The Attention Landscape · emerging evidence

The Anxious Search: What India Types When the Job Market Turns

Last reviewed 2026-08-10. Written by Chandranshu Kumar, Founder, Raveneye Global. · 10 min read

When India's job market tightens, the searches do not run toward the AI course. They run toward the government exam. A reading of India's monthly search-volume data, pulled in August 2026, finds that queries for the core government competitive exams, SSC, UPSC, railway recruitment, the NDA exam, and bank exams, total roughly 6.11 million a month. The entire category of AI-and-data courses, from prompt engineering to machine learning to generative AI, totals about 250,400 a month, a category roughly 24 times smaller. The single term "sarkari naukri" (government job) alone draws 1 million searches a month, about four times the volume of every AI or data course search combined. Explicit anxiety about AI taking jobs is close to invisible in the data: "will ai take my job" draws 170 searches a month nationwide. The pattern reads less as a rejection of AI and more as a vote, cast in search bars, for the job that still promises security over the skill that promises adaptability.

The measured gap

Raveneye Global pulled a set of India-wide monthly search-volume readings from Google's Keyword Planner in August 2026, covering government-exam terms, job-search terms, AI and data course terms, and a handful of explicit AI-anxiety phrases. The core government-exam terms alone total a striking figure. "Ssc" draws about 4,090,000 searches a month, "upsc" about 1,830,000, "railway recruitment" about 90,500, "nda exam" about 60,500, and "bank exam" about 40,500. Added together, the five terms come to roughly 6,111,500 searches a month, call it 6.11 million, for exams that lead to a government post.

Set beside that is the entire category of AI-and-data course searches: "data science course" (74,000), "ai course" (60,500), "data analytics course" (33,100), "artificial intelligence course" (18,100), "python course" (18,100), "prompt engineering course" (18,100), "generative ai course" (14,800), "machine learning course" (12,100), and "chatgpt course" (1,600). Every one of those terms combined comes to about 250,400 searches a month. Divide the two totals and the government-exam category runs about 24 times the size of the entire AI-and-data course category. This is not a close race.

Two things are worth being precise about before drawing any conclusion. First, Google's keyword tooling reports search volume in rounded bands rather than exact counts, so every figure here is an approximation, not a headcount. Second, this is a single month, in one country, read once. It says where search demand sat in August 2026, not where it will sit in a year, and it says nothing directly about how many people actually sat an exam, enrolled in a course, or changed their mind.

Part of the gap is structural, and worth naming directly. The government-exam category concentrates decades of demand into five well-known, single terms: everyone preparing for a central recruitment exam searches the same handful of words, because there is only one SSC, one UPSC, one set of railway boards. The AI-and-data course category is younger and far more fragmented; searchers spread themselves across course names, platform names, and language variants that no nine-term list can fully capture. Some of the emerging category's smallness is a real gap in demand, and some of it is simply the tax of measuring a newer market with fewer converging terms. Both are true at once, and this reading cannot cleanly separate them.

The word doing the most work

One term carries an outsized share of the picture on its own. "Sarkari naukri," the Hindi phrase for "government job," draws about 1,000,000 searches a month by itself, roughly four times the volume of the entire AI-and-data course category combined. No single AI or data-skill term comes close to that scale; the closest, "data science course" at 74,000, is not even a tenth of it.

Widen the lens to general job-search terms and the same shape holds. "Work from home jobs" draws about 450,000 a month, "government jobs" about 368,000, "remote jobs" about 135,000, "part time jobs" about 74,000, the two variants of "fresher jobs" and "freshers jobs" together about 44,400, "it jobs" about 9,900, and "software jobs" about 1,900. Add those to sarkari naukri and the whole job-search category comes to just over 2,083,200 searches a month, more than eight times the size of the AI-and-data course category.

Two of those terms are worth sitting with. "Work from home jobs" and "government jobs" both outdraw every other job-search term by a wide margin, and neither is really about a specific role or an industry. They are about a working condition, flexible location or public-sector stability, sought before any particular job title. "It jobs" and "software jobs," by contrast, are two of the smallest terms in the whole dataset. That is not evidence the technology sector is shrinking; it is far more likely evidence that people already working in or targeting tech roles search by specific company, specific job board, or specific role title rather than by the generic category name, so a broad term like "it jobs" was never going to capture that behavior well.

