By Lee Flanagan
HCLTech’s answer to India’s AI talent shortage is the Forward Deployed Engineer, a role built around sitting with the client and solving the business problem using AI and other advanced technologies. That is the tell. Nasscom’s own numbers explain why the distinction matters. More than 90 per cent of early-career technology professionals in India already use AI tools, according to a Nasscom report. Only 23 per cent qualify as what Nasscom calls ‘AI-native’ engineers, meaning they have the technical depth and independent problem-solving skills to build and deploy AI systems. Our read: that is the difference between prompting a tool and engineering one, and it is a gap most interview processes are not built to see. That gap between usage and depth is not a hiring pipeline problem. It is a screening problem, and Indian IT firms are about to spend years finding that out the hard way.
The 90 Per Cent Everyone Is Screening For
Nearly every junior candidate walking into an interview today can talk fluently about AI, cite the tools they have used, and pass a technical screen. Fluency with the tools is not the same skill as the ability to build and deploy the systems those tools sit inside. Nasscom estimates India could face a shortage of more than 600,000 AI professionals by 2027. We think that figure reflects a measurement problem as much as a genuine supply gap: screens built to test tool familiarity cannot distinguish a candidate who is close to the 23 per cent bar from one who is nowhere near it, and both end up counted inside the same shortfall. If your screen rewards familiarity with AI tools, you will keep passing the 90 per cent and missing the 23 per cent. We think that is a design flaw in your assessment, not a talent shortage.
Nearly 1.5 Million Graduates a Year Is Not the Constraint
India produces nearly 1.5 million engineering graduates every year, on paper more than enough to cover the country’s AI hiring needs. It is not, and as the Business Standard reporting notes, the problem has shifted from headcount to employability. Findability Sciences CEO Anand Mahurkar puts it plainly: “Talent is not the issue, the problem exists in the gap between academic education and corporate applications. Although graduates may be knowledgeable about machine learning or be good Python programmers, they do not have hands-on experience with AI development in practice.” Prof V N Rajasekharan Pillai, an elected Fellow of the Indian Academy of Sciences, said, “AI is evolving at a much faster rate than the traditional university curriculum revision cycles. By the time many programmes are updated, the industry has already moved to the next generation of tools and applications. Faculty development has not kept pace with this transformation, leaving many educators without adequate exposure to the latest AI tools and industry practices.” Competition for talent from global capability centres, startups and multinational firms adds cost, not clarity, but neither explains why the other 90 per cent, the tool users already applying, are the ones failing interviews rather than clearing them on an honestly lower bar. We read the split in Nasscom’s numbers as a cliff, not a gradient: fluent tool use on one side, technical depth on the other, almost nothing measured in between, and curriculum lag does not explain why your hiring process cannot tell who, among that 90 per cent, is close to that bar.
Wipro’s Skills-First Bet Is the Evidence
Wipro is increasingly weighting demonstrable capability over academic credentials, while Tata Consultancy Services has moved to smaller, targeted hiring built around AI-native talent and critical thinking rather than volume recruitment. These are not soft culture initiatives. They are companies admitting that their old filters, degree, pedigree, keyword match on a resume, cannot tell them who can actually do the job. When a major IT services firm changes how it screens rather than how many people it screens, we think the shortage was never about supply. Companies had the applicants. They lacked a way to tell them apart.
Why a Single Skill Test Cannot Surface This Profile
Quest Global’s Sonia Kutty describes the actual profile companies are chasing: “The most difficult roles to fill are those that combine AI expertise with deep industry and engineering domain knowledge. The challenge isn’t only technical proficiency in AI, but the ability to deploy AI solutions at enterprise scale while understanding complex engineering environments and client business contexts.” That is a three-part profile, and no single coding test can surface all three at once. ManpowerGroup’s Global Talent Shortage Survey 2026 found 82 per cent of employers globally report difficulty finding skilled talent, with AI model and application development, at 39 per cent, and AI literacy, at 38 per cent, ranking among the hardest capabilities to hire for. In our work with hiring teams, interviewers routinely ask candidates to describe an AI tool they have used, hear a confident answer, and move them forward without testing whether they can reason through a problem the tool cannot solve.
AI Hiring Is Rising. Screening Design Is Not.
Naukri’s JobSpeak report for June 2026 found AI hiring within India’s IT sector rose 16 per cent year on year even as overall IT job listings fell 3 per cent in the same period. That divergence shows real, growing demand for AI-specific capability. It says nothing about which employers are hiring for it well. Experts cited in the reporting estimate it could take three to five years to build an enterprise-ready AI workforce through reskilling and industry partnerships. That timeline is real for the structural fix. It is not an excuse to delay the screening fix, which is simply a decision about what your scorecards measure: applied judgment against an unfamiliar problem, not a rehearsed description of tool use. If you are still screening for AI-native talent by asking what tools someone has used, you are not short on candidates. You are short on a test that knows the difference between the 90 per cent and the 23 per cent.
Original reporting: Business Standard.
Frequently asked questions
What does ‘AI-native’ mean in Nasscom’s data, and why does the distinction matter for hiring?
Nasscom defines AI-native engineers as those with the technical depth and independent problem-solving skills to build and deploy AI systems, not just use AI tools. The distinction matters because more than 90 per cent of early-career tech professionals already use AI tools, but only 23 per cent meet the AI-native bar, so a screen that only checks tool familiarity will pass most candidates who do not actually qualify.
Does India’s AI talent shortfall mean TA teams should focus on expanding graduate pipelines?
India produces nearly 1.5 million engineering graduates a year, which should be more than enough volume, yet firms still struggle to fill AI roles. We read this as an employability and screening problem rather than a headcount problem, so expanding pipeline size alone will not close the gap.
Does curriculum lag in Indian universities explain the AI-native shortage on its own?
Curriculum lag explains why fewer graduates reach genuine AI-native depth, since Prof V N Rajasekharan Pillai notes that university courses and faculty training revise slower than AI itself evolves. It does not explain why hiring processes still cannot tell, among the tool users already applying, who is close to that depth and who is not.
How does Wipro’s skills-first hiring approach differ from traditional credential-based screening?
Wipro is placing greater emphasis on demonstrable capability rather than conventional academic credentials when hiring for AI roles, according to the reporting. We read this shift as evidence that fixing how you screen matters more than how many people you screen.
What should interviewers test for instead of AI tool familiarity when screening for AI-native roles?
Quest Global’s Sonia Kutty describes the hardest roles to fill as those combining AI expertise, engineering domain knowledge and business context. Interview panels should test a candidate’s applied judgment against an unfamiliar problem, not their confidence describing tools they have used.