AI cost savings are reshaping technology services contracts

Customers are pricing in AI savings before providers deliver them

Technology services customers are awarding new contracts on the assumption that artificial intelligence will cut delivery costs by at least 30%, and sometimes by as much as 50%. Many providers are accepting these expectations while still developing the systems and operating models needed to deliver such savings consistently.

The timing creates substantial financial risk because technology services contracts often run for three to five years. Once a provider agrees to lower prices, it must improve productivity enough to protect its margins throughout the contract. If those improvements arrive late, the provider absorbs the gap between its contractual commitments and actual delivery costs.

AI is reversing the usual sequence of technology-driven efficiency gains. Traditionally, providers invest in better delivery methods, establish productivity improvements, and then negotiate how to share the benefits with customers. Now customers increasingly expect those benefits when they sign, leaving providers to build the capabilities needed to fulfill the promise afterward.

The consequences reach across the technology services business model. Providers must rethink how they sell, price, staff, and deliver work when customers already expect substantial efficiency gains. The challenge is to turn anticipated AI productivity into reliable customer savings while retaining enough value for profitable growth.

Customer expectations are running ahead of AI delivery

The pricing pressure extends beyond new contracts. During renewals, enterprise customers are demanding efficiency improvements reaching 20%, forcing existing providers to deliver established services at lower cost. These demands create a different challenge from new contracts because providers must improve operations with staffing structures, workflows, and service commitments already in place.

New contracts give providers more freedom to design AI-enabled delivery from the start, but they also carry more aggressive cost assumptions. In both cases, providers make financial commitments before knowing whether they can deliver the necessary productivity consistently. The risk varies by type of work because AI’s potential differs substantially across technology services.

Bain & Company estimates that AI could improve productivity by 15% to 45% across different service lines. Bain is a management consulting firm with a commercial interest in enterprise transformation, including demand for the strategic and operational changes it recommends. Its analysis shows substantial variation in the gains available from different activities, making service-level economics important when providers negotiate contracts.

The main quantitative expectations describe different parts of the commercial problem.

Customer demands establish the savings providers must commit to, while productivity estimates describe potential improvements in delivery. The benchmark against existing delivery baselines adds another consideration for companies redesigning their operations. Because these measures serve different purposes, providers need to connect each contractual commitment to the productivity available in the relevant service.

A company can have effective AI tools and still face considerable financial exposure. Those tools may accelerate particular activities without producing equivalent savings across an entire engagement. When contracts assume broad improvements, the gap becomes a problem for delivery teams and finance functions.

Investors are questioning the economics of technology services

The same uncertainty is affecting financial markets. S&P Capital IQ financial data and Bain analysis show that technology services company share prices have fallen sharply since the first quarter of 2025. The pressure reflects investor expectations that AI will let providers maintain comparable service output with substantially fewer employees.

Those expectations raise questions about businesses historically built around large workforces and labor-intensive delivery. A smaller workforce can reduce operating costs, but it can also reduce revenue when customers buy services according to the labor required. Investors therefore need to understand how productivity gains will affect pricing, margins, and growth capacity.

Customer negotiations make that question more urgent. When buyers capture much of the expected productivity benefit through lower prices, providers may struggle to turn operational improvements into stronger financial performance. The market is consequently evaluating whether technology services companies can adapt their economics as AI changes the relationship between labor and output.

Delivery capabilities are still developing

Meeting those expectations requires operational maturity that the industry has yet to establish. Despite substantial AI investment, no technology services provider has fully demonstrated reliable AI-enabled delivery across its operations at scale. Companies have introduced platforms and tools that improve productivity in specific activities, but delivering consistent results across contracts is harder.

Scaling those improvements requires providers to integrate AI into workflows, reorganize teams, supervise automated work, maintain quality, and reproduce savings across different customer environments. Each requirement brings costs and operational dependencies that affect contract economics. The challenge grows when aggressive efficiency commitments are already built into the price.

The timing puts particular pressure on providers with limited investment capacity. They must fund changes to delivery operations while meeting customers’ immediate savings expectations. Companies heavily dependent on conventional staffing models face an additional challenge because their existing structures may struggle to deliver the productivity customers demand.

Even when AI produces the required operational gains, providers still need to retain some of the financial value. How much they keep depends on the commercial arrangement under which the work is sold.

