

| Regulatory context, not legal advice This article discusses the use of AI in recruitment, automated candidate assessment and UK data protection considerations, including the Information Commissioner’s Office (ICO) guidance and the Data (Use and Access) Act 2025. It provides practical talent acquisition perspectives rather than legal instructions. For specific decisions involving candidate data, automated decision-making or recruitment technology, a qualified specialist can advise you. |
A managing director once told me that her business had become exceptionally good at getting people through interviews.
The vacancy would go live on Monday. Applications would arrive by Wednesday. By the following week, three candidates would have been interviewed, and an offer would be on the table.
It looked like an efficient recruitment operation.
But six months later, she was still struggling with the same department. One employee had left during probation. Another was underperforming. The third was doing reasonably well, but the manager admitted that the person had been hired largely because they were available and interviewed confidently.
Her question was simple: why were we spending so much time recruiting without consistently getting the people we needed?
I have seen this pattern across UK businesses of different sizes. Recruitment activity is often mistaken for recruitment effectiveness. A busy hiring team does not automatically make good hiring decisions.
After more than 16 years in talent acquisition across the UK, Dubai, South Africa and Malaysia, I have learnt that the quality of a hire is heavily influenced by what happens before the first interview.
The hiring process needs to work as a connected business operation. Every stage should have a purpose, a decision to make and a clear owner.
Here is how I would structure it for a UK SME.
An effective hiring process moves from identifying a business need to attracting, assessing, selecting and successfully integrating the right person.
The sequence matters because every decision creates the conditions for the next one.
These are not ten administrative activities to tick off.
They are connected decision points.
For example, if the role requirements are unclear, screening becomes subjective. If screening is inconsistent, interviews become overloaded with unsuitable applicants. If interviews are poorly structured, the final decision can be driven by personality rather than evidence.
The result is a recruitment process that consumes management time without producing reliable hiring outcomes.
The CIPD’s Resourcing and Talent Planning Report 2024 provides useful context. Its survey of 1,016 UK-based HR and people professionals, conducted online between 4 and 22 April 2024, found that 69% reported increased competition for well-qualified talent and 56% said retaining talent had become more difficult.
The same research found that 41% of organisations that had selected candidates reported that new recruits always, mostly or sometimes resigned within their first 12 weeks.
Source: CIPD Resourcing and Talent Planning Report 2024
These findings demonstrate why the recruitment process should be designed around more than filling vacancies.
Start by identifying the business problem the new employee needs to solve, rather than immediately writing a job description.
This is one of the most frequently overlooked hiring process steps.
A department manager says they need another analyst. The business approves the vacancy. HR receives an old job description, updates the company logo and publishes the advert.
Nobody stops to ask whether the original role still reflects what the business needs.
Perhaps reporting requirements have changed. Perhaps automation has removed several administrative responsibilities. Perhaps the existing team has developed skills that could cover part of the workload.
Before opening a vacancy, I recommend a short role-requirement discussion with the hiring manager.
The distinction between essential and desirable criteria is particularly important.
A UK SME looking for a finance analyst might request five years of experience, a particular qualification and familiarity with three software packages. But the actual requirement could be someone who can reconcile financial data, identify discrepancies and communicate meaningful findings to management.
Those are different hiring specifications.
The first describes a preferred background. The second identifies the work that needs to be done.
A strong hiring process begins with the second.
A good job description explains the responsibilities, skills, working conditions and outcomes of the role clearly enough for relevant people to recognise their suitability.
A job description should not become a shopping list of every quality the hiring manager would ideally like to find.
For SMEs, this is particularly important because an unnecessarily restrictive advert can narrow the candidate pool before recruitment has even started.
I recommend building the document around five areas.
Role purpose: Explain why the position exists and where it sits within the organisation.
Key responsibilities: Describe the actual work, including the decisions and problems the employee will handle.
Essential skills: Identify the capabilities needed to perform the role effectively.
Success measures: Explain what satisfactory performance will look like during the first six to twelve months.
Employment proposition: Include salary information, working arrangements, location, reporting structure and relevant benefits.
One practical test is to give the draft to someone outside the hiring department.
Can they explain what the person will actually do?
If the answer is no, the advert needs more work.