The searches sitting next to those numbers are telling in a different way. "Resume format" draws about 450,000 a month on its own, the same volume as "work from home jobs" and roughly ten times the combined total of "upskilling" (9,900), "career change" (33,100), and "online certification" (210), which together add up to about 43,210. Put plainly, more people in India are searching for how to format a resume in a given month than are searching for how to change careers, get certified, or build a new skill, combined. The instinct visible in the data is to get hired now, not to retrain for later.

Why LinkedIn's 9.14 million searches are not counted here

"Linkedin" is, by a wide margin, the single largest term in this entire reading at about 9,140,000 searches a month, dwarfing every other figure in this piece. It is deliberately left out of the job-anxiety count. A search for "linkedin" is a navigational query: the person typing it wants to reach the LinkedIn website, most often to log in, not to express an unmet need about work or skills. Counting it alongside "sarkari naukri" or "government jobs" would blur a request to open a known site with a request to solve a problem, and the two are not the same signal. Excluding it is a conservative choice that, if anything, understates how much of India's online attention already sits inside job-related platforms.

Where AI shows up in the data, and how small it is

The AI-and-data course category is not zero, and it should not be read as zero. "Data science course" leads the group at 74,000 searches a month, followed by "ai course" at 60,500, "data analytics course" at 33,100, "artificial intelligence course" and "python course" tied at 18,100 each, "prompt engineering course" also at 18,100, "generative ai course" at 14,800, "machine learning course" at 12,100, and "chatgpt course" trailing at 1,600. Taken together these nine terms come to about 250,400 searches a month, a real and specific pocket of demand, just a small one against the scale of the government-exam category.

What sharpens the picture is what the demand for "learning" is actually doing when it does appear. Compare the AI-and-data course category to the coaching brands built around competitive-exam preparation: "physics wallah" draws about 1,500,000 searches a month, roughly six times the entire AI-and-data course category by itself. "Unacademy" draws about 368,000, on its own larger than every AI or data course search added together. "Byju's" draws about 110,000, smaller than the category total but still a meaningful slice of it. India's search-driven appetite for coaching and preparation is real and large; it is simply pointed almost entirely at the government exam, not at the AI skill.

Read next to each other, the two learning categories tell a coherent story rather than a contradiction. The country's search demand for structured preparation exists at real scale. It is funneled toward passing a fixed, known, credentialed exam that ends in a government post, not toward an open-ended, less legible skill that promises adaptability without a guaranteed outcome at the end of it.

The shape of the two categories differs as much as their size. The exam and coaching terms are a small number of very large words: five exam names carrying 6.11 million searches, three coaching brands carrying nearly two million more. The AI-and-data course terms are a larger number of much smaller words spread thin across nine separate names, none of which individually breaks 75,000. A concentrated few large searches against a spread of many small ones is itself a piece of information: the government-job path is a known, singular destination that the whole country searches for by the same names, while the AI-skill path is still a scattered, unsettled category without one obvious term everyone converges on yet.

The anxiety that barely registers

If AI-driven job displacement were a subject actively worrying large numbers of Indian searchers, it should show up somewhere in the query data. It barely does. "Will ai take my job" draws about 170 searches a month nationwide. "Jobs after ai" draws about 30. "Ai proof jobs" draws about 390. Added together, the three explicit AI-anxiety phrases total roughly 590 searches a month, against a core government-exam total of 6,111,500. The exam category outweighs the openly stated AI-fear category by well over ten thousand to one.

This does not mean the anxiety is absent from people's lives; it means it is not showing up in this particular, explicit phrasing. Search behavior captures what people type, not everything they feel. A person worried about being replaced by automation may never search the words "will ai take my job" at all. They may simply search for a government post, a work-from-home role, or how to format a resume, because that is the practical next step their worry produces, not because the worry itself went unsearched.

Read that way, the near-absence of explicit AI-fear queries and the scale of the government-exam and job-search categories are not two separate findings. They are the same finding seen from different angles. When work feels uncertain, the reflex visible in India's search behavior is not to search for reassurance about AI specifically. It is to search for the job that still comes with a formal guarantee attached to it.

There is also a plainer possibility worth stating rather than dismissing: AI-driven job worry may simply live somewhere search does not reach. A discussion in a WhatsApp family group, a comment thread on a news video, a conversation at a coaching center, none of that shows up as a typed query. This piece can only speak to what this dataset actually measured, a set of search terms in one month, and it would be overreach to claim the near-zero figures prove the worry itself is near-zero. What the data supports is narrower: whatever the worry is or is not, it is not, at this scale, translating into people typing the words "will ai take my job" into a search box.

