employee questions and routing tickets to resolving complex workflows, guiding managers, and taking action on behalf of employees and organizations.
Wisq Co-founder and CEO Jim Barnett sat down with SHRM at the HR Tech 2026 conference in Las Vegas to discuss how HR leaders can identify opportunities for agentic AI, balance AI with human oversight, and ensure AI agents apply workplace policies consistently. He also shared his vision for how the technology will reshape HR’s role and capacity in the years ahead.
SHRM: What distinguishes an AI agent that can resolve an HR case from a chatbot that simply answers questions or routes tickets?
Barnett: It’s all about the reasoning and resolution. Relatively simple rules-based technology allows people to retrieve answers to very basic queries. Examples of these questions would be: “Do we have Presidents Day off?” Or “What’s my PTO balance?”
On the other extreme, there is technology like ours which handles the messy HR workflows. These workflows are complex, multiparty — sometimes originating in a shared inbox — long horizon workflows that require real reasoning. That is the shift we are seeing, from rules-based chatbots to agentic AI that reasons and resolves matters and takes actions on behalf of the employee or the company. It is extraordinary how capable the AI is at doing certain things. I believe that AI will ultimately allow HR to do the best work of their lives.
HR will be given the capacity they need to be successful. When we think about value, we think about capacity. Giving leverage back to the HR team. We think about care. Over half of one client’s HR inquiries come in after normal business hours. Providing immediate responses to a diverse set of questions is showing care for your workforce. We think about capability. We help guide managers through high-value experiences like performance reviews, promotion requests, and coaching a difficult employee. Then there’s compliance. There’s a lot of shadow HR happening in workplaces. The benefit of having an AI teammate is that it is there to make sure employees and managers can get compliance policies consistently applied.
SHRM: How can HR leaders identify which workflows are good candidates for agentic AI? Should they start with high-volume routine requests, or could more complex processes deliver greater value?
Barnett: When we started building workflows, we went to customers and asked them which should we build? They often didn’t know. Now, we first ask a customer if we can evaluate their HR tickets, emails, and policies, and use a tool to analyze that material. We then generate a report outlining the workflows that they should focus on first. It’s mostly about hours saved. And then we tell them which integrations they will need for those workflows, and how much IT time that will take. Then we build the first version of those workflows, test them, and go in and optimize them. When deciding which workflows to focus on, companies could consider high-volume workflows, or something like leave of absence, which may be lower volume but high in complexity and takes a lot of time. We see customers approach the opportunity both ways.
My recommendation is to start with the HR front door, so all employees can see the value of the tool very quickly and the organization is aware of HR’s impact immediately. When you talk about some of the more complex workflows, many employees will never make use of them, but they still provide high value to organizations.
The great majority of HR workflows can be assisted by agentic AI at this time. Think about the majority of HR requests that have historically led to tickets. A high volume of them can be resolved by AI, with some medium or less difficulty. With another portion of more complex tasks, you may not want AI involved. For example — harassment claims, safety issues, highly sensitive matters with a high cost for getting it wrong.
SHRM: How can an AI agent learn an organization’s policies and decision-making practices without also reproducing the bad stuff: the inconsistencies, outdated guidance, or bias in historical case data?
Barnett: Harper [Wisq’s AI HR agent] can detect some of the bad stuff. As Harper starts to handle inquiries, we monitor the responses. We look for knowledge gaps, and work with the customer to fill in those gaps. Harper also looks for inconsistencies in the employee handbook and across an organization’s policies, analyzing hundreds of documents. In addition, HR can actually see Harper’s conversations, including the reasoning and documents used to support decisions.
We also built an orchestration layer and certain algorithms to ensure that Harper stays within guardrails.
SHRM: Where should human oversight begin and end when AI agents handle sensitive HR cases? Which decisions can safely be automated, and which should always require human review?
Barnett: In a deskless worker organization with a high amount of turnover, there are a lot of corrective action and disciplinary issues. Harper can be very helpful in managing the intake. But when you’re talking about termination, you obviously want a human to review and make the decision.
Take time-and-attendance violations, for example. We have a manufacturing customer that told us that their HR generalists spend an inordinate amount of time on time-and-attendance issues. We built a time-and-attendance workflow that coaches the manager in those situations, instead of just writing people up. The context matters. Then there are the times when disciplinary action is needed, and Harper assists with the creation and dissemination of writeups. For anything that requires not just reasoning, but human judgment, we believe that having a human involved is the right thing to do.
There will be some job displacement. You won’t need people answering basic help desk questions. AI will handle 90% of what goes to the help desk over time. But when you think about other workflows, AI is assisting in providing a better employee experience, and much better compliance, so the impact is just as likely that people will get back time to focus on more important things.
SHRM: Where do you see this technology going in the next few years?
Barnett: We are hearing that HR teams want more control of their AI platforms. We have an administrative dashboard today that lets HR see all the inquiries coming in and all the cases being responded to. We’ve got analytics based on the cases.
We’ve also created a no-code workflow builder that we will release to our customers so they can adjust some of their workflows. It’s tricky, because we’re on the line for the technology. If a customer creates a workflow that wasn’t properly QA’d and an employee has a bad experience, we take responsibility for that. That’s why we will also QA those workflows after changes have been made. We expect a really close partnership with customers to make sure any workflows we’re not building are perfect.
Was this resource helpful?