How AI productivity changes services revenue

Technology services companies have historically earned substantial revenue by supplying skilled labor to enterprise customers. Under time-and-materials pricing, customers pay based on the people assigned to an engagement, their billing rates, and the hours worked. Revenue therefore depends heavily on labor consumption, while delivery costs depend on the workforce needed to provide the service.

AI changes that relationship by reducing the human effort required for certain activities. Consider an engineering engagement where a smaller team produces the same output using AI-assisted development tools. Delivery costs may fall substantially, but fewer billable hours can also reduce revenue when the customer pays for labor consumed.

The financial benefit therefore depends on the contract. When revenue stays stable, lower delivery effort creates room for higher margins. When prices fall alongside costs, the provider must improve productivity enough to offset the revenue reduction. The same technical capability can produce very different financial results under different commercial arrangements.

Contracts that already include expected AI savings make the calculation harder. A provider might achieve significant productivity gains and still see little margin improvement because those gains are needed to meet existing pricing commitments. Managers must therefore evaluate four connected variables: the agreed customer price, actual delivery costs, achieved productivity, and the share of value the provider retains.

These variables put commercial design at the center of AI transformation. Providers need agreements that reward delivery efficiency while giving customers measurable financial benefits. Finding that balance also requires sales teams to understand how AI changes the customer’s operations.

AI transformation requires different sales capabilities

Investment in AI platforms has exposed a gap between technical capability and commercial execution. Many providers have built systems that support AI-enabled delivery, yet relatively few have established consistent ways to turn those capabilities into funded customer transformation programs. The problem often emerges when account teams try to sell the technology.

A provider may demonstrate a credible AI platform while its sales team struggles to explain how it would change a particular customer’s business. Enterprise buyers need to understand which workflows will improve, what organizational changes are necessary, how much value can be created, and how the provider will deliver it. Without those connections, technical interest may never become a commercial commitment.

Bain identifies stronger AI transformation sales capabilities as an important response. It recommends combining customer business knowledge, expertise in the workflows being changed, and enough AI competence to judge what can realistically be delivered. Sales teams with these skills can identify operational problems and develop proposals based on achievable improvements.

Those proposals also need credible financial assumptions. A sales team promising substantial savings must understand delivery costs, implementation requirements, and the changes customers must make. Otherwise, it may win a contract whose terms exceed what the delivery organization can achieve profitably.

Sales incentives can make the problem worse when they reward staffing volumes and billable revenue. A proposal that reduces labor requirements may create significant customer value while lowering the measures used to assess sales performance. Providers pursuing AI-enabled growth therefore need incentives that reward transformation opportunities and profitable delivery.

Pricing models must reward measurable outcomes

The next step is to connect compensation more closely to delivered value. Under outcome-based pricing, a provider receives payment according to agreed results, measured against defined performance or business objectives. This lets compensation reflect the value of a service even when AI reduces the labor needed to perform it.

Gain-sharing arrangements offer another option. The customer and provider agree to divide the financial benefits generated by measurable improvements, giving the provider a direct incentive to reduce costs or improve performance. Both structures can help protect provider margins while giving customers a clear financial return.

Their effectiveness depends on how the agreements divide responsibility. An outcome-based contract can expose a provider to risk when success depends on customer systems, decisions, or operating processes outside its control. Gain-sharing requires an agreed starting cost baseline, reliable measurement of subsequent improvements, and a way to determine which benefits came from the transformation.

These requirements make contract design a shared responsibility for commercial, delivery, and finance teams. They must agree on expected productivity, implementation costs, success measures, and how to divide financial gains and risks. A contract built on inconsistent assumptions can damage profitability even when the technology performs well.

Bain sees growing interest in outcome-based pricing and gain-sharing arrangements, although the industry has yet to establish a consistently successful approach. The opportunity is significant for providers whose current models depend heavily on labor utilization. Their pricing methods, sales incentives, and delivery structures must evolve together if AI-generated efficiency is to support sustainable margins.

Customer knowledge and specialization create competitive advantage

Better pricing models can help providers retain value, but customers still need a reason to choose one services company over another. As capable AI tools become widely available, access to the technology offers limited differentiation. Providers must show they can apply it to specific business problems more effectively than competitors.

Bain’s recommendations emphasize specialization, existing customer relationships, mature delivery platforms, and technology partnerships. These capabilities support different stages of the commercial process, from finding valuable opportunities to designing suitable solutions and delivering them consistently. Together, they help providers compete on service quality and economic value.