What this reading does and doesn't show

The direction of this finding is well supported. It is corroborated from two independent angles in the same dataset: the sheer scale gap between the government-exam category and the AI-and-data course category, and the near-total absence of explicit AI-fear phrasing set against millions of exam and job searches. Two separate cuts of the data point the same way, which is a stronger basis for the direction than any single figure on its own.

The specific numbers are a different matter and are marked accordingly. Every figure in this piece is a Google-banded estimate, drawn from one country, one month, and one reading of the tooling; a rerun next quarter, or a state-level rather than national-level pull, would very plausibly return different exact counts even if the shape held. The 24-to-1 ratio between the government-exam category and the AI course category, in particular, should be read as a rough order of magnitude, not a precise, load-bearing constant.

One claim genuinely sits in contested territory: whether this pattern is a reaction to AI specifically, or simply the continuation of a pull toward government employment that predates generative AI by decades. India's government-exam economy, and the enormous coaching industry built around it, existed at scale long before "artificial intelligence course" was a search term anyone typed. The fairest reading is that AI-driven job anxiety, where it exists, is very likely being absorbed into a search habit that already existed for other reasons, rather than creating a new one from scratch.

It would also be unfair to read the small size of the AI-and-data course category as proof that the category does not matter. Course searches of this kind typically carry more commercial and decision-adjacent intent than a broad job-search term does. Someone searching "data science course" is closer to spending money and time on a specific decision than someone searching "government jobs" in the abstract. The category is small today and is very likely growing off a low base; it should not be dismissed for being outweighed by a category that has had a multi-decade head start.

What the data is actually telling institutions and employers

None of this is a full census of every job-related search in India. Branded job portals, government recruitment sites by name, and skilling platforms were not measured directly here, so the true volume of employment-focused search almost certainly runs higher than the generic category terms captured in this reading. What the comparison shows is relative: where the weight sits between a small number of large categories, not the complete picture of every search a jobseeker might run.

For institutions trying to move people toward AI and data skills, the practical read is not that the appetite is fake. It is that the appetite is not yet where the attention already is. A course, a certification, or a reskilling program competes for search visibility against a government-exam and coaching apparatus that already commands the overwhelming share of related search demand, and against a much more immediate, resume-shaped need for a job right now. Reaching this audience means showing up inside the searches people are already running, sarkari naukri, work from home jobs, resume format, rather than waiting for a smaller, still-forming category of AI-anxiety searches to arrive on its own.

The same logic applies to any employer or platform trying to be found by India's workforce as it thinks about AI. The searches say where the country's attention is standing today: overwhelmingly on the secure, credentialed, government-backed job, and only lightly on the AI-skill course. Being visible where people are actually looking, rather than where the AI narrative assumes they should be looking, is the more grounded starting point.

It is worth resisting one tempting but unsupported leap from all of this: that India is somehow behind, or in denial, on AI. The data measured here is a search-demand snapshot, not a judgment on the country's readiness. A workforce searching for the secure job and the correctly formatted resume is a workforce responding rationally to what search itself is for, solving the nearest, most concrete problem. The gap this piece measures says more about where AI-skilling has not yet built the same pull, familiarity, and single obvious name that the government exam has spent decades building, than it says about anyone's appetite for AI itself.

The evidence

Key findings, with their sources

Core government-exam searches (ssc, upsc, railway recruitment, nda exam, bank exam) total about 6.11 million a month in India.

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
The entire AI-and-data course search category (data science, ai, data analytics, artificial intelligence, python, prompt engineering, generative ai, machine learning, and chatgpt courses combined) totals about 250,400 a month, roughly 24 times smaller than the government-exam category.

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
"Sarkari naukri" alone draws about 1,000,000 searches a month, about four times the volume of the entire AI-and-data course category combined.

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
Explicit AI-job-fear queries ("will ai take my job," "jobs after ai," "ai proof jobs") total about 590 searches a month combined nationwide, against 6.11 million government-exam searches.

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
"Resume format" draws about 450,000 searches a month, roughly ten times the combined total of "upskilling," "career change," and "online certification" (about 43,210 a month).

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
Exam-coaching brand "physics wallah" alone draws about 1.5 million searches a month, roughly six times the entire AI-and-data course category, and "unacademy" (about 368,000) outdraws that category on its own.

emerging Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, August 2026.
"Linkedin" draws about 9.14 million searches a month, the largest single term in the dataset, but is excluded from the anxiety reading as a navigational query rather than an expression of job-market worry.

established Raveneye Global's analysis of India search-volume data, applying an established navigational-versus-informational query distinction.
Search-volume demand is a signal of what people type, not a census of who is employed, underemployed, or reskilling; the gap between categories is a gap in attention, not a proven gap in behavior.

established Raveneye Global measurement practice.

