Diverse Shortlists Do Not Happen by Accident.
Good intentions are not enough. Building diverse candidate pools requires deliberate sourcing strategies and access to talent beyond the usual channels.
The Intention-Execution Gap in Diversity Hiring
Most recruitment agencies understand the business case for diversity. A 2023 survey found that 57% of UK businesses consider equality, diversity, and inclusion as strategic priorities. Yet the execution is moving in the wrong direction. HR Magazine reported in 2025 that 75% of UK employers do not use blind CV submissions, up from 53% in 2022. More than half of employers no longer specifically state an interest in diverse candidates in their job adverts.
The sourcing challenge is structural. Traditional recruitment channels tend to surface the same types of candidates from the same networks and institutions. When a consultant searches LinkedIn using standard filters, the algorithm surfaces profiles similar to those the consultant has engaged with before. This creates a reinforcing loop that narrows the candidate pool rather than expanding it.
The candidate side of the equation is equally important. Gem's diversity hiring research found that female candidates are 2.4 times less likely to receive outreach compared to male applicants. Meanwhile, 72% of jobseekers say that a respectful, inclusive workplace is one of the most important factors when applying for a job, according to HR Magazine. There is a clear disconnect between what candidates expect and what agencies deliver.
75%
of UK employers do not use blind CV submissions (up from 53% in 2022)
HR Magazine, 2025
2.4x
less likely for female candidates to receive recruiter outreach
Gem, Diversity Hiring Research
72%
of jobseekers prioritise inclusive, respectful workplaces
HR Magazine, 2025
How AI Changes the Process
AI does not solve bias by itself. However, it can be configured to actively counteract the patterns that narrow candidate pools. The CIPD found that 66% of UK organisations using AI in recruitment reported improved hiring efficiency, and part of that efficiency comes from reaching candidates that manual processes miss. For agencies committed to diversity, agentic sourcing workflows can scan professional networks, community groups, and alternative platforms to build candidate lists that reflect the full breadth of available talent.
Audit your current sourcing channels
AI analyses your existing shortlists and placements to identify patterns: which demographics are over-represented, which are absent, and where your sourcing channels create blind spots.
Expand the search perimeter
Instead of relying solely on LinkedIn and job boards, AI identifies candidates from professional associations, community networks, return-to-work programmes, and university alumni groups that standard searches miss.
Remove identifying information for screening
AI can redact names, photos, university names, and other identifying details from CVs before they reach the shortlisting stage. This helps ensure candidates are evaluated on skills and experience rather than background.
Monitor shortlist composition
Before a shortlist is sent to a client, AI checks the demographic balance against the available talent pool. If the shortlist skews in one direction, it flags the gap and suggests additional candidates to consider.
Track outcomes over time
AI measures diversity metrics across your placements: application-to-interview ratios, interview-to-offer ratios, and placement rates by demographic group. This data reveals where the funnel narrows and where intervention is needed.
The Numbers
3+ hours
saved per week
£175+
monthly saving
Based on Bullhorn GRID 2025 data: AI saves 4.5 hours/week on candidate searching. Diversity-specific sourcing represents a subset of total search time. Estimate 3 hours/week for agencies running dedicated diversity sourcing alongside standard pipelines. Monthly cost based on £30,000 average UK recruiter salary (£14.42/hr).
Frequently Asked Questions
Does AI introduce its own biases?
AI can reflect and amplify biases present in its training data. If an AI is trained on historical hiring data where certain demographics were under-represented, it may score those candidates lower. Mitigating this requires careful configuration, regular auditing of outputs, and human oversight of every shortlist. The ICO audited AI recruitment tools in 2023-2024 and issued 296 recommendations, many related to fairness and bias.
Is positive discrimination legal in the UK?
The Equality Act 2010 prohibits positive discrimination, meaning you cannot hire a less qualified candidate solely because of a protected characteristic. However, positive action is permitted: you can take steps to encourage applications from under-represented groups and, where two candidates are equally qualified, choose the one from an under-represented group. AI should support positive action, not positive discrimination.
How do we measure whether diversity sourcing is working?
Track the composition of your candidate pools, shortlists, and placements over time. Compare the demographic breakdown at each stage of the funnel. If diverse candidates are entering the pipeline but dropping out at interview stage, the problem is not sourcing. AI analytics can surface these patterns automatically.
What if our clients do not prioritise diversity?
Diverse shortlists benefit clients whether they explicitly request them or not. A broader candidate pool increases the probability of finding the best person for the role. Present diversity sourcing as a quality improvement, not an obligation. The data supports this: McKinsey has consistently found that companies in the top quartile for ethnic and gender diversity outperform their peers financially.
Can blind CV screening really reduce bias?
Studies show mixed results. Blind screening removes surface-level bias but does not address deeper structural issues. It works best as one component of a broader diversity strategy that includes expanded sourcing, structured interviews, and diverse interview panels. AI-powered blind screening is faster and more consistent than manual redaction.
What about the UK Responsible AI in Recruitment Guide?
Published by the Department for Science, Innovation and Technology in March 2024, this guide covers pre-procurement, during-procurement, and live operation of AI recruitment tools. It is written for non-technical audiences and provides a practical framework for deploying AI responsibly. Any agency using AI for diversity sourcing should review it.
See Where Your Agency Stands
Take our free AI Readiness Quiz and get a personalised score across 7 dimensions of AI adoption.