HR Has a P&L Problem. This Chemical Engineer Believes He Fixed It.

Manu Khetan has sat in enough boardrooms to know the moment the conversation dies. A CHRO has just finished presenting. The CFO leans back. And then comes the question nobody in HR can answer cleanly: what did we actually get for the millions we spent?

Khetan is the founder and chief executive officer of Rolling Arrays, a Singapore-based enterprise HR implementation firm he built from a one-man consultancy in 2009 into one of Southeast Asia’s most battle-tested HR technology practices. It was among the first SAP SuccessFactors partners in the region, with 150-200 large-scale deployments under its belt.

What’s surprising is that he is not an HR professional. He’s a chemical engineer who spent time at SAP, stumbled into HR technology through a freelance engagement, spotted a gap in the market for SAP HCM services in Singapore, and never left.

That outsider’s eye is exactly why what he’s built matters. “Whenever I asked CHROs after a couple of years — what is the return on investment on the few million dollars you paid — the answer was not consistent,” he says. “It was not crystal clear. It was not a concrete answer.”

That pattern, repeated across organization after organization, is what Khetan spent years trying to solve. The result is a framework he calls R7. And if it works the way he believes it does, it redraws the line between HR and the P&L permanently.

The three ways HR transformations die

Before getting to solutions, Khetan wants to be clear about the disease. Enterprise HR transformations don’t fail because of bad software. SuccessFactors, Workday, and Oracle HCM are mature, scalable, and increasingly capable platforms. The failure is upstream, structural, and almost entirely preventable.

He has watched the same three breakdowns play out across his career.

The first: organizations cannot articulate what they actually need before they go shopping. They know they want to “modernize HR.” They do not know whether their real problem is quality of hire, attrition in critical roles, statutory compliance across ten countries, or something else entirely. Without that specificity, no software selection is defensible, and no implementation has a measurable target.

The second: everything becomes a priority. “Customers basically want to do all P1, P2, P3 all together at once, and then the project basically elongates,” Khetan says. When every problem is urgent, transformation energy diffuses, and the project becomes an operational grind, with data migrated, systems integrated, and value never actually delivered.

The third is the most operationally painful: the wrong people show up. Process owners who don’t have time to define processes. Program managers who have never run a transformation at this scale. Data stewards who don’t exist. The software gets configured. The organization doesn’t change.

Underneath all three runs a structural accountability gap that no one in the industry likes to name out loud. Three parties (vendor, systems integrator, and customer) each carry different KPIs. The vendor ships the product; the integrator implements the transformation the customer wants. Nobody owns the outcome end to end, and the gap between those three positions is where projects go to die.

Talent as a supply chain: The R7 hypothesis

Khetan’s response to the ROI problem wasn’t a consulting methodology. It was a framework built on a single premise: what if talent operated exactly like a supply chain?

R7 maps the employee lifecycle across seven stages — Recruit, Rate, Retain, Redeploy, Redevelop, Release, Rehire — applied to a workforce segmentation of A, B and C players based on role fitment, not individual worth. That distinction matters to Khetan. The framework’s Redeploy and Redevelop stages exist precisely to act on that. “It’s when we are telling you that you are a B player, you may be an A player in another role,” he says.

The supply chain yields two board-level metrics: the percentage of A players in critical roles, and the average tenure of A players in those roles. From those two numbers flows a hypothesis Khetan is willing to defend in any CFO meeting.

Incoming A players in critical roles drive new customer acquisition: revenue. Retained A players in critical roles drive existing customer retention: profit. The talent decisions your HR team makes every quarter are either compounding or eroding both lines. The supply chain makes that visible.

“HR job is one thing only,” he says. “We increased the percentage of A players in critical roles, and we increased their tenure. And because of that, our customer acquisition and customer retention went up.”

That’s the answer to the CFO’s question. Khetan is now working with Professor Felix Oberholzer-Gee at Harvard Business School on empirical research into differentiated employee value propositions. It offers academic scaffolding for a framework that grew out of watching organizations fail to answer for their own investments.

AI readiness: The wrong question is being asked

Dozens of vendors now claim to offer autonomous HR workflows. Every major platform is embedding AI. And most enterprise buyers are evaluating those capabilities against assumptions formed years ago — which is precisely the problem.

“The way humans work is, we use software, we see what works and what doesn’t. But because of AI, whatever tool we are using is changing at a rapid pace,” Khetan says. “That assumption has to go.”

SuccessFactors, Workday, and Oracle HCM are all moving faster than their reputations suggest. Enterprise software will always trail the frontier, as it must be stable, auditable and responsible. But the gap is closing faster than most buyers realize. The danger, then, isn’t platform lag but that customers will stop checking.

The larger mindset shift is conceptual, not technical. Khetan’s challenge to every HR leader in the room: stop asking how AI can be applied to existing HR processes and start asking how AI enables fundamentally different ways to solve HR problems.

His example cuts through the noise. Instead of performance check-ins and pulse surveys, AI can now observe employee work directly, across meeting recordings, Slack, Notion, emails, and shared drives. It can then surface productivity patterns and skills in real time without a single form being filled in.

The foundation required to unlock that capability is simpler than most organizations expect. Get role definitions right and build a clean competency framework. “So that on 12 months of data, when the AI is advanced enough, AI can tell you this particular employee should be redeployed in this role because they have a better skill fitment.”

Khetan maps HR to 18 broad process categories, 73 subcategories, and 197 automation candidates. That’s 197 places where a properly deployed AI layer can replace manual overhead, with none of them accessible without the underlying structural work.

The coding tsunami and what it does to your HR vendor

The last thing Khetan wants to talk about is the one that keeps platform vendors up at night.

“There was a time when people were saying, ‘Is AI code real?’ But now that topic is already gone,” he says. “AI coding is for real.” One engineer with the right tools is, by his estimation, now capable of producing the output of ten, twenty or even thirty working alone. The math on what that does to software development economics is not subtle. It commoditizes code, and commoditized code means a tsunami of new HR applications, from specialist tools for performance and rewards to recognition and candidate assessment.

Some will be acquired. Others will be built natively inside the platform ecosystems. Rolling Arrays is already building AI agents on top of SuccessFactors using SAP’s cloud platform; Workday Extend enables the same model on its side. The pace, Khetan says, “will be very, very fast.”

His own workflow runs at roughly 90% AI-generated code, 10% human refinement with human in the loop, not human replaced.

The chemical engineer who spent fifteen years watching HR technology fail to prove its own value now thinks the tools to prove it finally exist. The supply chain is defined, the metrics are board-ready, and the AI layer is catching up.

What’s left is whether HR leaders will build the right foundations before the wave arrives — or spend another decade explaining, badly, what they got for the money.

Image credit: iStockphoto/Deagreez