AI can help draft responsibilities, simplify language and identify repeated requirements. However, the hiring manager should validate every criterion against the actual role.
An AI-generated job description can still reproduce poor assumptions. Technology does not correct an unclear business requirement simply by making the wording sound more professional.
Screen candidates against evidence of the skills and experience required for the role, rather than relying on superficial similarities with previous employees.
Screening is where many otherwise promising recruitment processes begin to lose quality.
Hiring managers often receive more applications than they have time to review properly. Under pressure, they begin looking for shortcuts.
University names, previous employers, job titles, years of experience and keyword matches can become convenient substitutes for assessing capability.
Some of these details may be relevant. None should automatically replace evidence of suitability.
I recommend developing a screening matrix before reviewing applications.
| Assessment criterion | What to look for |
| Project delivery | Evidence of managing projects from initiation to completion |
| Stakeholder management | Experience coordinating internal and external stakeholders |
| Budget responsibility | Evidence of financial planning and monitoring |
| Risk management | Examples of identifying and addressing project risks |
| Communication | Evidence of clear reporting and decision-making |
Each criterion should have an agreed assessment method.
This creates consistency between reviewers and makes it easier to explain why someone progresses.
AI can support this stage by extracting relevant information, organising applications and identifying evidence against predefined criteria.
However, automated screening introduces additional considerations.
The CIPD’s 2024 survey of 1,016 UK HR and people professionals, fielded between 4 and 22 April 2024, found that 31% of organisations used AI in recruitment, compared with 16% in 2022. Its findings also indicated that around four-fifths of SMEs were not yet using AI or machine learning in recruitment.
The opportunity is not simply to introduce more technology. It is to make the screening process more consistent while retaining meaningful human judgement.
Related reading: AI CV screening in the UK
I explore the practical considerations in my article on AI CV screening in the UK.
Use role-relevant assessments to establish whether candidates can demonstrate the capabilities they will need in the actual job.
A CV tells you what someone says they have done. An assessment gives you an opportunity to examine how they approach a particular task.
This distinction is important when hiring for specialist or commercially important positions.
Consider a UK SME recruiting a marketing manager.
Instead of relying entirely on previous job titles, the business could provide a short exercise asking candidates to review an existing campaign and recommend improvements.
For a finance role, a practical assessment might involve identifying discrepancies in a sample spreadsheet.
For a customer service position, candidates could be asked to explain how they would handle a difficult customer scenario.
The assessment should reflect genuine responsibilities rather than creating unnecessary work for candidates.
Keep the task relevant. Assess capabilities that are genuinely important to the position.
Define the evaluation criteria beforehand. Reviewers should understand what a strong response looks like before seeing the candidates’ submissions.
Apply consistent conditions. Give candidates comparable instructions, information and time allowances.
Assess the reasoning, not just the final answer. Understanding how someone approaches a problem can reveal more than the finished output.
Use the results alongside other evidence. A work sample should contribute to the overall decision rather than become an isolated pass-or-fail mechanism without context.
AI can assist with developing assessment questions, organising results and summarising evidence. Where AI is involved in assessing individuals, the business should also consider transparency, fairness and appropriate human oversight.
The objective is to make better decisions, not to introduce more assessment stages for the sake of appearing sophisticated.
Structured interviews produce more consistent evidence because candidates are assessed against the same core competencies and agreed evaluation criteria.
One of the most expensive mistakes I see is an interview process that changes depending on who is conducting it.
One hiring manager spends 45 minutes discussing technical experience. Another focuses on personality. A third asks candidates about their career ambitions.
When the interviewers eventually compare notes, they are not comparing like with like.
This creates unnecessary subjectivity.
A structured interview starts with the role requirements and translates them into questions designed to assess specific capabilities.
| Competency | Interview question |
| Problem-solving | Tell us about a difficult problem you resolved at work |
| Leadership | Describe a situation where you had to influence a team through change |
| Commercial awareness | How have you made decisions that improved business performance? |
| Stakeholder management | Tell us about a disagreement with an important stakeholder and how you handled it |
| Adaptability | Describe a time when priorities changed unexpectedly |
Interviewers should use the same core questions and a consistent scoring approach.
