AI value in HCM: Where to focus first

It’s well documented that early adopter "AI in HCM" organizations have been using AI (specifically Gen AI) for generating job descriptions, as well as other AI-powered software functionality for screening and matching job candidates, scheduling interviews, and sourcing passive candidates – and these are just a handful of the dozens of new capabilities AI in HCM affords. Moreover, within these progressive, arguably less risk-averse organizations, Learning and Development (L&D) teams are creating personalized training that reflects how each employee actually learns best and then tracking their progress through performance upticks. And personalization also extends to tailoring performance improvement plans and to having AI agents respond to questions and provide other types of support in a manner that takes into consideration an employee’s broader context at work and support history.

An abundance of potential AI use cases in HCM to pursue would seem like a universally good thing. So, what’s the problem?

The issue is not a lack of opportunity. It’s that many organizations are moving from experimentation to scale without a clear view of which AI initiatives are truly positioned to deliver business value.

In many cases, the same early adopter organizations were also the first to realize that deriving material value from AI depended largely on first assessing how ready they were to create these new value streams. They were trying to scale transformative new capabilities without clear success metrics, governance, or an updated operating model to ensure business benefit justified the investment… not to mention possible disruptions to processes or functions that had already been delivering value.

In short, AI in HCM adoption accelerated faster than oversight and surrounding process redesign. This unfortunately left companies needing to manage new business risks or challenges, not to mention the inflated expectations surrounding these new (often underperforming) AI investments.

Recent research helps explain why so many AI investments have struggled to move from promise to payoff.
 

What the data tells us

Deloitte research recently determined that true ROI from AI mostly comes when people and AI work in convergence, multiplying each other’s impact rather than just collaborating effectively. The research found that just 16% of business leaders interviewed stated their organizations have now fully re-designed roles, processes, and operating models to effectively integrate AI into daily work. The point is that an “AI-centered technology upgrade” alone rarely creates meaningful business impact. What matters more is embedding AI into redesigned roles, processes, and operating models as seamlessly as possible.

Boston Consulting Group (BCG) this year found that companies were poised to double their AI investment levels, moving from 0.8% to 1.7% of revenue. Additionally, more than 90% of CEOs stated they were committed to maintaining or increasing investment levels even if short-term ROI was falling short. This largely corresponds with the roughly 90% of CEOs who believe that AI will redefine what success looks like within their industry, but also believe it will likely take at least another couple of years.

That makes AI investment especially challenging: even the definition of success is still evolving, and many organizations are being asked to invest before the long-term shape of that value is fully clear.

Finally, recent research from Atlassian shows that “just 4% of organizations are seeing true ROI from AI. The remaining 96% are still struggling with value realization.”

For me, that raises an important question: What are the fortunate 4% doing differently?

In my view, and in line with years of observation in the HCM and HR technology market, measurable ROI tends to show up when organizational transformation and redesign happen first, or at least alongside the technology investment. That usually means starting with the desired business outcome and working backward to identify the people, process, governance, and technology changes required to support it.

Finally, with respect to managing the AI investment portfolio, this aspect of the AI experience for customer organizations (as well as for HR technology providers) is actually pretty similar to all other product and technology investment portfolios. This is because they often include the same two key elements:
 

• Diversifying to properly assess the value and business impact of various levels of investment, plus the ease and speed of adoption, and examining both of these across different types of AI use cases and delivery mechanisms

• Achieving a balance of near-and-longer-term investments, and the same on the returns side.

Setting out on the journey towards AI value: 3 key points

I recently saw on LinkedIn a review of the positions of the six largest consulting firms on deploying AI to see where they align on workforce transformation strategy. The results were not what I expected. As it turns out, these firms agreed on exactly one thing: AI transformation is a people problem, not a tech problem. Not much agreement after that.

Against that backdrop, three considerations stand out for organizations trying to turn AI activity into measurable HCM value.

1. Business and adoption readiness come before technology plans. Before expanding AI use cases, organizations need to ask whether the business is ready to absorb the change responsibly and effectively. The World Economic Forum warns that adoption is accelerating faster than oversight, creating new risks in autonomy, safety, system integration, and trust, which includes leadership and investment community confidence also eroding. Therefore, the risks for enterprises rushing to deploy this powerful technology are expanding, and the organization may be accountable for both employees and the cadre of autonomous agents. We are also now testing the notion that humans have the capability (from a skills and competency perspective) and the capacity (i.e., bandwidth) to manage agents in a way that prevents and mitigates the broad set of potential operating risks.

It would appear then that companies investing in both outcome-based models (based strictly on each newly designed and viewed business context) – but who also realize broadly deploying AI along with people resources is as much a risk management and governance topic as a software quality or fit-for-purpose topic.

2. Change readiness is often the overlooked factor in AI success. It sits at the front end of change management and determines whether new ways of working can actually take hold. This is usually a multidimensional issue where readiness aspects are not limited to the very visible concerns of skills needed and related gaps, but also to the presence of supportive attitudes. And another key dimension often overlooked, but nonetheless equally relevant, is the potential intersection with other strategic undertakings or events occurring at the same time within the organization. Common examples of other potential efforts underway that might help or hurt the one being assessed can be expanding to other regions or launching new partnerships.

3. The prioritization of HCM use cases that “make the invisible visible.” Some of the highest-value AI use cases in HCM are the ones that surface options, patterns, and tradeoffs leaders could not easily see before. For example, having real-time insights into the likely best option for addressing a resourcing or skills gap, knowing which skills are becoming more or less important as business plans or priorities change, and understanding the likely ramp-up times for someone with similar or related skills. To elaborate on this ‘best option’ point, since the other items are self-explanatory, we know there are usually five such options:

Moving forward

It’s a safe bet that many of the early adopter organizations that struggled with AI in HCM value realization (a percentage that’s likely coming down now from the 90+ level with each passing week as learning shapes better practices) have concluded that the lack of envisioned outcomes isn’t due to weak technology, but to a business not being fundamentally redesigned to leverage AI investments to the fullest. Moreover, the disappointing experiences of those early adopter organizations trying to scale AI capabilities and impacts without holding these investments to a clear ‘business value yardstick’ – one ideally applied within effective governance and redesigned operating models and contexts – are now influencing more organizations to get it right. Again, the transformative value comes when you don't just speed up an old process but, in many cases, eliminate it entirely and replace it with something fundamentally different.

I’ve already mentioned the lack of agreement among major consulting firms on the direction of AI deployment. Therefore, my recommendation to all organizations on the AI in HCM journey is to not get lost or overwhelmed in the never-ending stream of new research findings and takeaways since – as previously pointed out – no two operating contexts are exactly the same. Study your own organization: its context, business and adoption/change readiness, and ask, “What’s the invisible to be made visible?”

The first step is not asking where AI can be applied. It is asking where the organization is ready to create value from it. That is one of the most reliable paths to value realization from AI in HCM investments.