Cognizant and Pearson's 2026 study: AI already does 37% of entry-level tasks in India. The tasks juniors learned on are vanishing. What L&D should build.
Key takeaways
- The same research reports that 94% of HR leaders expect AI to create entry-level roles that do not exist today, and 96% expect entry-level work to shift toward supervising or managing AI systems within five years (PNI).
- 60% of respondents say their L&D programs cannot keep pace with how fast AI is changing jobs, and 95% say middle managers are critical to making AI adoption work (StockTitan summary, Pearson).
- The hidden risk is not fewer junior jobs. It is juniors who never build the judgement that the old grunt work used to teach.
- The fix is deliberate: design the apprenticeship that work no longer provides by accident.
Last updated · Bodhih Insights team · reviewed by Siddharthan VGJ, Founder
Think about how you learned your job.
Not in a classroom. On the first rung: reconciling the spreadsheet, drafting the minutes, cleaning the data, writing the first-cut report that your manager tore apart. It was repetitive, a little boring, and it taught you more than any course did. Somewhere in those hours you learned what "good" looks like.
Now those are exactly the tasks AI does best. A 2026 study by Cognizant and Pearson found that AI is already performing about 37% of entry-level tasks in India. The efficiency is real. The side effect is quieter and much less discussed: the way juniors used to learn is being automated away.
A note on sources: figures come from Cognizant and Pearson's press materials and secondary coverage. The 37% India figure appears in the India press release; we could not independently confirm which region each bar in the report's charts represents, and we have not reviewed the full report. Check it before quoting externally.
The Old Deal: Learn by Doing the Small Stuff
For decades, careers in offices worked on an unwritten deal. A junior did the routine work, cheaply and imperfectly. In return, they absorbed context: how a client thinks, why a number looks wrong, what a clean deliverable feels like. Seniors reviewed the output and, without calling it training, taught.
Call it apprenticeship by task. It never appeared in a learning budget. It simply happened, because the tasks existed and someone had to do them.
When AI takes a third of those tasks, the output still gets produced. The apprenticeship does not. And because nobody ever wrote the apprenticeship down as a program, nobody notices it has gone missing.
The Paradox: Judging AI Output Needs the Skill AI Replaced
Here is the uncomfortable loop.
The Cognizant and Pearson respondents expect entry-level roles to shift toward supervising AI. But supervising AI means spotting a plausible-but-wrong answer, and that takes experience that comes from having done the work by hand. A new joiner who has never built a model, drafted a contract clause or handled an angry customer has little to compare AI output against.
Our earlier posts describe the adult version of this problem: skills fade when AI does the work and AI output that looks finished but is not. For a junior, it is worse. You cannot lose a skill you never had the chance to build.
What the Same Study Says Employers Value Now
The reported numbers point to where the premium is moving. Around 97% of HR leaders in the study say soft skills matter more as AI spreads, and 67% say they now value liberal-arts backgrounds more (StockTitan). Read that plainly: when machines handle routine production, the scarce things are judgement, communication, curiosity and the ability to work with people.
Those are also the skills that are hardest to learn from a video and easiest to learn from a good manager. Which brings the weight onto a group that is already stretched. We covered that squeeze in the disengaged manager and in AI made India faster, and work heavier.
Five Ways to Rebuild the First Rung
- Map what juniors used to learn from each task. For every routine task now automated, ask: what did a person learn by doing it? Write it down. That list is your new curriculum.
- Keep some tasks human on purpose. Let new joiners do a first pass by hand, then compare it with the AI version. The comparison is the lesson. Treat it as a training exercise, not inefficiency.
- Teach "review before trust". Give freshers a simple checklist for AI output: is it accurate, complete, appropriate for this audience, and defensible if challenged? Practice on real examples with known errors.
- Turn managers into deliberate coaches. A manager who reviews AI-assisted work should explain why an output is weak, not just fix it. Short coaching habits matter more than long programs.
- Measure judgement, not just completion. Course completion tells you nothing about whether a junior can spot a bad answer. Use scenario-based assessments and skill checks to find the real gaps, as we argued in the proof gap.
A Quick Test for Your Organisation
Ask three questions this week:
- If you removed AI tomorrow, could your two-year-experience people still do the core work?
- Can your newest joiners explain why an AI answer is right, not just that it looks right?
- Who, specifically, is responsible for teaching them? If the honest answer is "whoever happens to sit next to them", you have an apprenticeship gap.
How Bodhih Helps
Bodhih builds programs for the skills this shift makes valuable.
- Campus to Corporate Training prepares fresh hires for workplace expectations, communication and collaboration, including working alongside AI tools.
- AI Literacy Training teaches people to use AI with judgement, including how to check and improve its output.
- Coaching and Mentoring Skills and First-Time Manager Training help managers teach through review, feedback and conversation.
- Problem Solving and Decision Making builds the reasoning habits that AI supervision depends on.
- Leading Human + AI Teams Training helps leaders design work so that people keep learning as tasks change.
- Skill gap analysis and assessments show where junior capability is real and where it is only assumed.
Corporate batches are private and tailored to your roles and tools. Open workshops are fixed-date public batches.
The Bigger Idea
Every generation of tools has removed some of the work that taught the previous generation. Calculators, spreadsheets and search engines each did it. Organisations adapted by changing what they taught.
AI is the first tool to remove work at the entry of careers at scale, in the same years it asks juniors to supervise it. If nothing replaces the old apprenticeship, companies will find in five years that they have plenty of AI and very few people who know when it is wrong.
The organisations that avoid this will not be the ones with the most automation. They will be the ones that decided, on purpose, how the next generation would learn.
Frequently asked questions
What did the Cognizant and Pearson study find about India?
The June 2026 India release reported that about 37% of entry-level tasks in India are already performed by AI, compared with roughly one-third globally. The study surveyed 750 HR leaders across the US, UK and India (Cognizant).
Will AI eliminate entry-level jobs?
The study's respondents mostly do not expect that. About 94% of HR leaders expect AI to create new entry-level roles, and 96% expect entry-level work to shift toward supervising or managing AI (PNI). The concern is how people build skills in those roles.
Why does automating entry-level tasks create a training problem?
Routine tasks were how juniors learned context and judgement. When AI does them, the learning that came with them disappears unless organisations replace it with deliberate practice, coaching and feedback.
What skills matter most for freshers in an AI workplace?
Judgement in checking AI output, clear communication, problem solving, curiosity and collaboration. About 97% of HR leaders in the study said soft skills matter more as AI adoption grows.
How can L&D teams respond?
Map the learning that automated tasks used to provide, keep selected tasks human as practice, teach structured review of AI output, equip managers as coaches, and assess judgement rather than course completion.
Where do the facts in this article come from?
From the Cognizant and Pearson study The AI Workforce Pulse: The Adaptability Imperative, as reported in the Cognizant press release, IT Brief, PNI, StockTitan and Pearson. We have not reviewed the full report.