Specialization makes value easier to demonstrate

Broad industry expertise offers limited differentiation when several providers make similar claims. Bain recommends concentrating investment in carefully selected micro-verticals, meaning narrow industry segments with specific business requirements. This focus lets providers develop deeper knowledge of the workflows, systems, and operating constraints that determine whether AI can produce useful improvements.

Choosing those segments requires three connected decisions. Providers should identify where customers are prioritizing spending, which narrow industry markets offer attractive opportunities, and which geographic regions they can serve effectively. Those choices determine where the company concentrates sales effort, technical development, and delivery capacity.

Concentrated expertise can improve both sales and execution. When a customer compares five providers with similar general capabilities, price may become the deciding factor. A provider that understands the customer’s operating environment more deeply has a better opportunity to demonstrate specific improvements and justify its commercial terms.

Customer experience offers one measure of whether that expertise creates meaningful differentiation. Bain identifies a sustained Net Promoter Score® of 50 or higher as a strong signal. The score measures customers’ willingness to recommend a company, indicating whether they consistently see value in the relationship.

Specialization also directs investment. Providers can concentrate funding, talent, and platform development in markets where their knowledge and customer relationships give them a realistic advantage. This focus matters especially when AI transformation requires substantial spending before commercial returns are established.

Existing customer knowledge supports transformation

Specialization becomes more valuable when paired with detailed knowledge of a customer’s technology environment. Incumbent providers often understand the applications, integration dependencies, data flows, business rules, operational exceptions, and accumulated technical debt involved in running that customer’s operations.

This knowledge matters because enterprise AI must work within established systems and processes. An independent demonstration may show that an AI capability can perform a task, but production deployment requires knowing where information originates, how decisions are made, and which exceptions affect execution. Providers already managing those environments can spot opportunities and constraints that a new entrant may take longer to discover.

The commercial opportunity is to use that knowledge to shape the customer’s transformation plans. A traditional “delivery partner” executes work the customer has already defined. A “reinvention partner” helps identify improvements, establish their business value, and decide how the customer’s operations should change.

This expands the provider’s role in the relationship. By identifying economically attractive workflow changes, it may create opportunities beyond its original engagement. The value of incumbency therefore depends on turning operational familiarity into specific proposals that customers are willing to fund.

Platform maturity determines delivery capability

Those proposals still need dependable technology. Providers increasingly offer AI orchestration layers, which coordinate how AI models, autonomous software agents, and other components work together. Many also maintain agent libraries containing reusable AI components and enterprise guardrails, meaning controls that govern system behavior and reduce operational risk.

The value of these components depends on how they perform in customer environments. An agent library becomes useful when its components are reliable, mature, and connected to services customers buy. Orchestration must handle complex enterprise workflows, while guardrails must keep automated activity within acceptable operational limits.

Platform maturity also means working with technology the customer has already selected. Some industries require service providers to operate within customer-controlled platforms, limiting the role of proprietary systems. Providers pursuing those engagements must apply their expertise, agents, and delivery methods within the chosen architecture.

Adaptability therefore becomes a competitive capability. A provider whose offering depends entirely on customers adopting its proprietary platform may face restrictions where enterprises already have established AI environments. Specialized components, integration expertise, and dependable delivery can create value across different platforms.

Partnerships can create new customer opportunities

The need for adaptability extends to technology partnerships. Traditional partnerships often help providers meet customer demand that already exists. AI creates opportunities for services companies and technology vendors to collaborate earlier, identifying applications and developing services around emerging business requirements.

Bain recommends expanding relationships with hyperscalers, the large cloud infrastructure providers, and companies developing frontier models, the most advanced general-purpose AI systems. These partners bring infrastructure, model capabilities, and technical expertise that services companies can combine with their knowledge of customer operations.

The right collaboration depends on the customer’s stage of AI adoption. An organization beginning its AI work may need help identifying suitable workflows and establishing technical foundations. A more advanced enterprise may require specialized agents, deeper systems integration, or changes to its operating model.

Providers and partners can use these differences to develop joint commercial opportunities alongside their technical work. Collaboration becomes more valuable when it identifies specific customer needs, produces differentiated offerings, and supports successful implementation.

Proving the commercial model before expansion

Specialization, customer knowledge, and partnerships can help providers win business. The next challenge is proving that the work can be delivered profitably and repeated across similar engagements. Bain recommends starting with a focused opportunity to test the relationship between customer demand, productivity, implementation costs, and achievable margins.

