Recruitment has always been a people business. Increasingly, it is also a data business.
Every hiring cycle generates hundreds of CVs, interview notes, assessments and decisions, much of it underutilised, while recruiters spend a disproportionate share of their time on administrative work. AI promises to change that. According to LinkedIn’s Future of Recruiting 2025 report, teams adopting AI tools reported saving approximately 20 per cent of their workweek.
But faster hiring is not the same as better hiring. The real opportunity is not to replace recruiters, but to free them from administrative burdens so they can focus on what actually matters: understanding requirements, evaluating candidates beyond their CVs, building relationships and making informed decisions.
The six stages and where AI fits
Workforce planning and job design
AI can draft and refresh job descriptions, flag exclusionary language and analyse internal data to forecast hiring demand. What it cannot do is know that the team is about to lose its only senior engineer, or that the hiring manager’s brief is actually describing two roles. That context lives in conversations, and understanding it remains the recruiter’s responsibility.
Sourcing
AI excels at semantic searches across talent pools, resurfacing past applicants and personalising outreach at scale. The risk is that sourcing models learn from historical data, and historical data often reflects historical bias. Automated outreach also has a credibility problem. Candidates increasingly recognise machine-written messages, and a flood of them can erode the employer brand rather than strengthen it.
Screening
This is where AI is most widely used and perhaps most misunderstood. It can parse CVs, rank applicants and handle routine screening questions efficiently.
But screening models can also penalise career gaps, non-traditional education and candidates from certain demographic groups.
There is a subtler problem. Candidates now use AI to rewrite their CVs to mirror job descriptions, creating an arms race in which two models effectively negotiate with each other. Meanwhile, recruiters are left with a stack of near-identical documents that reveal very little about who can actually do the job.
AI should sort and surface, not reject. Human oversight must remain central to decisions about which candidates progress and which do not.
Hiring engineers? Look beyond the CV
ProfoundIQ combines targeted sourcing with technical evaluation by former CTOs and senior engineering leaders, helping organisations identify candidates with the technical capabilities and practical experience their roles demand
Assessment and interviewing
Scheduling, structured interview guides, transcription and objective skills tests are all useful applications of AI. Tools that claim to infer personality or cultural fit from facial expressions or tone of voice are a different matter. The scientific basis for such claims remains contested, and their use raises concerns about fairness, privacy and candidate experience.
For senior roles, especially in engineering, assessing architectural judgement, problem-solving approaches and technical trade-offs requires experienced human evaluation. And the candidate is assessing the organisation just as closely.
Selection and offer
AI can consolidate feedback, benchmark compensation and flag candidates who may be at risk of declining an offer. But the hiring decision itself must remain a human and accountable act.
Negotiation is similarly human. A candidate weighing two offers is responding not merely to compensation, but to how valued they feel, how honestly their questions are answered and whether they trust the person on the other end of the phone.
Onboarding
AI can answer routine questions, generate personalised learning paths and automate paperwork. What it cannot do is ensure that a new hire feels expected and welcomed.
Whether their manager showed up on the first day, whether a colleague took them to lunch, or whether someone made the effort to help them settle in, these are the moments that shape the early employee experience. No AI assistant can substitute for them.
The risks that cut across every stage
Bias is not automatically eliminated when hiring decisions are supported by AI. When models learn from historical data without appropriate auditing and oversight, they can reproduce and even amplify existing inequalities.
Data privacy is another concern. In India, the Digital Personal Data Protection Act, 2023 and its implementation framework are bringing greater scrutiny to how candidate information is collected, processed and reused. Employers need to ensure that AI recruitment tools handle personal data lawfully, transparently and with appropriate safeguards.
International regulations add further complexity. The EU AI Act classifies certain AI systems used in recruitment as high-risk, while some US jurisdictions impose specific requirements on automated employment decision tools. Indian employers hiring for global organisations may also encounter these obligations through contractual arrangements.
Employers cannot assume that responsibility for discriminatory or unlawful hiring decisions rests solely with their technology vendors.
Candidate trust is equally important. An efficient recruitment process that feels impersonal may lose precisely the candidates who have other opportunities.
Measure quality, not just speed
A 50 per cent reduction in screening time means little if hiring managers subsequently spend more hours interviewing unsuitable candidates. The organisation has merely shifted the workload rather than eliminated it.
More meaningful measures include shortlist-to-interview conversion, interview-to-hire conversion and quality of hire, assessed through performance, retention and hiring-manager satisfaction after joining.
For specialised roles such as engineering, organisations should also examine how much technical interview time is required to identify a successful hire.
These measures help determine whether AI is improving the quality of recruitment decisions rather than simply increasing the speed at which applications are processed.
The pattern is consistent
At every stage, AI is strongest at work that is repetitive, high-volume and data-heavy. At every stage, the decisive moments remain human: understanding what the business really needs, deciding who deserves a second look despite an unconventional CV, and closing an offer through trust.
At ProfoundIQ, we have observed this closely in engineering hiring. AI improves sourcing and screening efficiency, but assessing whether a candidate can actually meet the technical demands of a role requires experienced judgement. That is why we involve former CTOs and senior engineering leaders in candidate evaluation, not to slow the process down, but to ensure that speed does not come at the cost of quality.
The future of recruitment will not be defined by how much organisations automate, but by how effectively they combine AI efficiency with human expertise.
Shrihari Shastry is the founder and consulting CTO at ProfoundIQ, a technology-consulting firm specialising in technology leadership hiring, engineering hiring and engineering staff augmentation. He works with businesses on technology strategy, technical evaluation and building high-performing engineering teams.
