Revolutionizing HR with Generative AI

Since exploding onto the scene in early 2023, generative AI (GenAI) has gone from a mere curiosity to a technology commonly used throughout organizations to save time, boost efficiency and enhance product quality. Human resources has been affected dramatically, rapidly moving from experimentation to growing adoption of GenAI for a wide range of HR tasks, including service delivery, onboarding, recruiting, learning and development, performance management and survey design.

Recent research from Gartner found that nearly well over a third of HR leaders (38 percent) are now either piloting, planning implementation or have already implemented GenAI in their departments—up significantly from just 19 percent in June 2023.

A 2024 joint study from SHRM and The Burning Glass Institute, Generative Artificial Intelligence and the Workforce, found that 26 percent of HR professionals now work in departments that use AI, up from 15 percent in 2022. Another 28 percent (up from 9 percent) are planning to apply AI in the foreseeable future.

To go beyond the statistics and explore the daily impact of GenAI on HR professionals, we gathered these stories of how different organizations of varying sizes are using artificial intelligence for common workplace functions. The following case studies offer a glimpse into these companies’ successes and lessons learned regarding the use of this still-embryonic technology.

GenAI’s Wide-Ranging Uses

When Karla Thommen, SHRM-SCP, needed to create a job description for a new position in her organization, she decided to bypass the traditional, time-consuming practice of stitching together research from a host of external and internal resources in favor of trying out a promising new technology she’d been learning about.

“The first thing I did was go to ChatGPT and say, ‘Provide me with a job description for a training and development manager,’ ” says Thommen, managing director of people operations for The Partners Group, a consulting, insurance brokerage and financial services company in Portland, Ore. “By using some follow-up prompts, I had a first draft of the job description within 20 minutes, and it gave me about 98 percent of what I needed. Without ChatGPT, that draft likely would have taken me a few hours.”

Thommen and her team have been using OpenAI’s ChatGPT and other GenAI tools to not only to save time on creating job descriptions, but also for a growing number of other tasks, including writing performance improvement plans for employees, creating drafts of new training programs, generating interview questions for job candidates, and helping craft messages celebrating staff milestones and anniversaries.

“For example, for an employee struggling with something like effective time management, I might go to ChatGPT and ask for suggestions of what specific training or new habits might look like to help managers coach the individual in improving that skill,” Thommen says.

Thommen also has used GenAI to create outlines for leadership development and change management training courses. “In most cases, it quickly produces an outline that is 90 percent right on in a first draft,” she says. “I use additional prompts to generate specific content ideas for each component of the outline.”

The technology also has had an unanticipated benefit: helping Thommen communicate more concisely. “I’ve been known for writing lengthy and involved emails, but by jotting down the basic thoughts I want to convey and pasting them into ChatGPT with a prompt of  ‘reduce the size of this by 40 percent and make it more succinct,’ it helps me create shorter, clearer communication,” Thommen says, noting she takes care to first scrub the content of any confidential or identifying company information.

Given the well-established risks of using public large language models (LLMs) like ChatGPT, Thommen’s organization has created a subcommittee to develop a company AI use policy, as well as a detailed tip sheet to ensure employees use the technology responsibly and ethically.

“The policy and tip sheet includes guidelines, such as, if you’re using the technology to craft a performance improvement plan, you obviously shouldn’t put employee names, the company name, intellectual property or other confidential data into the tool,” Thommen says.

The AI policy also stresses the importance of human review of ChatGPT’s outputs, she says, adding that users can’t just simply cut and paste content from ChatGPT. “The outputs must be read for accuracy, tone, intent and potential bias. At the end of the day, humans in HR are ultimately responsible for the words on the page or screen.”

Thommen and her staff learned to use ChatGPT by first setting up free accounts, experimenting with the tool and networking with HR peers in other organizations for their ideas on good use cases, how to write effective prompts and pitfalls to avoid. She’s helped her own team learn the technology by providing them with a list of best-practice prompts created by an external source for a wide variety of HR tasks.

