Don’t Start from Scratch: How HR Can Use Its Own History to Put AI to Work

When HR leaders look to put AI to work in their organizations, knowing where to begin can be a challenge. But they don’t have to start from scratch. Their organizations’ own history can offer a useful starting point for identifying tasks where AI can make a difference, says Clayton Holz, head of product at Wisq.

Holz covered this important topic during a From Day One webinar titled, “Don’t Start from Scratch: How HR Can Use Its Own History to Put AI to Work.” Wisq is an AI platform for HR that uses the world’s first AI HR teammate known as Harper to automate repetitive HR work so employees can focus on more strategic priorities. 

Wisq recently surveyed approximately 100 senior HR leaders on AI and HR, and the top three responses were as follows: We don’t know where to start, we don’t have the capacity to set up and maintain it, and we’re worried about wrong answers and compliance risk. Holz shared what HR leaders can do when they face these first two issues.

How AI Can Review Historical Data

AI can support a range of HR responsibilities, from answering general organizational questions and providing personalized responses to taking action based on previous interactions and managing multistep processes, such as HR workflows. 

While many organizations use AI primarily for basic questions and answers, applying it to HR work requires going a step further by putting it to work on more complex tasks, says Holz. 

Getting started requires a few steps. Before analyzing ticket data or a shared inbox, HR leaders should get approval from their legal, privacy, and IT teams to export the data. Aim for 12 months of emails, including replies, to give AI enough information to identify patterns. There’s no need to spend time reformatting the data, since AI can work with a variety of file formats. From there, use the company-approved AI account, start with lower-risk content, remove any unnecessary information, and delete the raw export once the analysis is complete.

Once the data is obtained, there’s a process to follow. When using AI for this task, Holz emphasizes the importance of using the high or extra high thinking level of AI for the most comprehensive, accurate results.  

The first step is to clean and standardize the data, removing personal information such as names, employee IDs, and addresses. With a normalized dataset, HR leaders can examine work history to see where their teams spend the most time, classify common tasks, and rank them by effort. This helps identify where AI could make the biggest difference. 

Once HR leaders identify the tasks that require the most effort, they can build workflows to handle them. For example, a leave management workflow can outline the steps for reviewing, approving, or denying requests and apply that process consistently. Holz also pointed to job changes as an example. By analyzing conversations in the data related to job changes, AI can identify common reasons for requests and help build a workflow outlining the steps HR should follow.

AI can also help identify inconsistencies between written policies and actual HR practices. By comparing employee questions and concerns in emails and support tickets with the organization's policies, HR leaders can see where practices differ from what is documented. This comparison can help ensure policies are accurate and consistently applied.

“We can do this on a per inbox basis and look for things like unwritten rules, contradictions, gaps, inconsistencies, and more,” said Holz. “AI is not going to be perfect here. I don’t want to claim that this is a panacea and it’s going to catch every single issue with every single one of your policies, but it certainly can help be a starting point and kind of put you on the path to fixing some of these things.”

After HR leaders find out where they’re missing policy, the next step is to decide when a person should take over. This can be achieved by mapping every handoff with steps to complete along the way. Finally, create a definitive ask list for the IT department. Make it concise, create a priority list, and let them know what you need from their team. 

What to Do with AI Analysis Results

So, what can you do with the information? Even if you never deploy AI, you can use the AI analysis benefits to do the following: publish the answers in FAQs pages; streamline your organization workflows; clean up your policies; codify the handoff rules; provide the brief to your IT department; and automate the workflows.

Wisq was built to replace the work that used to be done by hand for their customers. Autopilot reads the entire history and keeps reading it while Harper handles the workflow in the systems. 

Holz also addressed how HR leaders can communicate the purpose of analyzing employee emails and support tickets, emphasizing that the goal is to streamline workflows and improve organizational efficiency, not audit individual employees. “Ultimately, I think it’s not about finding mistakes that people are making or casting blame, but it is about helping to offload some of the repetitive stuff that’s in there,” said Holz. “And my experience has been that that resonates with most people.”

Editor’s note: From Day One thanks our partner, Wisq, for sponsoring this webinar. 

Kristen Kwiatkowski is a professional freelance writer covering a wide array of industries, with a focus on food and beverage and business. Her work has been featured in the Bucks County Herald, Eater Philly, Edible Lehigh Valley, Cider Culture, and The Town Dish.

(Photo by Kindamorphic/iStock)