I also recommend documenting what different score levels mean. For a more consistent evaluation process, use our Interview Scorecard Template for UK SMEs to help interviewers assess candidates against predefined competencies, record evidence and compare scores objectively.
A score of five should represent clear, specific evidence of strong capability. A score of three might indicate acceptable evidence with some development areas. A score of one should reflect little or no relevant evidence.
Most importantly, interviewers should record evidence rather than impressions.
“Excellent personality” is an impression.
“Explained how they resolved a supplier dispute, renegotiated delivery terms and prevented a project delay” is evidence.
That difference improves the quality of the final hiring conversation.
It also makes it easier to identify where interviewers disagree and why.
The final decision should be based on the candidate’s overall evidence against the agreed role criteria, not on who made the strongest personal impression.
This is where consistency matters most.
A candidate might perform exceptionally well in an interview but demonstrate limited practical capability. Another might be less polished but provide stronger evidence of delivering the outcomes the business needs.
Neither should automatically win.
Before making an offer, bring together the assessment evidence.
The hiring manager should be able to explain why the selected candidate is the strongest appointment against the role requirements.
It is also worth discussing the realistic expectations of the position with the preferred candidate.
Salary, reporting arrangements, responsibilities, working patterns and performance expectations should be understood by both parties.
This conversation can prevent misunderstandings later.
If two candidates are closely matched, return to the original success criteria. Avoid introducing new preferences at the final stage simply because the decision feels difficult.
The selection decision should remain connected to the business need that initiated the recruitment process.
AI can improve administrative efficiency, organise recruitment information and support consistent assessment, while people remain responsible for meaningful hiring decisions.
For UK SMEs, the most useful AI applications are often found in repetitive activities.
Consider where time is currently being spent:
These are areas where appropriately selected technology can reduce administrative pressure.
However, there is an important distinction between using AI to support a decision and allowing an automated system to determine a candidate’s outcome.
The ICO’s March 2026 Recruitment Rewired report examined evidence from more than 30 UK employers who discussed their use of recruitment automation between March 2025 and January 2026.
The report highlights concerns around transparency, discrimination and the extent of meaningful human involvement in automated recruitment decisions.
The legal framework has also changed. Section 80 of the Data (Use and Access) Act 2025, which came into force on 5 February 2026, replaced the former Article 22 provisions of UK GDPR with Articles 22A–22D.
The new framework addresses automated processing, significant decisions and safeguards. The precise implications depend on the processing and decision involved.
For employers, the practical point is that introducing an AI tool does not remove the need to understand how it influences candidate outcomes.
Source: ICO Recruitment Rewired
A qualified data protection specialist can advise on your particular technology, data processing and decision-making arrangements.
My recommendation is straightforward: automate repetitive administration where appropriate, but retain accountable human review of the evidence and decisions that affect candidates.
The hiring process should continue beyond offer acceptance, because the quality of the employee’s first 90 days influences whether the appointment delivers its intended value.
This is the point where recruitment and retention become closely connected.
A candidate accepts an offer. HR sends the paperwork. The hiring manager assumes everything is progressing.
Then the first day arrives, and the employee discovers that their equipment is not ready, their manager is unavailable and nobody has explained what they are expected to achieve.
The recruitment team has completed its task. The employee experience has barely started.
The CIPD’s survey of 1,016 UK HR and people professionals, conducted from 4 to 22 April 2024, found that 27% of organisations that had selected candidates reported that new starters always, mostly or sometimes failed to arrive on their first day.
The same research found that 41% reported new recruits always, mostly or sometimes resigning within their first 12 weeks.
These are important signals for employers.
A successful appointment needs a planned transition from candidate to employee.
Before joining: Maintain communication, confirm arrangements and introduce the employee to their manager and team.
First week: Establish responsibilities, provide access to systems and explain how the organisation operates.
First 30 days: Agree priorities and provide regular feedback.
Days 31–60: Review progress, identify obstacles and address any gaps in training or support.
Days 61–90: Evaluate performance against the original expectations and discuss longer-term development.
The hiring manager should remain involved throughout.
Related reading: Employee onboarding for UK SMEs
For a more detailed approach, read my guide to employee onboarding for UK SMEs.
