14 min readA technology leader opens a renewal proposal from an Indian IT vendor. The vendor promises faster delivery with AI. Yet the annual price has not changed.
If the work takes fewer hours, who gets the savings?
The contract may hold the answer.
That question captures one part of the AI impact on Indian IT services. The sector is still growing. In February 2026, Nasscom estimated that fiscal 2026 revenue would rise 6.1% to $315 billion. Headcount was expected to grow only 2.3%, while clients increasingly wanted a share of AI-related productivity gains.
But pricing is only part of the story.
AI is also changing how providers develop software, maintain systems, structure teams, and compete for work. Some firms are creating new AI-related revenue while established services face pricing pressure.
This article separates reported results from forecasts and shows what AI is changing in delivery, pricing, jobs, vendor competition, and contract decisions.
What AI Is Doing to Indian IT Services So Far
Three developments are happening at the same time.
They can point in different directions, which is why a single growth or layoff headline does not explain what is happening across Indian IT services.
AI-related revenue is growing
Nasscom estimated AI-related industry revenue at $10 billion to $12 billion, or roughly 3% to 4% of total industry revenue. It also noted that the estimate did not capture AI revenue at every company.
Individual providers are reporting larger contributions under their own definitions.
TCS reported $3.1 billion in annualized AI revenue for the September 2026 quarter, representing more than 10% of its revenue on that run-rate basis.
Annualized revenue is important here. It converts a quarterly figure into an annual run rate. It does not mean TCS collected $3.1 billion of AI revenue during the previous 12 months.
Established services face pricing pressure
The growth in AI-related work does not mean every part of the industry is expanding at the same rate.
Reuters reported that TCS, Infosys, Wipro, and HCLTech grew only around 1% to 3% in the April-to-June quarter as clients pushed providers to deliver more work at lower prices.
Those are reported results.
The more aggressive estimates of future AI-related revenue pressure are forecasts, which should be treated differently.
Revenue is growing faster than headcount
Nasscom estimated FY26 industry revenue growth of 6.1%, while workforce growth was expected to be 2.3%.
That suggests the historical link between revenue and headcount is becoming weaker.
It does not prove that AI alone caused the change. Weak demand, hiring discipline, utilization, automation, offshoring patterns, and changes in the mix of services can also affect headcount.
Still, the direction matters.
Indian IT providers may be able to support more revenue without increasing staffing at the same pace as before.
The tension is becoming clearer: vendors can generate new AI-related revenue while simultaneously facing pressure on the price of established services.
That makes delivery economics the next part of the story.
Looking to Outsource Your IT Projects?
Enosis Outsourcing helps technology leaders scale their teams with expert offshore engineers. Get a free consultation today.
Get a Free ConsultationHow AI Changes Day-to-Day Service Delivery
AI can help teams draft code, generate test cases, analyze logs, prepare documentation, and automate parts of routine maintenance.
Picture a team updating a payments application.
An AI tool drafts a code change in minutes. Engineers still have to confirm that it handles failed payments, protects customer data, integrates with the existing application, and behaves correctly under edge cases.
The draft may be faster.
The finished release still depends on requirements, integration, testing, security review, and approval.
The work can move faster, but review and ownership do not disappear.
Two terms increasingly matter here.
An AI agent is software that can plan or perform parts of a task using approved tools and permissions.
An agentic workflow allows one or more agents to choose and carry out steps rather than following only a fixed sequence.
Both still need limits.
Buyers should know:
what an agent can do without approval
what requires human review
what systems or data it can access
who validates the result
who is accountable if the result is wrong
As routine work becomes easier to automate, engineers can spend more time on requirements, architecture, test design, output validation, security, integration, and release decisions.
Productivity claims still need scrutiny.
Infosys told investors that it migrated 3 million lines of Hertz COBOL code to microservices at 60% lower cost and in 60% less time. Amazon has said AWS Transform processed more than 1.1 billion lines of code and saved more than 810,000 hours.
Those are substantial reported outcomes.
They are also vendor-reported examples rather than independent measurements across the industry.
A successful demonstration or pilot does not automatically prove that the same productivity can be achieved safely in production.
Faster output also raises questions about AI-generated code quality, because more code only creates value if teams can review, test, and maintain it reliably.
The same distinction matters when an organization tries to move an AI pilot to production. A promising demo and a production-ready system face very different requirements around reliability, security, integration, and accountability.
