From streamlining recruitment to transforming leaders’ conversations, AI is helping Connex Community thought leaders reshape what it means to be a successful, efficient, and rewarding place to work.
Fridriksdottir is the EVP of People, Strategy, and Sustainability for Embla Medical, a medical mobility solutions business that includes the second-largest prosthetic company in the world. They only have approximately 4,000 employees despite operating within 36 different countries, meaning maximizing efficiency is crucial to their success. “Microsoft’s Copilot is probably my favorite commercial tool so far, and it’s something I use every day at work. We’ve trained our entire team to utilize it for repetitive tasks and to enhance brainstorming. It really allows us to do more strategic work,” she continued, “and it allows us to spend more time working with people rather than our data.”
As commonplace as Copilot has become, Embla’s embrace of it places them in a rapidly growing – but still relatively exclusive – club of AI adopters. According to SHRM surveys, nearly two-thirds of HR leaders say their organization is unaware of, still learning about, or outright avoiding AI tools. Only 28% of those same surveyed HR leaders have implemented AI’s most accessible use case, generative AI. This lags considerably behind the wide adoption of tools like ChatGPT by US workers, more than half of whom were already using GenAI tools for work-related tasks by September of 2023.
Lack of knowledge is a leading driver of AI hesitancy according to other surveys conducted that same year, which in the form of fear echoes some of what Fridriksdottir has experienced so far. “People are sometimes worried when they see all kinds of tasks being performed by AI in a few seconds that used to take days before,” she added. “But the reality is when you mix AI and the human mind, you get something powerful. The world continues to develop. I mean – I’m not sending faxes anymore,” Fridriksdottir joked. “Technology changes, as does the way that we do work, and [AI] is really groundbreaking technology.”
A Use Case for Every Problem
Global professional services firm Genpact understands that sentiment well, having committed to a particularly ambitious AI creation and deployment strategy across their various internal functions. “[Our work on AI] has two benefits,” explained Steve Woolwine, VP of Human Resources at Genpact. “It creates credible use cases that can be showcased to clients, and it helps us reimagine our support operating models with a more efficient vision.” Genpact provides a range of BPO, managed services, data-tech, and AI transformation services, meaning innovation is key to their organizational
DNA. Ahead of the pack, they’ve already deployed AI tools across the employee hire-to-retire lifecycle, all of which help improve org effectiveness and the employee experience. Starting at the beginning, Genpact is leveraging an AI-powered job matching engine to cross-reference applicant resumes within their database against job descriptions. This allows them to zero in on the most suitable candidates more quickly. Generative AI is similarly helping Genpact improve those very job descriptions, creating better, more accurate, and more equitable versions that can attract a diverse pool of talent. AI-led tools are integrated into the interview process as well, such as those that analyze video interview footage through a neuroscience lens to pick up on inflections in tone, facial expressions, and body language. “AI is even helping us generate competency maps that augment and guide our human recruiters’ judgments,” added Woolwine.
“Besides significantly improving the employee experience, AI enables us to rethink our HR operating model, further enhance productivity, and create more value for the business,” summarized Woolwine. “Furthermore, it’s also about doing things ‘right’.” Woolwine went on to explain how another digital assistant they’ve launched – AMBER – routinely interacts with all employees at predetermined intervals, asking questions that are relevant to that individual’s role, experience, and career path. “NLP within the tool gives us meaningful insights into employee needs,” clarified Woolwine, “while also flagging which employees have issues that need resolution through a direct meeting with HR or their manager.” That continuous read on employee sentiment enables Genpact to more quickly, meaningfully, and specifically address growing employee concerns before they balloon into much less manageable problems.
Augmenting the Human Experience
Paradoxically, one of AI’s best use cases within the workplace might just be its ability to humanize those sorts of one-on-one conversations. “In the moment, managers sometimes lack the vocabulary to have a meaningful discussion about performance or succession,” observed Michael Latsko.
Latsko is the Chief Human Resources Officer for Arizona State University (ASU), one of the nation’s largest public research institutions, and responsible for the education of more than 140,000 students. They’re fundamentally future-focused by design, so it’s no surprise that even their HR teams are piloting new, ambitious ways to push the envelope. Latsko is overseeing two AI capstone projects focused on the talent lifecycle, one of which endeavors to make it easier for managers to engage with their direct reports. “My thought was that AI could help create resources for managers to use that would enable more productive conversations with employees about how they’re doing,” explained Latsko, “and what their potential is. Are they on a path to a bigger, better, or just different job within the organization?”