This approach limits the risk of investing heavily in an operating model whose economics are still developing. An initial engagement lets a provider test commercial assumptions, identify delivery constraints, and establish which capabilities can be reused. Expansion then becomes a decision based on demonstrated performance.

Start with a defined customer opportunity

The first decision is where to focus. A provider should select a promising market opportunity and establish an ideal customer profile: the characteristics of organizations most likely to benefit from the proposed service and buy it. This profile turns industry specialization into a practical guide for sales and investment.

The company can then develop an offering around those customers’ workflows, business requirements, and technology environments. Domain knowledge and proprietary capabilities help establish differentiation, while the customer’s operating constraints determine what the solution must accomplish. The commercial proposition becomes clearer when the provider connects a defined business problem to a measurable improvement.

Delivery design follows from that proposition. The provider must decide which activities AI can perform, where human expertise remains necessary, how to maintain quality, and how to measure performance. Pricing must develop alongside those decisions so contractual promises match the cost and productivity assumptions of the intended delivery process.

The initial engagement tests the complete model. The provider needs to demonstrate measurable customer savings and improved margins for its own business, then determine whether those results can be reproduced. Once the relationship between delivery performance and financial outcomes is established, management has a stronger basis for pursuing the ambition of double-digit growth.

Design technical delivery and customer adoption together

The initial engagement must also account for changes to the customer’s operations. A functioning AI system creates productivity potential, but financial savings depend on whether the customer can incorporate it into everyday work. Technology implementation and organizational adoption therefore need to advance together.

Consider a provider developing AI agents to perform activities previously handled by a customer-facing service team. The agents must execute their assigned tasks reliably, while the customer must decide how employees will supervise them, manage exceptions, approve decisions, and handle changed responsibilities. Delays in these operating changes can prevent technical productivity from becoming financial savings.

Bain recommends developing the technical delivery system and the customer’s transition process together. Providers should understand the existing environment, define the intended operating model, and plan the organizational changes needed to reach it. This connects technical implementation directly to the customer’s ability to realize the expected benefits.

The connection also matters when setting contractual targets. Savings commitments must reflect both the activities the provider can improve and the changes the customer must make. Clear responsibilities help prevent situations where the technology works successfully but the agreed financial results remain out of reach.

Delivery operations require structural redesign

Customer adoption is only one side of implementation. Providers must also change their own delivery organizations to capture larger productivity gains. Adding AI tools to existing processes can accelerate individual tasks, but conventional team structures and workflows may limit the benefits across an engagement.

Bain argues for redesigning team composition, workflows, global delivery arrangements, and pricing structures around AI-enabled execution. These changes affect how work is assigned, which capabilities teams need, and how human supervision is organized. Providers must also determine where technical expertise adds the most value after AI reduces the effort required for routine activities.

The decisions have direct financial consequences. Different team structures change labor costs, management responsibilities, and how work is distributed across delivery locations. Global delivery models therefore need to reflect the new balance between automated activity and specialized human expertise.

Cost assumptions must change with the operating structure. A provider planning future margins around conventional staffing ratios may misjudge the economics of an AI-enabled engagement. Delivery architecture, organizational design, and contract pricing must share consistent assumptions about the work performed and the resources required.

AI changes workforce economics

These structural changes are especially important for junior-level work. AI can increase productivity in activities traditionally performed by entry-level employees, changing how providers assemble teams and develop future specialists. Companies must reconsider which tasks offer useful learning opportunities and which capabilities employees will need as automation spreads.

Bain anticipates growing importance for forward deployed engineers, technically skilled professionals who work closely with customers to implement AI systems in their operating environments. These roles combine engineering expertise with knowledge of customer workflows, integration requirements, and deployment constraints.

Their importance follows from the demands of enterprise transformation. Employees must understand how technology can improve a business process and what changes will make that improvement work in practice. Domain expertise, creative thinking, and AI proficiency therefore become more valuable in hiring and professional development.

Career progression may also need to depend more on demonstrated capabilities than conventional role structures. As work changes, providers must develop employees who combine technical judgment with business knowledge. This affects the future talent pipeline as well as the cost of current engagements.

Pricing must reflect the resulting mix of capabilities. Conventional staffing roles may become less useful as the basis for commercial agreements when delivery increasingly combines specialized professionals with AI-enabled systems. Providers may instead price around the capabilities needed to achieve an agreed result.

