Technology Services in the AI Era: Rethinking How Work Is Delivered

Technology Services leaders are facing a difficult timing problem. They are making decisions about teams, services, and pricing while the uses of AI are still changing.

Software development shows why this is hard. AI can help engineers complete parts of existing work faster, but a productivity gain on one task does not tell a firm how to staff, price, or deliver an entire project. Leaders must learn where AI works reliably in client environments while making decisions about the people and capabilities those clients need now.

At TechAlpha, Arkadiy Dobkin, Executive Chairman and Principal Founder of EPAM Systems, discussed this challenge with Zinnov President, Sidhant Rastogi. The conversation drew on the experience of leading a Technology Services firm through change and on what its teams are learning from AI projects. It raised a practical question for the industry: How do firms change the way they work when neither providers nor customers have a settled model to follow?

The answer has to come from decisions made in live projects, and from working with customers as those decisions change what can be delivered.

How Can Technology Services Firms Adapt Their Delivery Models for AI?

The approach will vary by firm. A provider’s existing services, revenue mix, and customer base determine where AI creates immediate pressure and where it may open new work. A company seeing productivity gains in software engineering faces different choices from one whose customers are beginning to explore AI in their core operations.

The first task is to understand what is happening in projects. Which activities are taking less time? Where does the team still need substantial effort to validate outputs or make an approach work in production? What new requests are customers bringing to the provider?

Those findings have consequences beyond the project. If a task takes less effort, the existing staffing plan may no longer fit. If a customer wants the provider to take responsibility for a broader result, the team may need different skills. Changes to delivery affect the talent a firm needs, the services it can credibly offer, and how it can price them.

These decisions also affect how firms organize delivery. Established staffing structures and project estimates may need to change as AI alters the effort required for different activities. Leaders must determine which changes can be introduced into current engagements and which require a different approach to execution.

The priority is to make these adjustments based on what current projects demonstrate, rather than assume that productivity gains in individual tasks will translate directly into improvements across entire engagements.

How Can Technology Services Teams Navigate Uncertainty in AI Projects?

Technology Services firms have long helped customers reduce uncertainty. They define requirements, estimate effort, and organize teams to deliver against an agreed plan. Those capabilities still matter, especially when customers depend on reliable execution.

An AI project can also expose a problem the plan did not anticipate. A promising approach may fail to meet a production requirement. The customer may discover that the original request addressed only part of the problem. A new capability may make a different solution possible.

The team then needs someone who can assess the situation with the customer, decide what should change, and carry that decision into delivery. Following the original plan more efficiently will not necessarily solve the problem.

This creates a talent question that goes beyond training teams to use AI tools. Firms need to identify people who can work with incomplete information, understand the customer’s operating context, and make sound decisions as requirements evolve. They also need to examine whether those people have the authority and incentives to act before a project loses momentum.

The implication for leaders is to identify where this judgment already exists within their organizations and give those responsible for delivery the authority to respond when project requirements change. Technical expertise remains important, but so does the ability to make decisions when established approaches no longer apply.

How Can Technology Services Firms Build Continuous AI Learning?

A firm can train thousands of employees on AI and still find that much of the material needs updating as tools and delivery practices evolve.

That makes teams working on live client projects an important source of learning. They can explain where an approach succeeded, why another failed, and what it took to move beyond a demonstration. Capturing these lessons while the experience is fresh helps other teams apply them to similar problems.

The difficulty for a large provider is moving learning from individual projects across the organization. Training programs need to incorporate what client teams discover in production, including unsuccessful approaches that others can learn from.

This requires a learning process that can keep pace with delivery. Firms need ways to identify useful practices, update training materials, and make relevant project experience accessible across teams. The value of that learning depends on how effectively it reaches employees working on similar challenges.

What Happens When Customers Want Outcomes but Still Buy People?

As enterprises explore AI-driven delivery, a gap is emerging between what they expect from Technology Services providers and how they contract for that work.

Customers are beginning to ask for AI agents while continuing to specify the number of full-time equivalents on a project. They may also expect providers to take responsibility for business outcomes while requiring fixed teams of specialists. These requirements can limit how providers structure teams and introduce new approaches to delivery.

That mismatch cannot be resolved by introducing a tool. The provider and customer need to agree on what will be delivered, what each party controls, how success will be measured, and how changes to the team or process will be handled. Pricing follows those decisions.

These conversations require trust. Customers need confidence that changes to team structures and delivery responsibilities will not compromise the outcomes they expect. Providers can build that confidence by demonstrating what works in existing engagements and agreeing on how new approaches will be evaluated.

Client adoption is therefore a critical part of the Technology Services transition, influencing how quickly providers can introduce new delivery models and commercial arrangements.

How Can Technology Services Firms Prepare for AI-native Applications?

Much of today’s AI activity improves how existing software and products are built. The next phase could involve applications in which AI plays a central role in how software functions, rather than primarily improving the process of developing it.

This would introduce different delivery requirements. Teams may need to design applications around AI agents, monitor their behavior during execution, manage exceptions, and establish when human intervention is necessary. The responsibilities associated with maintaining these applications could also differ from those of conventional software.

The scale and pace of this shift remain uncertain. Providers therefore need to assess which emerging capabilities are relevant to their customers, what expertise they will require, and how much investment is justified before demand becomes established.

Preparing for AI-native applications will require choices about future service offerings and technical capabilities, even as firms continue to support today’s software environments.

What Will Define Successful Technology Services Delivery in the AI Era?

AI is changing the assumptions on which Technology Services firms have organized delivery for years. The implications extend across project execution, talent, organizational learning, customer contracts, and the capabilities required for future applications.

For leaders, the challenge is deciding how quickly to change each of these elements while continuing to deliver on existing commitments. Some decisions will be informed by proven applications of AI, while others will need to be made before the longer-term direction becomes clear.

The ability to make those decisions, act on them, and adjust as the industry evolves will be central to how Technology Services firms navigate the transition.