Home / Learn

Assessing AI-written applications and candidate experience

Assess applications against evidence of the work the role requires. Similar wording or a close match to the job description cannot establish a person's ability or authorship. Use consistent follow-up questions, test relevant skills and add direct sourcing when the applicant pool leaves important parts of the market unexplored.

In short

  • Similar application wording gives reviewers less information about the differences between candidates' actual experience.
  • Ask candidates to explain specific work and assess their answers against the same criteria.
  • Combine advertising with direct research when suitable people are unlikely to apply themselves.

A well-written application can explain relevant experience clearly. It can also repeat the job description while saying little about what the applicant did. AI writing tools make that kind of tailoring easier, so a close verbal match gives a hiring manager only a starting point for assessment.

The useful question is whether the candidate can do the work. Answer it by connecting each important requirement to an example, a follow-up question and an appropriate assessment. Separately, check whether advertising reaches the people you need. Application quality and market coverage require different decisions.

Similar wording across applications

AI tools make it easier to adapt an application to the language of a job description. Candidates can use them to organise their experience, improve phrasing or generate claims with little detail behind them. Similar prose does not tell you which happened.

A 2025 BCG survey of UK business leaders found that many respondents felt confident spotting AI use while also finding selection harder. Those responses describe employers' perceptions; they do not test whether the employers identified authorship correctly.

For a reviewer, repeated wording creates a practical problem: several applications may appear equally close to the specification. Return to the evidence that could distinguish them. Look for the work each person owned, the context in which they did it and the result they can explain. These give the next conversation a purpose.

Authorship and evidence of experience

A CV that mirrors the job description gives you a reason to ask for detail, without proving who wrote it. Familiar phrases, polished grammar and a detector score are insufficient grounds for a hiring judgment on their own.

Research by Liang and colleagues in 2023 found that tested detectors misclassified some non-native English writing as AI-generated. The study concerns those tools and samples; it does not measure every current detector or establish a way to verify an individual CV.

Treat an unsupported claim as something to examine. If an application says the candidate led a migration, ask what moved, what they personally decided and what went wrong. Compare that explanation with the role's requirements. Cruvo researches candidate profiles and career evidence; its research service does not identify who wrote an applicant's CV.

A consistent application review

Use the same job-related criteria and follow-up questions for applicants to the same role. Define the evidence required for each criterion before reviewing applications, then record what is supported and what needs a conversation.

For example, a payments team may need someone who owned service reliability. “Improved platform performance” leaves important questions open. Ask the candidate to describe one incident, their responsibility and the change they made afterwards. Assess the answer for relevant experience and judgment, using the same standard across candidates.

A short, relevant work sample can provide further evidence. CIPD's guidance on selection methods supports structured assessment and clear information about what candidates should expect. Explain the task, time involved and permitted assistance in advance. If AI use is allowed, specify what the candidate must explain themselves so reviewers can assess their contribution.

Job adverts and external recruitment

A job advert is useful when suitable people are likely to see the role and apply. It gives interested candidates a way to approach you and can remain part of the search while the team investigates other channels.

Different methods of external recruitment reach people in different ways. Referrals draw on existing relationships. Direct sourcing identifies people from relevant companies and career histories. An agency can take on agreed parts of candidate search and engagement. Choose the method according to the market and the work your team can undertake.

Review the applications against the brief before deciding that the channel works. If few applicants show the necessary experience, consider whether the advert explains the role clearly and reaches the right audience. If suitable people are unlikely to be looking, add direct research. More applications will help only if they include people with the experience the role needs.

Research beyond the applicant pool

Direct sourcing lets the team investigate people whose experience fits, whether or not they applied. Start with the work the person must do, then identify companies and adjacent sectors where that experience could develop.

For example, a team hiring for payments infrastructure can investigate engineers responsible for similar systems across several industries. The research should explain each person's relevant experience and the questions that remain. Your team can then make an informed approach and assess interest in the role.

Cruvo is one way to obtain that research. It provides a reviewed, ranked candidate long-list with the reasoning behind each name, while your team handles outreach and hiring. The candidate sourcing guide compares delivery models. Use that distinction to decide whether the missing capacity is research, candidate engagement or assessment before choosing support.

Each requirement needs evidence that the team can explore and assess consistently across candidates.

Sources

  1. Boston Consulting Group. Title: UK Employers Seek Typos to Spot AI-Written Applications. Year: 2025. URL: https://www.bcg.com/publications/2025/uk-employers-seek-typos-and-grammar-errors-to-spot-ai-written-job-applications
  2. Liang and colleagues, arXiv. Title: GPT detectors are biased against non-native English writers. Year: 2023. URL: https://arxiv.org/abs/2304.02819
  3. CIPD. Title: Selection methods. Year: 2026. URL: https://prod.cipd.org/en/knowledge/factsheets/selection-factsheet/
  4. Cruvo. Title: Candidate sourcing services. Year: 2026. URL: https://www.getcruvo.com/candidate-sourcing-services

Talk about a hiring brief

If applications are leaving questions about the people you need, we can talk through the role and the search so far. We can discuss where candidate research could help, whether you have started hiring or are still considering the brief.

Or email hello@cruvo.app directly.