Bridging the Enterprise AI Implementation Gap: Where Value Is Moving in IT Services

AI is creating an unusual dynamic in the IT Services and Tech-Enabled Services markets. The same technology that threatens to automate large parts of traditional outsourced development and business-process work is also creating what may be the largest new services opportunity in decades: bridging the enterprise AI implementation gap.

Enterprises now have ready access to extraordinarily capable AI models. What most lack is the ability to deploy those models securely and productively across their businesses.

Enterprise data is fragmented and poorly structured. Legacy applications weren’t built for AI. Proprietary data has to be located, cleaned, governed and made accessible. AI applications need to be woven into existing workflows. Security and permissions have to be established, outputs monitored, and employees retrained on how they work. Once an AI application is deployed, someone has to operate and continuously improve it.

The gap between what AI can theoretically do and what enterprises can actually implement is becoming one of the defining opportunities for technology services companies.

The data backs this up. EY counted 449 IT Services M&A transactions worth roughly $14.8 billion of disclosed value in the first half of 2026 — essentially matching the 456-deal pace of H1 2025 despite economic uncertainty and AI-driven disruption. Cloud, data and analytics, cybersecurity and managed services led transaction activity.

More tellingly, enterprise buyers see outside services providers as necessary to AI deployment. KPMG’s 2026 Managed Services Outlook, based on 1,224 senior leaders at large global organizations, found that more than 90% view managed services as important to agentic-AI delivery. Among U.S. respondents, 93% said managed services matter for agentic AI delivery, and AI capability is now the top consideration when selecting a managed-services provider.

The strategic conclusion for services-company CEOs: the opportunity isn’t just to get more productive with AI. It’s to build the capabilities enterprises need in order to use AI themselves.

The Services Business Is Changing, Not Disappearing

AI is a real threat to the traditional IT Services model. A company that earns revenue by putting 100 developers on customer projects and billing for their time has a problem if AI lets 60 developers produce the same output — under a conventional time-and-materials model, most of that productivity gain flows straight to the customer through fewer billable hours.

But a services company that understands a customer’s systems, data and processes, has specialized AI engineering talent, owns proprietary accelerators and connectors, and runs the resulting system under a recurring managed-services contract sees a very different outcome. AI productivity can raise margins, increase the number of customers each employee supports, and convert human expertise into scalable intellectual property.

The dividing line in services M&A is increasingly between companies that sell labor and companies that use specialized labor to build scalable capabilities and outcomes.

Several transactions TechStrat has managed over the past five years illustrate the building blocks of this emerging model.

1. Own the Customer Relationship: Solvd / EastBanc Technologies

One of the most valuable assets in the AI implementation market may be something decidedly nontechnical: a trusted, long-term relationship with the enterprise customer.

In 2024, TechStrat advised EastBanc Technologies on its sale to Solvd, a portfolio company of Siguler Guff. EastBanc brought AI consulting capabilities to a larger global software-engineering and digital-transformation platform, and Solvd described the deal explicitly as repositioning for the AI era — pairing its own engineering capabilities with EastBanc’s AI expertise.

AI implementation rarely starts with a customer calling an unfamiliar provider to redesign the enterprise around AI. It starts inside an existing relationship. A provider that already maintains applications, develops software or manages infrastructure for a customer has learned how those applications work, where the data lives, which systems talk to each other, where the technical debt is, and which workflows cause problems. That knowledge becomes extraordinarily valuable in the AI era — the incumbent can spot opportunities an outside AI consultant would never see, and move from maintaining the customer’s environment to helping redesign it around AI.

The strategic objective should be to deepen enterprise relationships rather than commoditize them. Long-term customers aren’t just a source of recurring revenue; they’re the installed base through which a services company introduces AI assessment, data preparation, workflow redesign, application modernization, AI implementation and, eventually, managed AI services. Customer tenure itself becomes part of an AI strategy.

2. Own Unique Data: SOLVE / MBS Source

If customer relationships create the opportunity to implement AI, proprietary and hard-to-replicate data supplies the raw material that makes the resulting AI valuable.

TechStrat served as exclusive M&A advisor to MBS Source in its 2025 acquisition by SOLVE, a provider of data, analytics and predictive pricing for fixed-income markets. MBS Source brought specialized mortgage- and asset-backed securities data and analytics to SOLVE’s fixed-income platform — aggregating real-time information from more than 200 dealer and institutional relationships, with 18 years of historical pricing spanning more than 150 million records. SOLVE is now applying AI-enabled parsing and analytics to that dataset.