Among Thommen’s own lessons learned from using ChatGPT is the value of providing the “human touch” to the technology’s outputs to avoid relying on generic content, and adding some originality to communication birthed by GenAI. “No employee likes to receive robotic messages or sentiments from their manager, for example,” Thommen says. “An unintended consequence of using AI can be subtracting from the employee experience and the relationship HR has with the workforce if we miss adding that personal touch.”

Reshaping Learning and Development

It’s not a stretch to say the arrival of GenAI has reshaped the learning and development strategy at Genpact, a global professional services firm based in New York City. Since adopting GenAI in early 2023, the learning function at Genpact has used it for a growing number of purposes, including creating learning assessments, developing curricula, building quizzes and condensing long-form learning content into bite-size modules, all hosted on its proprietary internal learning platform called Genome.

“About 10 percent of all new learning content on Genome is now created using generative AI,” says Shalini Modi—Genpact’s senior vice president and global leader of employee learning—who is based in Noida, Uttar Pradesh, India.

A new GenAI-based chatbot called AI Guru is also core to the company’s approach to learning and employee performance support. AI Guru answers employees’ learning-related questions, provides coaching on difficult performance issues and offers other services around the clock. The chatbot uses Microsoft Azure OpenAI infrastructure featuring GPT-3.5 Turbo. AI Guru is trained on the collective intelligence of Genpact’s internal experts, on over 15,000 learning content assets, and on the responses to thousands of employee questions captured over the past few years, Modi says.

“The GenAI-enabled approach allows us to implement a master-apprentice model at scale, and learners can strike up a conversation with AI Guru at any time to get answers to their skill-related queries,” Modi says.

The chatbot also functions as an easily accessible performance support tool when employees encounter challenges in the flow of their daily work, Modi says. For example, first-time managers have called on AI Guru for tips on how to effectively manage remote teams or how to give candid feedback to direct reports.

While Genpact does use some external resources for learning, it prioritizes leveraging internal expertise to help continuously educate its workforce. “We’ve identified 600 ‘Master Gurus’ throughout our organization to harvest their knowledge for our learning programs,” Modi says. “These are senior leaders who are experts on specific subjects.”

Master Gurus volunteer 5 percent to 10 percent of their time to build curriculum, curate content, teach classes and more. AI Guru was developed in part to replicate the expertise of these Master Gurus—but also to reduce the time required of them.

“We’ve combined AI tools for learning content with the AI Guru chatbot to potentially save up to 50 percent of time for our subject matter experts,” Modi says, noting that this frees those human experts to focus more of their time on strategic learning planning and activities.

Supporting Middle Managers

Not long after ChatGPT first became publicly available, Veronica Knuth, chief people officer at Quantum Health, a consumer health care navigation company based in Dublin, Ohio, held a meeting where she asked whether anyone on her HR staff had used the new technology.

“I only had one person out of 40 raise their hand,” she says. “So I put the challenge out that at every such future meeting, I wanted everyone to report back how they’d used ChatGPT in some fashion. I wanted people to experiment with it, learn it and get comfortable with it. Now, people on our team are using ChatGPT for a variety of HR tasks, and each meeting we share successes and lessons learned from that use.”

Like many of her HR peers, Knuth is using ChatGPT to save time on creating job descriptions. “It’s not uncommon to see hiring managers or HR business partners spending anywhere from two to four hours creating a job description,” she says. “But when I recently used ChatGPT to create a description for a new senior-level role we have open, it spit out a draft in about 30 seconds that I spent about 30 minutes refining. That’s a significant time savings.”

GenAI also proves adept at condensing or summarizing training content for other purposes, Knuth says. “You could put eight hours’ worth of training content into ChatGPT and ask it to provide a one-page overview of that content,” she explains, which is useful for marketing a class internally or educating executives inquiring about courses.

One of the top priorities for Knuth and her HR team at Quantum Health is providing support for line managers whose roles have grown more complex and challenging. “The role of middle manager has become so complicated and busy, that anything we in HR can do to support them with tools or resources to help save them time or create new ­efficiencies is a win for the company,” she says.

With that goal in mind, Knuth turned to ChatGPT to reduce the time it takes to develop resources that support managers in their daily work. For example, Knuth uses the technology to create coaching frameworks for managers to help them better manage their teams’ performance.