The objective is not simply to welcome someone into the business. It is to help them become effective, confident and connected to the team.
Measure hiring quality alongside speed and cost, so recruitment performance reflects the outcomes the business actually needs.
Many businesses know how long it takes to fill a vacancy. Fewer can explain whether their recruitment decisions are producing employees who perform well and stay.
I encourage SME owners and HR leaders to establish a small recruitment performance dashboard.
| Metric | What it tells you |
| Time to hire | How efficiently vacancies progress |
| Application-to-interview ratio | Whether attraction and screening are working |
| Interview-to-offer ratio | Whether shortlisted candidates meet expectations |
| Offer acceptance rate | Whether offers are competitive and expectations are aligned |
| First-90-day retention | Whether new appointments are staying |
| New-hire performance | Whether employees are delivering expected outcomes |
| Hiring manager feedback | Whether the recruitment process meets operational needs |
The CIPD’s Resourcing and Talent Planning Report 2024, based on 1,016 UK HR and people professionals surveyed between 4 and 22 April 2024, also highlights a gap in recruitment measurement.
Only 24% of organisations reported measuring the return on investment of their recruitment processes.
This matters because recruitment activity can look successful while the underlying business outcomes remain disappointing.
For example, reducing time to hire is not necessarily an improvement if the business subsequently experiences higher early attrition.
Likewise, a high offer acceptance rate does not automatically indicate that the selected candidates are performing well.
Review your results by role, department and hiring manager where the available data allows.
Look for repeated patterns.
Are particular roles producing higher early attrition? Are candidates withdrawing at a specific stage? Are some managers consistently struggling to identify suitable applicants?
Those questions help turn recruitment reporting into practical business decisions.
Start by examining the last five appointments and identifying where the process produced weak decisions or avoidable delays.
You do not need to replace your applicant tracking system or introduce an expensive recruitment platform to improve hiring quality.
Start with the process you already have.
Take five recent appointments and review each one.
Then identify the stage where the greatest weakness occurred.
If the business repeatedly attracts unsuitable applicants, review the job description and attraction strategy.
If good candidates are rejected during screening, examine the selection criteria and how they are applied.
If interviews produce inconsistent decisions, introduce structured questions and scoring.
If employees leave shortly after joining, examine the connection between recruitment promises, manager expectations and onboarding.
Make one improvement, monitor its effect and then address the next weakness.
This is how smaller organisations build a more mature recruitment operation without creating unnecessary bureaucracy.
After all, the objective is not to make hiring more complicated.
It is to make every hiring decision more useful to the business.
The strongest hiring process is one that consistently identifies the capability your organisation needs, selects people on evidence and gives them the conditions to succeed once they join.
That is how we move beyond reactive hiring.
And that is how UK SMEs begin building a future-ready workforce.
The seven broad stages are identifying the vacancy, defining the role, attracting candidates, screening applications, assessing candidates, selecting the preferred candidate and onboarding the successful applicant. Each stage should have clear ownership and objective decision criteria.
Defining the role and its success requirements is arguably the most important starting point. If employers do not understand the capabilities they need, subsequent screening and selection decisions are unlikely to identify the right person.
There is no universal timeline. The appropriate duration depends on role complexity, candidate availability, assessment requirements and internal decision-making. Employers should remove unnecessary delays without compromising assessment quality or candidate experience.
AI can support job description development, CV screening, interview scheduling, assessment administration and recruitment reporting. Employers should retain appropriate human oversight, particularly where automated systems influence decisions about individual candidates.
Use skills-based selection criteria, structured interviews, relevant practical assessments and consistent scoring. Review new-hire performance and early retention to establish whether recruitment decisions are producing the intended business outcomes.
| About Sabiha Sabiha is a Talent Acquisition Director, speaker and author with over 16 years of international hiring experience across the UK, Dubai, South Africa and Malaysia. Shortlisted for Best Career Coach UK by the CDI, she has helped over 300 businesses improve their hiring and retention strategies. Her forthcoming book with Trotman, How to Use AI to Win Talent and Retain People, explores how UK employers can build future-ready workforces through AI and human insight. |

Global Talent. Ethical AI. Strategic Hiring. Sustainable Retention.
WhatsApp us