How AI Is Changing IT Services Pricing
Return to the renewal proposal. The vendor may now complete some routine work faster. That does not mean the customer's invoice automatically falls.
Lower delivery effort and lower customer prices are different things.
Who receives the financial benefit depends largely on the contract.
With time-and-materials work, fewer approved hours may reduce the bill.
With fixed-price work, the vendor may retain the productivity gain unless the contract includes a price review, shared-savings mechanism, or scope adjustment.
Outcome-based contracts create another option by linking part of the fee to measurable results.
Recent contract discussions show how that pressure is developing.
Example | What is changing |
|---|---|
Persistent | Its CEO said clients are asking for the same work at 25% to 30% lower cost and with faster delivery |
Cognizant / Daimler Truck | A reported deal structure splits AI-related cost savings between vendor and client |
HCLTech / E.ON | A reported cloud agreement links later payments to efficiency gains |
TCS | Its CEO said about 80% of contracts in finance, HR, and other business services are now outcome-based |
Some of those deal details came from people familiar with the agreements rather than confirmed contractual disclosures. That distinction matters.
Other reports point in the same direction.
Kotak Institutional Equities has said some clients in multi-year managed services agreements are asking providers to price future AI productivity gains into the contract upfront.
HFS Research has also described clients reopening contracts relatively early when they believe AI has changed the cost structure.
This creates an important distinction between three different outcomes:
a lower contract price
revenue pressure on existing work
a decline in total industry revenue
They are not the same.
Kotak estimates that AI could create an annual revenue drag of roughly 3% to 3.5% relative to a baseline in which that disruption did not occur, even after accounting for some new AI-related spending.
That is a forecast of pressure against a hypothetical baseline.
It is not evidence that Indian IT services revenue has already fallen by 3% to 3.5%.
Jefferies has separately modeled scenarios where annual growth could be several percentage points lower than it otherwise would have been.
Again, slower growth is not the same as an outright decline.
Pricing model | What the buyer pays for | What AI may change |
|---|---|---|
Time and materials | Hours or people | Fewer approved hours may reduce the bill, although rates still require negotiation |
Fixed price | A defined scope | The provider may retain productivity gains unless the price or scope is reopened |
Outcome based | A measurable result | Savings can be shared, but measurement and accountability must be clear |
The same productivity pressure is also changing the broader outsourcing revenue picture as clients ask providers to share more of the savings created by AI.
Which Service Lines Face the Most Pressure
AI-related pricing pressure does not affect every service line equally.
The nature of the work matters.
Kotak identifies application development and customer-experience outsourcing among the more exposed areas. Infrastructure management and more specialized business-process work appear less exposed.
The revenue mix makes that important.
Jefferies estimates that application services account for a large share of revenue at leading Indian IT companies. Reuters has reported that application work represents around 55% to 60% of revenue at several providers, while HCLTech has a lower share.
Application managed services are also important because repetitive maintenance and support tasks are increasingly exposed to automation.
Service line | AI pressure | What still requires people |
|---|---|---|
Application development | High, according to Kotak | Architecture, requirements, integration, and review of AI-written code |
Customer experience outsourcing | High, according to Kotak | Complex cases, judgment, customer escalation |
Application maintenance and managed services | High in Jefferies' outlook | Incident ownership, legacy knowledge, architecture, and production decisions |
Software testing and QA | Significant task-level automation potential; no comparable published ranking | Test strategy, edge cases, exploratory testing, and release sign-off |
Infrastructure management | Lower, according to Kotak | Reliability, security, incident response, and operations |
AI implementation and modernization | Mixed | Data engineering, integration, workflow redesign, governance, and security |
AI implementation deserves particular attention.
Kotak argues that enterprise AI still requires work in:
data engineering
workflow redesign
integration
security
testing
governance
legacy modernization
That means AI can reduce demand for some repetitive work while creating demand for different technical capabilities.
Published comparisons across service lines remain limited.
The table above combines published analyst views with editorial interpretation, so it should be treated as directional rather than as a measured industry ranking.
These shifts also change the economics of software testing, where automation may reduce repetitive execution without removing test design, review, or release judgment.
The same pattern appears in software maintenance and support, where routine work can be automated while complex incidents, architecture decisions, and system knowledge still require engineers.