“ChatGPT quickly provides ideas and content for managers coaching employees struggling with particular issues,” Knuth says. “I might prompt the tool with ‘give me a positive coaching framework for employees who need to improve their communication skills,’ or request information on coaching individuals who may not be treating peers in a fair or respectful manner.”

Responses from ChatGPT can prove helpful in choosing the proper language and tone for interactions that otherwise may feel threatening to employees, Knuth says. “It provides clear and simple guidelines to serve as a starting point for difficult conversations managers or HR have to have with employees,” she says.

Knuth and her team distributed a list of 30 best-practice prompts for Quantum Health’s front-line managers to use on their own with ChatGPT in areas such as setting performance goals, constructive coaching, time management and more. Those lists were sent with accompanying guidelines about responsible and ethical use of the technology.

“Those best-practice prompts can be a tremendous help to people leaders who are new to their role or perhaps not as advanced in their leadership capabilities,” Knuth says. “But just like its use within HR, we stress not putting any sensitive or proprietary information into ChatGPT like employee names, specific titles, pay ranges and more.”

Taking HR Service Delivery to New Heights

Chatbots have long been integral to the employee experience at Publicis Sapient, a digital business transformation company based in Boston. These virtual assistants efficiently answer employee queries related to payroll issues, information technology, company policies and more, around the clock.

But when the organization decided to supercharge its chatbots with GenAI, the technology’s ability to answer HR-related questions went to the next level, says Kameshwari Rao, global chief people officer of Publicis Sapient. GenAI enables traditional chatbots to handle more complex or nuanced employee queries, and these enhanced bots can learn from the vast datasets that they analyze, and then iteratively improve their responses.

“With GenAI, we’re evolving our chatbots to new heights by providing more personalized answers tailored to individual needs,” Rao says. “We expect as our chatbots grow smarter each day, they’ll be a more indispensable resource for our hybrid workforce, enabling employees to perform more actions, such as filling timecards, scheduling meetings or submitting expenses.”

The company’s use of GenAI-infused chatbots for HR service delivery reflects a growing trend. The 2024 Gartner study found that the most frequent way HR leaders report using GenAI is for employee-facing chatbots in service delivery, with 43 percent reporting they prioritize chatbots’ use to improve how quickly or effectively they respond to workers’ HR-related questions or issues.

Rao’s HR function isn’t just using the technology to improve service delivery. GenAI has transformed aspects of the organization’s recruiting process as well.

“Initial candidate screening is a time-consuming task for our talent acquisition teams,” she says. “GenAI now speeds up this process with customized pre-screening that’s aligned with job prerequisites. We also leverage the technology for candidate sourcing and creation of comprehensive job descriptions.”

The use of AI enables the company’s recruiters to more quickly assess candidates’ qualifications and job fit, Rao says, as algorithms analyze and screen resumes to identify potential candidates with the desired skill sets.

Publicis Sapient also deploys its own internally built GenAI tools to enhance the capabilities of its proprietary learning and performance improvement platform called Marcel. The platform leverages AI to aggregate and organize vast amounts of internal data and knowledge within the organization. “That allows our employees to more easily access relevant information and insights, enabling better decision-making,” Rao says.

The Marcel platform also uses AI to connect employees in the organization across different agencies, geographies and areas of expertise, Rao says. “That use promotes collaboration, enabling teams to work together more seamlessly, share resources and leverage our collective intelligence.”

Lastly, AI embedded in its learning platform helps Publicis Sapient identify critical skills gaps in the workforce and recommend training programs to fill those gaps, providing personalized learning paths and suggesting relevant training materials based on individual employee profiles.

GenAI has grown so central to operations at Publicis Sapient that most of the workforce has taken formal training in how to use the next-generation technology, Rao says. That instruction covers basics like prompt engineering, but also places an emphasis on using the technology responsibly and ethically, with more specialized GenAI training customized to unique job roles.

“The GenAI revolution requires skills transformation across the enterprise,” Rao says. “Whether someone is in HR, is a project manager, a coder or a data scientist, they need to understand and master the technology for the future.”

Dave Zielinski is a freelance business journalist in Minneapolis.