When completed, ASU’s new tool will pull together a variety of inputs – such as teammate feedback, manager appraisals, previous discussions, and employee skillsets – to evaluate current performance and chart future growth opportunities. More importantly, managers will receive a tailored set of talking points that make it easier to discuss where the employee is headed, and why. “That’s usually the missing piece in a performance conversation – the employee starts to ask detailed questions, and without a concrete readout, the whole thing gets really subjective,” emphasized Latsko. “The hope is that this will make the whole process more objective, meaningful, and productive.”
“Not just that,” he continued, “but then we’ll use the same tool to examine all the learning assets we have access to that might be relevant to that employee’s growth and advancement. We can say, ‘here are the skills you’re missing, here’s a pathway to improvement, and here’s all the learning assets you need to fill in those gaps.’ So, managers will get a script for a more meaningful and actionable performance conversation with employees, who can then go off and start following the roadmap they’ve been provided.”
Embla Medical and Fridriksdottir have successfully experimented with a similar approach. “We use AI coaching to help leaders through conversations on difficult topics,” added Fridriksdottir. “It gives them a chatbot to practice with, so they don’t have to risk upsetting a real person as they learn. It’s a game changer,” she lauded, “and it’s always available.”
Genpact has also seen success with AI-enabled leadership, giving Woolwine and his team plenty to look forward to. “We’re using what we call HRPedia,” explained Woolwine, “which combines the power of GenAI with our own internal knowledge base of process, policies, tools, and the like. It’s been helpful for information retrieval, and more importantly, for just-in-time coaching and guidance on nuanced topics such as career pathing, networking, and learning
Overcoming Roadblocks
That success isn’t to imply that AI and its adoption have been without their fair share of problems, concerns, and challenges. Sometimes, those are as simple as tool developers failing to understand the needs and nuances of HR. “They’re often not HR professionals,” explained Latsko, “so when you talk to them about taking in information for something like the commonly used 9-box tool, there’s often a lot of confusion. You have to collaborate with them to clarify what you need, and again to then synthesize that down into a usable format for the layman employee or manager expected to use the tool.”
“There’s also concerns about data privacy,” clarified Latsko. “There’s a wealth of information in people’s emails, Slack messages, and the like – can those be mined for data points without violating privacy concerns?” Similar fears have been voiced in nearly every Connex Community discussion on AI and its use cases – if usable information across communication and data sources can be collected ethically, how must it then be stored, processed, and safeguarded to ensure one’s organization stays out of national headlines?
“Responsible use of AI is a guiding principle for us,” added Woolwine. “All of our AI and analytics initiatives go through rigorous legal, information security, and data privacy assessments to ensure we collect only relevant data. We also remove details that might negatively bias our resources,” continued Woolwine, “and we ensure there’s always a human in the loop to vet AI’s outputs before they’re approved for use.”
However, the biggest challenge can sometimes be those same humans – few people are routinely and consistently comfortable with rapid and extensive change. “We’ve done an excellent job upskilling and enabling our workforce on GenAI,” explained Woolwine, “but getting the right mindset to develop and adopt these tools is still a journey.” Woolwine cautioned that most of their tech and AI experiments don’t initially yield the intended returns simply because of that mental disconnect. “We sometimes approach AI projects with a traditional mindset – we expect perfection, but imperfections come with the probabilistic nature of AI.”
Taking the Leap
Despite AI adoption sometimes feeling like a truly daunting undertaking, all three of our interviewees adamantly recommended starting down the path as soon as possible. “Just start using it or, or you’re going to be left behind,” cautioned Fridriksdottir. “It’s already becoming very accessible, so try it out and work on asking the right prompts.”
“You definitely need to think carefully about your prompts,” Latsko agreed. “I remember asking ChatGPT a question that used the term ‘coaching’, and while anyone in HR would know I meant leadership or executive coaching, it returned information tailored to sports coaching. The lesson being, don’t get frustrated when you get the wrong answer,” Latsko summarized. “The answer is only as good as the prompt that goes in.”
Emphasis on prompt accuracy, clarity, and improvements aside, Woolwine added that HR leaders shouldn’t allow perfection to be the enemy of good. “Don’t treat AI tool implementations like software implementations. Get comfortable with launching an imperfect version to users so you can source their feedback and iterate,” said Woolwine.