Do You Actually Trust Your People Data?

HR Technologies 2026: What the industry revealed, and why it still matters

HR professionals spend billions on people technology. They build workforce plans, deploy AI agents, and make high-stakes hiring decisions every day. And yet, when you ask how confident they are in the data underpinning all of it, the honest answer across organisation after organisation is the same: not very.

That was the thread running through HR Technologies 2026. Four sessions, different industries, different problems. The same uncomfortable question kept surfacing regardless of the topic or the speaker: how much can you actually trust your people data?

75% of CVs Contain Inaccuracies. You Are Still Trusting the Candidate to Tell You the Truth.

Start with the most basic act in hiring: verifying where someone worked and when. It sounds like a solved problem, but the reality for most HR teams tells a different story.

Traditional employment verification is built on a broken model: trust by assumption. Employers ask candidates what they did and where. They collect references from people the candidate chose. They wait and at the end of it, they still cannot be certain the picture they have is accurate. Research suggests up to 75% of CVs contain inaccuracies. The manual reference process was never designed to catch them.

The alternative is AI-powered Employment Verification. Rather than relying on what a previous employer is willing to say, it connects directly to verified payroll data: the most authoritative employment record that exists. Employer, dates, earnings, confirmed from source, with no employer sign-off required. The candidate consents and shares their own verified data, transparent and GDPR-compliant.

Every check produces a timestamped, audit-ready record. Employment verification that previously took days or weeks can now run in minutes, without sacrificing any rigour.

"The organisations winning on talent right now are those that have replaced friction with confidence. Faster decisions, zero document chasing, zero doubt."

James McBriar, Head of Sales, IDGateway

References still have a place. Insight from someone who has managed a candidate is valuable. But using manual reference collection as the primary mechanism for verifying employment history, when payroll-verified data already exists, is a process inherited from a previous era. The technology to replace it is already there.

The Skills Gap Costs £96 Billion a Year. The Real Cause Is a Data Problem.

Gordon McFarlane from the London Borough of Newham opened with a stark figure: the UK's skills gap costs £96 billion a year, and his argument was that the root cause is a data failure, not an HR strategy failure.

That figure does not come from vague inefficiency. It comes from the fact that people cannot accurately articulate their skills when moving between roles, and organisations rarely hold the data to help them. Skills go undocumented. Capabilities go unverified. By the time someone moves on, the institutional picture of who they are and what they can actually do has largely gone with them.

The argument was straightforward: human capital data is still far weaker than financial or operational data in most organisations. Job title and education level are no longer sufficient as hiring criteria. The gap is a data infrastructure problem as much as it is an HR strategy problem.

A practical note from the session: be honest about what you are actually measuring before you set targets. Social workers in Newham reported better work-life balance after AI managed their case note writing. The cost savings were smaller than projected. But the impact on the people doing the work was real. Define the problem first, then measure what changes.

AI Processed 1.7 Billion HR Transactions Last Year. Most Organisations Are Not Ready for It.

Richard Doherty, Senior Director of HCM Product at Workday, came with numbers that are difficult to ignore: 1.7 billion AI transactions processed in the last year alone, across more than 6,000 customers. One customer cut time-to-hire by 42% using a recruitment agent. These were not projections or pilots.

But the more useful contribution was not the scale. It was what Doherty said about what happens when you deploy AI without building structure around it first.

Frontier AI models trained on the internet know nothing about your organisation. Without a framework defining what AI can and cannot do in a given workflow, you are not accelerating good decisions. You are accelerating whatever quality of data you already have. The quality of your data determines the quality of every decision that follows. Better foundations, better outcomes. Weaker ones, and AI simply accelerates the errors you already have.

"AI in HR is not a future problem to plan for. It is a current decision about how much structure you are willing to put in place before something goes wrong."

Richard Doherty, Senior Director of HCM Product at Workday

That starts earlier in the process than most organisations realise. The data AI uses to make hiring decisions is only as reliable as the verification that fed it. If employment history was never properly confirmed, the model has no way of knowing. The capability is real. What you build around it determines whether it delivers value or creates liability.

One Company Lost Historical People Data It Could Never Get Back. The Warning Signs Were There.

The Hibob-moderated panel brought things back to ground level and was the most relatable session for anyone who has lived through a painful HR systems migration.

The warning signs that your HR technology is not working show up well before the crisis point. Wayve implemented an HRIS later than they should have and lost historical people data they could not recover. Grind found their hospitality platform could not serve their commercial and supply chain teams at all. In both cases, the signals were there beforehand: admin burden sitting entirely on the people team, managers quietly not using the system. Easy to overlook when you are busy. Much harder to fix once the data is gone.

Kenny Nicholson, Head of People Operations at Wayve, was direct about the underlying issue.

"Garbage in, garbage out."

Kenny Nicholson, Head of People Operations at Wayve

Build on shaky foundations and everything you add on top inherits those problems.

The Real Problem Is Not Your Technology. It Is the Data You Are Feeding It.

Four sessions covered different problems, different organisations, and different contexts. But the same answer sat underneath all of them.

As AI takes on more of the hiring process, the quality of your underlying people data becomes the deciding factor between decisions you can stand behind and decisions that come back to bite you. The skills gap is a data problem. AI governance is a data problem. HRIS failure is a data problem. And inaccurate employment verification? That is where the data problem starts, at the very first point of hire.

HMRC-sourced employment verification is not a marginal improvement on the status quo. It is a shift from a process built on assumption to one built on verified fact, at exactly the point in the hiring journey where accuracy matters most.

VettingGateway gives HR teams in regulated industries a single place to run DBS, Right to Work, HMRC employment verification, references, and FCA checks, with a clean audit trail for every hire.

If your background check process is still relying on people responding to emails, it is worth seeing what verified data looks like in practice.

Book a demo at vettinggateway.co.uk