What Is Happening to Jobs, Hiring, and Skills
The change in service demand leads to another question.
What happens to the people doing the work?
The employment impact is real, but the available evidence is more nuanced than many layoff headlines suggest.
Across the industry, Nasscom still expected total workforce growth in FY26.
The largest vendors showed a different picture.
The five largest Indian IT firms, TCS, Infosys, Wipro, HCLTech, and Tech Mahindra, recorded a combined net headcount reduction of 6,981 in FY26, compared with a net addition of 12,718 in FY25, according to The Times of India.
TCS accounted for much of the reduction.
But net headcount reduction is not the same as layoffs.
It also reflects:
hiring
voluntary departures
restructuring
utilization
slower replacement hiring
TCS separately announced restructuring-related job cuts connected to changes in skills and operating needs.
Entry-level hiring is changing too.
Nasscom's president has said campus recruitment has fallen significantly. TCS initially planned roughly 25,000 fresher hires for FY27, compared with approximately 40,000 to 42,000 in earlier years.
Those plans can change during the fiscal year and should not be treated as final hiring totals.
Not all of this can be attributed to AI.
Weak client demand and delayed technology spending also affect hiring.
The more durable shift may be in which skills become more valuable.
As AI handles more routine production work, providers still need people who can:
design architecture
understand business requirements
integrate complex systems
engineer and govern data
review AI-generated output
manage security
test high-risk workflows
redesign business processes
govern AI systems
make production and release decisions
That does not mean every role automatically becomes more senior.
It does mean that judgment, integration, domain knowledge, and accountability become harder to separate from delivery.
This is why large forecasts of future job losses should be treated cautiously. Different estimates use different assumptions about demand, productivity, redeployment, new services, and attrition.
Looking for a Team That Can Turn AI Into a Real Product?
Explore companies with experience building AI and machine learning solutions around practical business requirements.
Find AI Development TeamsHow AI Affects Different Indian IT Services Providers
The shift in skills and demand also changes how providers compete.
Some mid-sized firms are currently growing faster than the largest Indian IT vendors. That does not automatically mean scale has become a disadvantage.
Reuters reported that Persistent's revenue rose 16% and Coforge's by roughly one-third in the April-to-June quarter, while four of the largest providers grew around 1% to 3%.
That is one quarter and a limited group of companies.
It does not prove that market share is permanently shifting from large providers to smaller firms.
HFS Research's Phil Fersht told Reuters that some smaller providers are winning work by deploying senior leadership quickly and offering more flexible commercial models.
Large firms still have advantages when buyers need:
global delivery capacity
large transformation programs
broad technology coverage
long-term operational support
deep enterprise relationships
Evidence for very small providers remains thinner.
AI may lower some barriers to demonstrating technical capability, especially through pilots. But running a pilot is different from supporting a large production environment over several years.
Competition also comes from outside traditional IT services companies.
Cloud platforms such as AWS increasingly sell AI-assisted modernization and agent-based tools directly.
Provider type | What AI changes | What buyers should check |
|---|---|---|
Large providers | Pressure on some legacy maintenance revenue and pricing | How AI savings are shared and what AI systems actually run in production |
Mid-sized firms | Some are winning work through faster senior involvement and flexible pricing | Capacity for larger programs and depth of AI capabilities |
Specialists | Domain knowledge may be more defensible than routine delivery capacity | Evidence of production results in the buyer's industry |
Small providers | AI can make it easier to compete on pilots | Financial stability, security practices, delivery depth, and long-term support |
The buyer's decision should still come back to project fit. A faster-growing provider is not automatically a better provider.
AI therefore makes choosing a software development company less about size alone and more about proven capability, governance, delivery depth, and fit with the project.
What the Changes Mean for IT Services Buyers
The next proposal from an Indian IT vendor may promise:
faster delivery
fewer people
lower effort
a new pricing model
AI-enabled productivity
Each promise needs a different check.
Buyers appear to have more leverage in many current negotiations.
Everest Group's CEO told Reuters that the current environment favors clients. That is one analyst's assessment, and leverage will vary by service, contract, and provider.
Still, the contract evidence shows that clients are actively pushing providers to demonstrate and share AI-related productivity gains.
Before signing or renewing, work through these checks.
