What reward leaders can learn from Davos: the missing data foundation

Investing better in people, particularly through benefits, from wellbeing to the overall employee experience, requires a stronger data foundation. And when it comes to human and benefits data, that foundation is still missing in most organisations.

World leaders, CEOs, and policymakers gathered in Davos in January to debate the forces reshaping the global economy. AI, productivity, cost pressure, and geopolitical uncertainty dominated the agenda. But one theme cut across all of them: people.

It is telling that one of the World Economic Forum’s five core themes this year centered on a deceptively simple question: How can organisations invest better in people?

At the same time, the Financial Times hosted a session focused on improving AI performance by closing the human data gap.

On the surface, these conversations appeared to tackle different challenges. One was about people investment; the other was about technology.For many organisations, they are the same problem.

Your AI is only as good as your data 

Without a trusted data foundation, even the most advanced AI cannot deliver strategic value. Yet many organisations are racing to deploy AI across global benefits, layering new tools on top of data they do not fully own, govern, or trust.

That data is typically spread across local brokers, regional systems, PDFs, and spreadsheets. It sits across countries, languages, and vendors. It is fragmented, inconsistent, and often controlled by third parties rather than the organisation itself.

Leadership expects fast, confident answers. HR teams are left navigating blind spots.

Origin’s Global Benefits Intelligence research, based on insights from over 500 senior HR and Reward leaders at multinational organisations, highlights the scale of the issue:

• 82% are concerned about the lack of visibility into their global benefits inventory
• 48% struggle to compile a complete global overview of benefits data
• 40% cannot confidently validate their benefits spend

The uncomfortable truth is this: applying AI to fragmented benefits data is not transformation. It is acceleration without direction.

Questions AI can't answer without a data foundation

Consider some fundamental questions every global organisation should be able to answer with confidence:

• What is our total benefits exposure across all markets?
• Where are costs rising fastest, and why?
• Are we meeting minimum health insurance coverage requirements in Italy?
• Which benefits deliver the strongest return for retention, wellbeing, and engagement?

If answering these questions takes weeks, relies on estimates, or requires manual work across multiple regions, that is a strong signal to focus on data foundations before layering AI on top of the existing structure.

In enterprise benefits management, critical information often sits with brokers, consultants, and insurers, scattered across PDFs, spreadsheets, portals, and contracts, frequently in multiple languages and formats. The result is a fragmented data landscape with no centralisation, structure, or governance.

For most HR functions today, aggregating this data soup into a verifiable single source of truth is extremely difficult if not impossible.

This is where, and why, many AI promises start to break down.

AI can only ever produce outputs as good as the data it is working with. When AI is applied to benefits data that is fragmented, unstructured, or simply inaccessible, the answers it generates simply mirror those weaknesses.

This is where the risk emerges. Certainty without truth (answers that sound confident and complete, but are built on data that is partial, outdated, or unverifiable). When leaders take those outputs at face value, decisions are made with confidence that cannot be defended. In a benefits context, where cost, compliance, and employee outcomes are tightly intertwined, the consequences of such decisions are high.

A tier-one employee experience is the visible tip of the iceberg. Beneath the surface sits the work that makes it possible: a tier-one system of record that is trusted, structured, and governed.

From spending to investing

So how do organisations invest better in people?

Invest with intent.

After payroll, benefits are typically the largest people-related investment an organisation makes. Yet many leaders still struggle to explain where that investment goes, what it delivers, or how it supports long-term business outcomes.

To invest better, and to use AI responsibly, organisations need to rethink how they approach benefits data. That starts with three shifts:

• First, treat benefits data with the same discipline as financial data. Finance teams can see spend, exposure, and performance in near real time. Benefits, one of the largest people investments, should be no different.
• Second, create a single, governed source of truth. Not spreadsheets that are outdated the moment they are shared. Not broker-controlled portals. A unified, structured view that the organisation owns and controls.
• Third, move from instinct to insight. When coverage, costs, utilisation, and outcomes are visible across all markets, decisions stop being reactive and start being strategic.

What leaders can do next

The conversations in Davos reflect what is happening inside organisations every day. The challenge is rarely recognising that something needs to change. It is knowing where to start.

Here are practical steps HR and benefits leaders are taking to close the gap:

• Get clear on where your benefits data actually lives. List every system, broker portal, spreadsheet, and document store involved. Note who controls access and how often information is updated.
• Measure how long it takes to answer basic questions. Simple queries about spend, coverage, or renewals are a useful test. If answers take weeks, that delay has a real cost.
• Identify where lack of visibility creates the greatest risk. This may be compliance exposure, renewal surprises, hidden broker fees, or an inability to explain cost changes to leadership.
• Start with one market or region where better data would change decisions. A complex or high-cost region often makes the value visible fastest.
• Apply the same standards you expect elsewhere in the business. If fragmented, unverifiable data would not be accepted in finance, it should not be accepted in benefits.
• Be honest about AI readiness. Before adding new tools, ask whether you trust the data they will rely on, and whether you can stand behind the answers they produce.

This is why Origin exists. We did not build Origin to add another layer of technology to an already complex ecosystem. We built it to solve the foundational problem first.

With a strong data foundation in place, the picture changes. Teams can answer critical questions quickly, plan with confidence, and lead at group level with decisions that stand up to scrutiny.

This blog was originally published on the REBA website. You can find the original here.