Leadership responsibilities change alongside workforce development. Executives and account leaders need to understand customer economics, organizational change, and the financial implications of AI-enabled delivery. That knowledge becomes particularly important when negotiating contracts based on productivity improvements that delivery teams are still learning to achieve consistently.

Repeatability establishes commercial viability

These operating changes require investment before returns are assured. Providers must develop AI capabilities, redesign processes, train employees, strengthen sales functions, and help customers manage organizational transitions. Meanwhile, customers are already demanding financial benefits, putting pressure on the resources available for transformation.

Resource-constrained providers face a particularly difficult choice. Companies unable to finance the transition may struggle to meet demanding contract terms, while those relying on generic platforms without specialized customer expertise may find it hard to support attractive pricing. Management must therefore judge carefully when an initial success justifies expansion.

A profitable pilot provides useful evidence, but repeatability requires a deeper understanding of delivery economics. Providers must identify which activities can be standardized, which need customer-specific work, and how those differences affect costs. They also need consistent methods for measuring customer savings and dividing the financial benefits.

These measurements connect sales promises to delivery results. If an engagement produces attractive margins because it received exceptional attention or unusually high investment, management needs to understand how its economics will change under normal delivery conditions. Extensive customization creates a similar challenge when every new customer requires substantial additional work.

Bain recommends building a complete commercial approach around one high-potential opportunity, demonstrating its economics, and using the validated model for expansion. This gives management a more useful investment measure than the number of AI tools deployed or engagements launched. The important evidence is whether additional customers can be served while preserving the expected relationship between customer savings, delivery costs, and provider margins.

Providers with trusted customer relationships may have an early advantage as AI creates new categories of technology services. Those relationships can help establish demand, but sustained expansion depends on operating models that deliver consistent results as the number and variety of engagements grow.

Growth demonstrates the opportunity in AI services

The need to prove delivery economics has not stopped technology services companies from expanding. Coforge, EXL, and Persistent are among the providers that recorded growth reaching 30% in 2025. Their performance shows that substantial business expansion remains possible while AI changes customer expectations and service delivery.

Coforge offers a particularly relevant example of industry specialization. The company reported 29% growth in its last fiscal year and has undertaken more than 150 AI engagements concentrated in four industries: banking and financial services, insurance, travel, and healthcare. This focus shows how a provider can combine sector knowledge with a substantial portfolio of AI-related customer work.

The example also shows how AI creates new commercial opportunities alongside pressure on established labor-based activities. Enterprise customers are investing in AI and related transformation, creating demand for providers that can connect technical capabilities to industry-specific requirements.

Growth figures, however, establish a different outcome from delivery profitability. Coforge’s reported expansion does not establish that its AI engagements caused its revenue growth. Similarly, the growth recorded by EXL and Persistent does not demonstrate that these companies have perfected outcome-based pricing, gain-sharing arrangements, or AI delivery at scale.

For executives evaluating these businesses, the financial distinction matters. Revenue growth shows that customers are buying services, while contract-level economics reveal how much value providers retain after meeting delivery commitments. Understanding those economics requires examining achieved productivity, implementation costs, customer savings, and margins across repeated engagements.

As providers expand their AI portfolios, these measures will carry more weight in investment decisions. A company that reproduces attractive economics across customers has a stronger basis for funding new offerings and entering additional markets. One whose engagements require extensive customization or exceptional delivery resources faces a different calculation, even when demand remains strong.

Concluding thoughts

AI is changing the economics of technology services faster than providers can adapt their business models. Customers are demanding substantial cost reductions before AI-enabled delivery has demonstrated consistent results at scale. For executives, the immediate challenge is to meet these expectations without committing to savings that undermine profitability.

Long-term growth will depend on more than deploying AI tools or reducing labor costs. Providers must align pricing, sales incentives, workforce capabilities, and delivery operations around measurable customer outcomes. Industry specialization and deeper customer relationships can help differentiate their offerings, but those advantages must translate into repeatable delivery performance and sustainable margins.

Leadership teams should prioritize proving the economics of AI-enabled services before expanding investment. That means establishing realistic productivity targets, measuring implementation costs, and determining how much value the provider can retain under different contract structures. Growth opportunities remain substantial, but revenue expansion alone will not demonstrate that the underlying business model is working.

The defining competitive advantage will be the ability to deliver AI-driven productivity consistently while capturing enough of its value to fund continued growth. Providers that establish this balance will be better positioned to meet customer expectations, protect profitability, and compete as technology services move away from traditional labor-based models.