Ask how AI savings will be shared
Possible approaches include:
lower rates
fewer hours
smaller teams
expanded scope
gain sharing
outcome-based fees
Do not assume productivity gains automatically flow to the buyer.
Negotiate a price-review mechanism where appropriate
Define:
when pricing will be reviewed
the productivity baseline
what evidence is required
who validates the evidence
how verified savings affect price or scope
An automatic price reduction will not fit every contract.
A review mechanism is usually more flexible.
Define outcomes before agreeing to outcome pricing
Outcome-based pricing only works when the result is measurable.
Record:
the baseline
the desired outcome
acceptance criteria
the measurement method
who validates performance
Outcome-based pricing works better when the scope of work already defines those baselines, responsibilities, acceptance criteria, and measurement methods.
Separate pilots from production
Ask for production references.
A successful pilot may prove that a vendor can build something quickly.
It does not automatically prove that the system can operate safely at scale.
Ask for measured production results such as:
delivery time
incident rates
defect rates
availability
operating cost
human review requirements
Providers also define AI revenue differently, so a large AI-revenue figure alone does not prove delivery maturity.
Set AI security and accountability rules
Document:
permitted AI tools
permitted data use
confidentiality rules
system access
approval thresholds
logging requirements
human review
responsibility for AI-generated errors
Security, accountability, vendor dependency, and unrealistic productivity assumptions remain part of the wider IT outsourcing risk picture.
Be careful with unusually low bids
Tech Mahindra's CEO has said some competitors are pricing deals around assumptions of extremely large productivity gains over several years.
Infosys has also said it walked away from contracts it considered economically unattractive.
Very low bids can be attractive.
They can also depend on productivity assumptions that have not yet been proven in your environment.
Consider whether some work should stay in-house
Reuters has reported that some clients are reconsidering which tasks should remain internal.
The right answer may differ by activity.
Core architecture, proprietary workflows, sensitive data, strategic product knowledge, or high-risk decision-making may justify internal ownership even when execution is outsourced.
AI can make the in-house versus outsourcing decision more granular because some activities may benefit from external capacity while strategic knowledge and accountability stay internal.
Need Help Building Your Vendor Shortlist?
Tell us what you're looking for and get help identifying companies relevant to your requirements.
Get a Free ConsultationOutlook for India's IT Services Industry
The next competitive advantage may come from proving results rather than simply supplying more people.
Traditional application work faces pricing pressure.
At the same time, new demand may come from helping enterprises:
modernize legacy systems
integrate AI with existing software
prepare and govern data
redesign workflows
secure AI systems
connect AI agents to enterprise processes
Those projects can require specialized engineering and deep knowledge of a client's operations.
They are real opportunities.
They are not a guarantee that new AI revenue will completely offset pressure on established services.
Kotak expects pressure on some traditional service lines to continue through FY28, with the possibility of stronger new demand later.
That is an analyst forecast, not an observed industry outcome.
TCS's September 2026 quarter also shows that AI-related revenue and healthy operating margins can coexist.
One company and one quarter are not enough to establish a long-term industry trend.
Commercial models may continue moving toward:
measurable outcomes
shared productivity gains
domain expertise
integration capability
AI governance
smaller but more specialized teams
That does not mean time-based pricing disappears.
Time-and-materials and fixed-price contracts will remain useful where results are difficult to define or the scope is uncertain.
Over the next few years, buyers should watch:
how AI savings appear in renewal rates and scope
how often contracts are reopened early
the share of outcome-based deals
how outcomes are measured
how vendors define AI revenue
changes in entry-level hiring
competition from cloud-platform AI tools
whether AI-related revenue offsets pressure in established services
India's experience is also part of the broader shift in how AI is changing outsourcing across delivery models, pricing, staffing, and vendor relationships.
Final Takeaway
The main effect of AI on Indian IT services so far is not a collapse in demand. It is pressure on how work is delivered, staffed, and priced.
Industry revenue is still growing. Headcount is growing more slowly. AI-related services are creating new opportunities while buyers expect vendors to share productivity gains from established work.
For buyers, that creates both leverage and responsibility.
Define outcomes clearly. Ask for production evidence. Verify productivity claims. Watch for bids built on unproven assumptions. Decide how AI savings should be shared before the savings appear.
Most importantly, choose providers that can explain how AI changes cost, quality, risk, and accountability.
Then put those answers in the contract.