Calibration

What is proven, what is promising, what is unproven

Evidence tier	Tactics	What the evidence says
established	The direction: under job-market pressure, India's search demand runs toward the secure, credentialed government job and the immediate need to get hired, not toward the AI-reskilling course.	Corroborated from two independent cuts of the same dataset, the scale gap between the exam and course categories, and the near-total absence of explicit AI-fear phrasing, both pointing the same way.
emerging	The specific figures and ratios: the 6.11-million exam total, the 250,400 course total, the 24-to-1 ratio, and every individual term-level number in this piece.	Each is a single-month, single-country, Google-banded estimate from one reading of keyword-planner tooling, not a settled, repeated census.
contested	Whether the pattern is a reaction to AI-driven job anxiety specifically, rather than the continuation of a government-job pull that predates generative AI by decades.	India's exam-coaching economy existed at large scale well before AI-course search terms did; attributing today's pattern to AI specifically is a plausible reading, not a demonstrated one.

Reference

Glossary

Sarkari naukri: Hindi for "government job." One of the single highest-volume job-related search terms in India, and used throughout this piece as shorthand for the pull toward secure, credentialed government employment.
Search-volume band: Google's Keyword Planner reports monthly search volume as a rounded estimate rather than an exact count, so every figure in this piece is an approximation of demand, not a precise headcount.
Navigational query: A search where the person is trying to reach a specific, known site or brand, such as "linkedin," rather than expressing an unmet need. Treated differently here from an informational or transactional query.
Commercial or decision intent: How close a search sits to an actual decision or spend, distinct from raw volume. A course-related search can carry more decision intent than a much larger, more general job-search term.
Government-exam category: In this piece, the combined search demand for ssc, upsc, railway recruitment, the nda exam, and bank exam, the five terms used as the core measure of demand for government employment.

Straight answers

Frequently asked questions

Does this mean Indians are ignoring AI and reskilling?

Not quite. The AI-and-data course category is real, at about 250,400 searches a month, and it likely carries more decision intent than a general job-search term does. What the data shows is a matter of relative scale: under job-market pressure, the search reflex runs about 24 times harder toward the government exam than toward the AI course, not that AI interest is zero.

Why was LinkedIn's 9.14 million searches left out of the reading?

Because a search for "linkedin" is navigational: most people typing it want to reach the site itself, often to log in, not to express a specific worry or need about jobs and skills. Counting it alongside "sarkari naukri" or "government jobs" would mix two different kinds of search intent.

What exactly counts as "government-exam demand" in this reading?

Five terms: ssc (about 4,090,000 a month), upsc (about 1,830,000), railway recruitment (about 90,500), nda exam (about 60,500), and bank exam (about 40,500), for a combined total of about 6.11 million searches a month.

Is this data a prediction of what will happen to jobs in India?

No. It is a single reading of search-volume data from one month in one country. It shows where attention was sitting in August 2026, not where employment, AI adoption, or reskilling will land in the future. Treat the specific figures as a snapshot, not a forecast.

If AI-course search demand is so much smaller, does it matter at all?

Yes. A smaller category is not a meaningless one, especially where it likely carries higher decision intent and is probably growing off a low base. The finding here is about where the bulk of search attention currently sits, which matters for anyone trying to be found by this audience, not a claim that AI-skill demand does not exist.

Provenance

Sources

Raveneye Global's analysis of India search-volume data from Google's Keyword Planner, India, August 2026 (emerging)
Public mandate documentation of India's central government recruitment bodies: the Staff Selection Commission (SSC), the Union Public Service Commission (UPSC), the Railway Recruitment Boards, the National Defence Academy exam, and the Institute of Banking Personnel Selection (IBPS) (established)
Established search-behavior distinction between navigational queries (seeking a known site or brand) and informational or transactional queries, applied here to exclude "linkedin" from the job-anxiety reading (established)
Raveneye Global, "The Engineer's Reckoning: India's Entry-IT Escalator Splits in Two" (emerging), for the labour-market pressure this search behavior responds tosource
Raveneye Global, "India's Human-Capital Question" (emerging), for the credentialing and reskilling contextsource
Raveneye Global measurement practice note: search volume is read here as a demand signal, not a census of employment status or of who does or does not reskill (established)

Every figure above is attributed to a real, dated source and tagged with its evidence tier. Where a claim could not be verified to a primary source, it is not stated as fact.