Insights & perspectives on modern recruitment

The Promise Hits a Wall

The numbers sound extraordinary.

Organisations using AI-powered hiring software report 85% faster screening and up to 70% savings in recruiter resources.

87% of companies now use AI somewhere in their recruiting process, up from 26% in 2024.

By almost every measure, AI adoption in recruitment is accelerating at a pace that would have seemed implausible three years ago.

So why are so many recruiting firms still not feeling it?

The answer, more often than not, has nothing to do with the AI itself.

36% of firms say they are not seeing productivity gains from AI because their data is a mess. That is the dividing line. AI does not work on incomplete candidate profiles, outdated job records, or inconsistent workflows.

The tools are not the bottleneck. The foundation underneath them is.

What "Bad Data" Actually Looks Like

It is tempting to imagine data quality as an enterprise IT problem, the kind of thing that requires a six-figure systems overhaul to fix. In practice, it looks much more familiar.

It is a candidate record in your ATS with a phone number from 2019 and no notes from the last three conversations. It is a job order where the role title has changed twice but the original brief was never updated. It is a shortlist that was emailed as a PDF attachment, leaving no searchable trace in your system. It is five years of placements recorded in a way that makes it impossible for any AI to draw meaningful patterns from them.

Deloitte's 2026 Global Human Capital Trends report highlights that 95% of executives worry about candidate skills and capability data accuracy, yet only 5% of organisations say they are making significant progress improving the quality and trustworthiness of their workforce data.

That gap between concern and action is where recruitment productivity quietly bleeds out.

The Compounding Cost

Poor data does not just hamper AI tools. It compounds across every part of the workflow.

The larger shift in 2026 is from individual AI tools toward AI embedded across a firm's entire operating model.

That shift only delivers value when the data running through that operating model is structured, consistent, and current. A fragmented database does not just slow down one tool; it undermines the entire system.

Top recruitment firms are doubling down on recruiter productivity, not by adding more hires, but by removing friction from how their existing team works.

The firms making the most progress understand that data hygiene is not a one-time cleanup project. It is an operational discipline, as fundamental to a modern recruitment business as any sourcing strategy.

There is also a less obvious cost: missed relationships. When candidate records are stale or incomplete, reverse marketing becomes guesswork. Spec campaigns go to the wrong people. Warm leads sit unrecognised in a database that looks like a graveyard. The intelligence that should be your competitive advantage becomes noise.

What Good Data Discipline Actually Requires

The good news is that building cleaner data habits does not require a new system. It requires a different relationship with the one you already have.

A few principles hold across almost every recruitment firm that has made meaningful progress here.

Structure at the point of entry. The cheapest time to capture quality data is when a candidate or client interaction first happens. Every call, every submission, every client response should be logged in a structured way, not buried in an email thread or left in someone's head.

Treat the database as a living asset. Candidate records decay fast. Someone who was actively looking six months ago may be settled and happy today. Regular, systematic re-engagement, even at a light-touch level, keeps your database reflecting reality rather than history.

Measure what your AI actually sees.

The right question to ask of any AI tool is whether it has access to complete candidate context across sourcing, applications, and interviews, or only sees limited data passed through integrations, and whether it understands job requirements contextually or just matches keywords.

If the answer reveals gaps, those gaps point directly to where your data needs attention.

AI is increasingly absorbing the execution side of recruiting rather than replacing the judgment at its centre. As research and administration become cheaper, relationship building, assessment, persuasion, calibration, and advisory work become a larger part of the recruiter's value.

But that shift only benefits recruiters whose systems can handle the execution reliably, which means the data has to be there.

The Opportunity Hidden in the Problem

Firms that prioritised automation were 57% more likely to hit revenue targets last year.

That figure will only grow as AI capabilities deepen. But the firms capturing those gains are not just the ones that bought better tools. They are the ones that did the less glamorous work of making their existing data trustworthy enough for AI to act on.

This is genuinely good news for recruiters who are willing to treat data quality as a strategic priority rather than an IT task. Most of your competitors have the same tools. Fewer of them have clean, structured, current databases. That gap is closable, and closing it compounds quickly.

Platforms like Floats are built on the assumption that candidate data should be dynamic, shareable, and always current, because static records sitting in a PDF or a locked ATS field are the first place recruitment value goes to die.

The recruiters who pull ahead in the next 18 months will not necessarily be the ones with the most sophisticated AI. They will be the ones whose AI has something worth working with.