Quick answer
Economic Survey 2025-26 estimates a 39.5% relative export gain for AI-exposed services after AI diffusion; software and business services lead, with important caveats.
AI can complement services exports by raising productivity and lowering trade costs, even while automating individual tasks. Box VII.2 finds stronger post-diffusion export performance in AI-exposed services, not evidence that every job or service category benefits.
Concept diagram
AI and the Future of India’s Services Exports
A human worker and AI co-pilot collaborate on digitally delivered services.
- The scene shows complementarity, not guaranteed gains for every worker.
- Survey estimates differ across software, business and financial services.
Reading an AI impact study without overclaiming
Digital delivery
- Digitally deliverable services can cross borders through computer networks; software and business services are important examples.
- AI can lower translation, coding and search costs while trade itself diffuses technological knowledge.
Difference-in-differences
- DiD compares the before–after change in a treated group with the before–after change in a control group.
- Two-way fixed effects account for category and time effects; the causal interpretation still relies on identification assumptions, including parallel trends.
Box VII.2: how the Survey defines its comparison
| Design element | Specification |
|---|---|
| Data | Quarterly RBI balance-of-payments data; Q1 FY18–Q2 FY26 |
| Treated group | Software, business and financial services; presumed AI-intensive |
| Control group | Travel, transport, insurance, government services n.e.c. and communication |
| Intervention | Q2 FY23 |
| Overall estimate | Post × treatment coefficient 0.333; significant at 1%; implied increase about 39.5% |
| Pre-trends | Formal tests find no evidence of differential pre-intervention trends |
Data chart
Estimated gains after AI diffusion: relative, not annual growth
Source: Economic Survey 2025-26, Box VII.2, Tables A–B (author’s calculations) · Approximate implied gains reported in the Survey. The overall group overlaps the subcategories; these are not additive. Financial services are excluded because their positive estimate is statistically insignificant.
Show the data (Per cent)
| Period / item | Implied export increase |
|---|---|
| AI-exposed overall | 39.5 |
| Software | 44 |
| Business | 67 |
Global context: forecasts and associations are not Indian outcomes
| Evidence in Box VII.2 | Reported magnitude | Interpretation |
|---|---|---|
| WTO global trade scenario | 34–37% higher by 2040 | Potential AI-enabled expansion, not a realised increase |
| Digitally deliverable services scenario | About 42% expansion | Global projection, not an India export target |
| Trade and knowledge diffusion | 10% more digitally deliverable trade associated with 2.6% more cross-border AI patent citations | Association; not a universal causal rule |
| Financial services coefficient | 0.12; statistically insignificant | No statistically established category-specific gain in this sample |
Turn task disruption into higher-value exports
Reskill around tasks
Build AI literacy alongside domain expertise rather than assuming all service occupations disappear together.
Move beyond labour arbitrage
Compete through product engineering, analytics and client-specific solutions where AI complements judgement.
Build trusted delivery
Strengthen data governance, security and accountable human oversight to sustain overseas client confidence.
What the AI result does not establish
Statements examiners like to twist — learn the correct version.
AI caused a 39.5 per cent annual rise in all Indian services exports.
This is an approximate relative post-treatment estimate for treated categories compared with controls, not aggregate annual growth.
A coefficient of 0.333 means exactly 33.3 per cent higher exports.
For a log specification, 100 × (exp(0.333) − 1) is about 39.5 per cent, the Survey’s implied effect.
An insignificant financial-services estimate proves zero impact.
It means the sample does not establish a statistically significant effect; it does not establish absence of any effect.
Prelims focus
- Treatment starts Q2 FY23; data cover Q1 FY18–Q2 FY26.
- Software and business estimates are significant at 1%; financial services are not.
Mains angle
- Discuss AI as both a task-displacing technology and a productivity-enhancing export opportunity for India.
- Use the DiD evidence with caveats: export gains do not directly measure wages, job quality or employment effects.
Flip & recall
Answer in your head first, then tap the card to check.
Practice MCQs
Original Prelims-style questions with explanations.
0 / 3 correct · 3 left
Q1.Which is included in the AI-intensive treated group in Box VII.2?
Q2.The 39.5 per cent estimate is best described as:
Q3.A positive but statistically insignificant financial-services coefficient means:
