Tech services and generative AI: Plotting the necessary reinvention

Through most advances and innovations in enterprise technology, IT or technology services providers (also known as system integrators [SIs] or managed services providers [MSPs]) have, for the most part, been able to rely on one constant: their customers continuing to look outside their four walls for help stitching together and overseeing the ever-changing, complex web of hardware, software, networking, and storage products that drives their businesses. Now, enterprises are funneling more of their technology spending into generative AI (gen AI) and leveraging its capabilities to streamline or automate some of these same IT management services. Services providers could be forgiven for wondering what the future holds for them.

As it turns out, the future could be quite bright. Gen AI is also increasing demand for a wide range of new services, which represents a significant opportunity for providers to reimagine and recharge their business. Just how significant? We estimate the emerging market for services relating to gen AI/AI could be worth more than $200 billion by 2029. If tech services providers can succeed in claiming a decent share of that incremental value, their profitability could grow by as much as 30 percent. Reaching those heights will require providers to reinvent how they do business. That challenge will include transforming service offerings and how they are delivered, embracing new go-to-market (GTM) and commercial models, and upskilling teams while finding new sources of talent.

The next 12 to 18 months will be pivotal. Enterprise customers are already exploring new ways to manage some core IT work themselves while ramping up all manner of gen AI pilots and initiatives, and a wide range of other tech players (from hyperscalers to hardware and software companies) are beginning to make or contemplate moves into the burgeoning AI services market. Traditional services providers that don’t begin to reimagine their value proposition in this arena risk losing some relevance—and potentially more than 15 percent in revenue and profit.

To help services providers reposition themselves to thrive in the gen AI era, we surveyed 100 top industry executives and technology decision makers, interviewed a number of experts in the sector, and conducted an analysis of the state of gen AI adoption in a representative sample of the Forbes Global 2000 ranking of the largest enterprises. This article, which is also informed by our experience in the market, examines how the new technology is affecting enterprise technology spending patterns, what new gen AI services providers can offer to enterprise customers at different stages of gen AI adoption, and what critical steps providers can take to position themselves for this new era.

Gen AI’s impact on enterprise tech spending and provider economics

Most companies have been working to implement and scale traditional AI and automation solutions for more than a decade. However, the launch of OpenAI’s ChatGPT solution in late 2022 has led to a paradigm shift in enterprise AI priorities. Organizations are now turning to gen AI to help power and reinvigorate traditional AI initiatives while launching entirely new gen AI efforts in multiple functions. But after the initial waves of excitement and hype that greeted gen AI’s arrival, the enterprise customer base is now squarely focused on seeing the transformative technology live up to its billing. The goal is to move from piecemeal efforts, isolated pilots, and proofs of concept (POCs) to scalable solutions that can be deployed across organizations.

As part of this gen AI reset, organizations have a better understanding of the strategic and financial commitment required to create significant impact. As one executive told us, “Scaling AI is hard. It requires new skills, new processes, and new ways of working. It’s a transformational challenge for most organizations.”

In the latest McKinsey State of AI survey of enterprise customers, 67 percent of respondents said they expect their organizations to spend more of their technology budgets on AI/gen AI over the coming three years. Based on our analysis, we estimate that overall tech budgets will grow at a CAGR of about 4 to 6 percent over the next five years. But where that is spent may change considerably, with traditional value pools changing and new ones (and new competitors) fast appearing. Gen AI is already causing various tech ecosystem players to expand their range of offerings; for instance, hyperscalers are providing integrated development environments to build, train, customize and deploy large language models (LLMs), while some hardware vendors are venturing into services that enable development of custom LLMs and micro-services for faster deployment of gen AI applications.

New competitors are just one aspect of the disruption. With enterprises funneling a greater portion of tech budgets to AI/gen AI efforts, spending on services providers’ core offerings could drop or stagnate across a few traditional areas of spending. The three main categories of tech services spending to feel this effect are as follows:

• Outsourced traditional services. The long-standing foundation of tech services providers will likely see a decline of 8 to 10 percent as advancements in automation and AI from cloud platforms solve traditional IT challenges.
• Insourced services. Although internal IT teams will leverage low-code/no-code platforms, new data/AI apps, and infrastructure-management-automation tools to grow their portfolio, overall spending on this category will still likely stay flat, with the cost of additional workloads largely offset by productivity gains.
• On-prem, co-location (colo), and private cloud. The secular shift to public cloud and cloud-based graphics processing units (GPUs) is likely to cause a slow, steady decline in this area across most workloads. However, a few specialized and sensitive AI workloads (for example, public sector, healthcare) may drive some spending growth.

At the same time, a wealth of new AI services opportunities are developing thanks to increased spending across a number of categories. The largest of these include the following:

• Outsourced AI services. These services are expected to enjoy double-digit increases as enterprises grappling with the scarcity—and expense—of gen AI talent and their own inexperience developing and implementing gen AI solutions seek external help. This increased spending will focus on AI foundational services (for example, AI for IT operations [AIOps]), gen-AI-led productivity solutions (for example, coding copilots), and vertical-industry solutions (for example, clinical trial acceleration in life sciences, hyperpersonalized B2C solutions for telecom).
• Outsourced digital services. Increased demand for digital services like cloud, legacy modernization, and data and analytics will likely drive growth by 8 to 12 percent as enterprises leverage gen AI across existing digital initiatives.
• Enterprise applications. These applications should see solid growth as enterprises adopt a new generation of gen AI apps and solutions that solve traditional IT challenges.
• New AI stack solutions. This new spending category (largely for LLMs and related spending) is likely to grow rapidly as enterprises increase their adoption of gen-AI-native offerings such as foundational models, tooling, and data architecture.
• Public cloud and infrastructure as a service. Gen AI should drive increased enterprise cloud migration and consumption (for example, LLM training on cloud), fueled by the rise of AI-specialist cloud providers.
• Compute hardware (for example, GPUs). This category is likely to experience growth as enterprises use more advanced AI/gen AI custom chips, with some of this spending captured by public cloud providers making their own investments in the advanced hardware.

The full impact of gen AI on enterprise tech spending could be dramatic. According to our analysis, the disruption could unlock a $200 billion-plus market opportunity for tech services providers in the next five years, primarily centered around AI foundational services, AI-first horizontal solutions, and vertical-growth solutions. Companies are likely to overhaul their tech spending and reallocate their budgets along the lines of the global enterprise macro shifts laid out in Exhibit 1.

The alternative to seizing that potential opportunity may be even more daunting. Our analysis highlights the risk of inaction for both the top and bottom lines of companies in the sector (Exhibits 2 and 3), with a 15 percent cut in both a real possibility. The growth in automation and insourcing of certain workflows ushered in by gen AI is likely to spur a sizable reduction in demand for providers’ traditional services, while productivity gains and added competition could lead to pricing pressure. By contrast, embracing and adapting to the new technological era offers the prospect of not only maintaining the industry’s 3 to 5 percent historical growth trend but also improving its financial position, with a further 2 to 4 percent revenue boost and a potential 30 percent profit bump, based on the provider’s starting point.

The enterprise gen AI journey

In the relatively short period of time since enterprises started working with gen AI, they have already begun to evolve how they navigate the technology. For starters, they are now facing up to the myriad challenges many have already experienced in attempting to adopt gen AI: high costs, particularly from trying to leverage LLMs on a large scale; insufficient cloud and data readiness of the tech stack; the reliability of gen AI output; and uncertain risks related to regulation, intellectual property, ethics, and more. Instead of relying on one vendor, enterprises are increasingly using multiple model providers to develop their gen AI use cases. They’re also turning to more affordable open-source models that they can control and fine-tune rather than attempting to build custom models from scratch, taking the approach of what previous McKinsey research has dubbed a gen AI model “taker” or “shaper” rather than the more costly and complicated path of a “maker.”

While these types of shifts are widespread, large enterprises are at very different stages of gen AI adoption. Based on our analysis, organizations can be classified into the following three stages of gen AI readiness and adoption along their journey (Exhibit 4):

• Observers. The bulk of enterprises (50 to 60 percent) are “observers” focused on AI readiness (setting up their data and cloud foundations) and conducting small-scale POCs primarily for internal use cases (for example, text summarization, knowledge management) that are largely anchored on driving higher productivity.
• Front-runners. Another 30 to 40 percent are “front-runners” with a clear vision for using AI/gen AI to reduce costs at scale through use cases such as AI-enabled service desks and legacy-code upgrades. These organizations have already made changes to their organization and operating model, such as creating AI centers of excellence (COEs) or new roles like chief AI officer.
• Innovators. Less than 10 percent are true AI-first “innovators” with an integrated strategy to achieve cost reductions and unlock the true potential of gen AI through cutting-edge use cases like gen-AI-enabled product design and development that can propel future business growth.

Just as companies are at different stages of their respective gen AI journeys, they also require a range of tech services to propel them forward. To satisfy that demand and launch their own gen AI journey, providers should strongly consider rolling out three distinct types of AI services over an initial 12- to 18-month period. They are as follows:

• Reimagined traditional and digital services lines that leverage gen AI for improving delivery productivity and innovating current services offerings (for example, knowledge management, agent copilot).

• Foundational AI/gen AI offerings that prepare the enterprise tech stack for cloud and gen AI readiness (for example, data on cloud) and/or gen-AI-native services related to LLMs or tooling (for example, building comprehensive, end-to-end services ecosystems related to new players such as OpenAI, Cohere, and others).

  Importantly, the second, foundational class of services may provide a critical entry point for providers to expand their relationships with enterprises and help shape their innovation agenda. Services providers can now target the “new AI stack” of tooling/security, foundational models, and data and databases sandwiched between the applications and infra/cloud layers, which were previously the main focus areas of the traditional services providers’ play (Exhibit 5). While services providers may encounter difficulty gaining ground as new entrants in these areas, they are more likely to develop extensions that make existing solutions more powerful, cost effective, and easier to implement than to attempt building their own models or tools from scratch.

• Vertical solutions that target product/process innovation and revenue growth initiatives of enterprises through deep, vertical-specific use cases (for example, gen-AI-enabled connected products and manufacturing ecosystems in the industrial segment, insurance claims processing platform in finance). Services providers’ strong track record of targeting verticals with data and analytics services should prove helpful.

Because of the relatively small number of enterprises that already qualify as gen AI innovators, the vast majority (about 80 percent) of more than 10,000 AI services deals expected over the next year will probably fall in the first two categories of services. They will likely be largely focused on foundational AI readiness and cost reduction POCs through horizontal use cases involving knowledge management, customer service, and text summarization. Our research shows that there is a high degree of correlation between enterprise AI readiness and cloud readiness, so it will be critical for services providers to position themselves as end-to-end, holistic AI transformation partners, starting with getting the core data and cloud elements in place. Given that cloud transformation has been a key part of services providers’ growth in recent years, they should be well positioned to guide customers to the next fundamental digital overhaul with AI and gen AI.

Still, the growth potential of these categories of services is limited. There is a strong possibility that the market for productivity improvements and cost reduction will fast become commoditized as both enterprises and providers increase investments in an area that lacks much room for differentiation. For a provider to thrive in this space, it will want to seriously consider building ready-to-implement, packaged AI-readiness offerings that enable rapid and at-scale margin realization, modeled on the platform approach they have used successfully with cloud solutions over the past decade.

As more organizations become comfortable and competent with implementing gen AI at scale over the next couple of years, AI services deals will likely focus as much if not more on product innovation and revenue growth. The providers that can build full-stack, vertical-specific platform solutions and lighthouse industry use cases focused on these two areas will be best positioned to stand apart. Maintaining that advantage for the long term, however, will demand more; providers will need to prove that they can continuously adapt and respond to the accelerating pace of gen AI innovation while enabling sustained impact for their enterprise customers.

Rewiring providers for the gen AI age

The prospect of providers turning themselves into gen-AI-services-driven dynamos is compelling, but turning that vision into a new reality will require them to make a series of complex, challenging shifts across their entire organizations.

Build a new, broad-based AI-services catalog, including ready-to-deploy accelerators

As stated earlier, service providers can jump-start their gen AI business by rolling out a wide array of new offerings that include foundational-AI-readiness solutions across both data and gen AI tools, packaged cost-reduction solutions that can enable rapid and at-scale margin realization, and vertical AI use cases in pilot industries that focus on product innovation and revenue growth initiatives. The true differentiator, however, is developing ready-to-deploy accelerators or solutions (for example, prompt library, source attribution tools) across the gen AI tech stack to accelerate the development and adoption of existing gen AI solutions on the market. Providers have had success with this model for cloud, making implementation of other vendors’ products faster, easier, and more affordable, but it doesn’t happen overnight. It takes at least 15 to 20 individual client implementations to fine-tune the accelerators and equip them with the proper functionality and systems as they evolve and mature through repeated usage. Those players that can develop effective accelerators early on will likely have a decided advantage as the gen-AI-services market starts to take off.

Rethink GTM and commercial models

The sector’s time-tested approaches to landing, structuring, and delivering on deals aren’t likely to have the same success in this burgeoning era. Services providers may no longer be able to wait for sufficiently large deals to materialize. To stake an early claim to the emerging gen-AI-services opportunity, they should strongly consider embracing volumetric deals—experimenting and conducting small-scale POCs with a number of clients, as leading providers have already begun to do. Learning on the go will be key for both providers and their customers as the technology (and use cases, solutions, implementations, etcetera) evolves at a breakneck pace. At the same time, providers will want to work with certain existing customers to create larger, pioneering gen-AI-led transformation deals. Many if not most of these newfangled deals will be outcome based with a significant gain-share component. As IT providers compete more with internal development teams and off-the-shelf SaaS solutions, apps, and low-code/no-code platforms, they will be forced to prove their worth more than ever before. To win deals and gain market share, services providers may need to embrace innovative commercial models; for instance, fees would be linked to the number of tickets resolved by a customer-support gen AI system, or to the time saved per salesperson through gen AI augmentation of customer relationship management tools. Developing systems that can accurately measure the productivity gains attributable exclusively to gen AI will be critical.

Develop a new AI talent model across build, sell, and deliver functions

Services providers will need to create a wide range of new roles, such as a responsible AI lead to establish policies, principles, operating models, and controls to govern and ensure the ethical and safe use of AI across the enterprise. Other new positions may include AI product managers on large-app-build programs, while sales leaders will need to have a broader technical and consultative skill set and be capable of answering customer inquiries on a range of topics, such as how to calculate the ROI of gen AI implementation or whether to approach custom foundation model development through prompt engineering, retrieval-augmented generation (aka RAG) or fine-tuning.

Adopt a new partnership and M&A approach to bring together the emerging gen AI ecosystem

Most enterprises access LLMs through APIs (leveraging cloud services providers, model providers, or data cloud players) instead of hosting them themselves, which makes it critical for services providers to strengthen partnerships with this growing roster of players. But while services providers have traditionally only formally linked up with vendors once they have reached a certain scale with enterprise customers, the unusually fast pace of gen AI adoption and experimentation dictates that they take a different tack. If they want to play a key role in bringing “best of breed” ecosystems spanning the “chips to apps” tech stack, they will have to partner with smaller or fledgling software companies as well as the larger, more established ones they are more used to working with. At the same time, although services providers haven’t widely employed programmatic M&A in recent years, that is likely to change, given their need to quickly acquire key gen AI capabilities, such as LLMOps (LLM operations), gen-AI-led app development, and AI-led FinOps (financial operations).

Modernize delivery model through AI COEs to enable higher efficiency for customers

Early evidence suggests that gen AI can enable 20 to 40 percent productivity improvement in tech delivery across traditional service lines through new solutions such as coding copilots or agent copilots to aid knowledge management in application development and maintenance services. Still, services providers will need to develop a holistic approach to realize the full potential of gen AI on their delivery model. The linchpin of that approach will be a new COE to effectively coordinate and scale cutting-edge AI services delivery efforts across the organization. Staffed with a mix of technical and industry experts who collaborate with leaders in different functions and parts of the business, COEs have typically been used for providers’ cloud and digital and analytics initiatives. But gen AI’s rapid pace of change and complexity means that this new COE will require greater depths of expertise and levels of collaboration than previous similar efforts. Most large providers have already launched initiatives in this direction.

Even if services providers manage to pull off these various transformations, there are still fundamental risks to grapple with as they embrace gen AI. Perhaps most important, as enterprises struggle to understand how to scale this fast-evolving technology to fuel new growth, is that services providers will have to deal with the potential for significant scope creep and, as a result, possible reputational damage. Specific areas they will want to pay careful attention to include a lack of responsible AI controls as enterprises overlook or deprioritize important practices such as toxicity mitigation, ethical considerations, and regulatory compliance; escalating costs and unclear ROI as a lack of clear business value alignment and ROI tracking threaten to turn certain projects into unsustainable financial burdens; insufficient customer focus and readiness, with enterprises trying to implement gen AI across too many functions and revenue streams without thorough prioritization and evaluation of their own organizational preparedness; and a lack of requisite customer experience, causing some to rush into utilizing too many or simply ineffective tools and systems available on the market that could prove very costly down the line.

As with most industries and aspects of business, gen AI offers tech services providers both daunting challenges and significant opportunities. With enterprises reallocating tech spending to harness gen AI to take over some of the traditional IT services that providers initially built their business around, providers now have to reimagine their value proposition, business models, and organizational structures. It’s a sizable undertaking, but if they can pull it off, gen AI has the potential to fuel the sector’s next S-curve of growth, with the prospect of a $200 billion–plus AI services market over the next five years. But even as they undergo a major transformation to gain their share of that new opportunity, providers will also need to rely on their traditional role as “neutral” technology advisers, bringing customers the best-of-breed offerings in an evolved, complex technology stack.