This gets at one of the key distinctions between generic and differentiated AI: access to the same foundational model rarely provides a sustainable edge, since competitors can buy access to it too. Access to data competitors don’t have can. That’s especially true in opaque markets like fixed income, where proprietary historical pricing, quotes and trading activity meaningfully improve analytics and predictive capability.

For Tech-Enabled Services companies, the implication reaches well beyond financial markets. Companies should inventory the information flowing through their businesses and ask what they see, because of their customer relationships, that others don’t. Years of providing a service tend to generate proprietary benchmarks, transaction histories, equipment data, pricing, operating statistics, claims data or other specialized datasets — much of it incidental at the time it was collected. Properly permissioned and governed, that data becomes a strategic asset in an AI environment.

3. Vertical Data Can Be Particularly Powerful: HData / Insight Engine

The same principle shows up in a completely different market. In 2025, TechStrat advised Advanced Energy United and Insight Engine on the sale of Insight Engine to HData. HData pairs purpose-built vertical AI with a centralized library of regulatory filings, updated continuously and extending back decades. Insight Engine added regulatory and legislative intelligence, expanding the combined company’s reach across utilities, power producers, regulators, financial institutions and energy companies.

A generic LLM knows a lot about energy. A system built to answer commercially important questions about energy regulation needs decades of regulatory filings, current proceedings, legislative developments, specialized classification and reliable source material — a far harder thing to assemble.

The most defensible AI strategies may not involve building better foundational models at all. They may involve combining commercially available models with proprietary vertical information and domain expertise. For CEOs: what proprietary or alternative data exists in your vertical, and can you acquire, license, collect or gain access to it? M&A is often the most effective tool for answering that — an acquisition doesn’t need to add revenue or headcount; it can add a dataset that materially increases the value of the acquirer’s existing services and technology.

4. Own the Talent Pipeline: HTEC / Cognits

AI doesn’t eliminate the importance of human capital — it changes which capital is scarce.

TechStrat advised Cognits on its 2025 sale to HTEC, an AI-first global digital product-development and engineering company. Cognits brought a long-established delivery and recruiting operation in Guatemala and a growing talent base across Latin America, giving HTEC a platform for regional expansion.

Acquiring more developers might look inconsistent with a thesis that AI reduces the value of outsourced development, but the asset here wasn’t cheap engineering capacity. It was an established mechanism for identifying, recruiting and developing technical talent in a strategically important region, with same-time-zone delivery and real experience serving sophisticated U.S. clients.

In an AI-driven services market, the scarce resource is shifting from generic development capacity toward people who can do AI and data engineering, integrate models with enterprise systems, design agentic workflows, modernize legacy applications, build data infrastructure, implement cybersecurity and AI governance, and pair technical skill with industry knowledge.

Being a source of developers is becoming less differentiated. Being exceptionally good at finding, developing and deploying scarce AI-era talent remains highly valuable — the recruiting engine, training methodology, geographic footprint, university relationships and employer reputation behind the engineers may ultimately matter as much as current headcount.

5. Before AI Comes Data Readiness: Resultant / Teknion

Much of the current conversation about enterprise AI starts too late. Before an enterprise can deploy sophisticated AI, it needs usable data.

In 2022, TechStrat advised Teknion Data Solutions in its acquisition by Resultant. Teknion had spent more than 20 years on enterprise data problems — data warehouses, integration, governance, data quality, interoperability, visualization and business automation — and brought more than 150 recurring client relationships. Resultant specifically flagged the opportunity to add Data Governance and Data Quality as a Service to its managed-services offering.

The deal predates generative AI’s breakout, but its relevance has only grown. Intelligent models don’t repair poor enterprise data on their own. KPMG points to fragmented data, inconsistent integration and technical debt as structural barriers to scaling AI, and reports that 98% of senior leaders now consider AI implementation a critical capability they demand from managed-services providers. “Garbage in, garbage out” has become more consequential, not less.

Services companies that can help customers discover, integrate, normalize, govern and maintain their data occupy an important position in the AI value chain. This work is also hard to separate from the customer’s underlying business — understanding what the data means, and which version is authoritative, requires real institutional and industry knowledge. That creates room for services businesses to move upstream into enterprise data architecture and governance, and downstream into ongoing managed data services.

6. Data Must Be Made Useful, Not Merely Accessible: Calligo / Decisive Data

TechStrat saw a version of this thesis even earlier. In 2021, TechStrat advised Decisive Data on its acquisition by Investcorp-backed Calligo. Decisive Data specialized in data analytics, data science and visualization; Calligo folded those capabilities into its existing managed data, privacy and security offerings to move from traditional managed IT toward higher-value managed data insights.

The problem then sounds remarkably contemporary now. Enterprises had enormous amounts of data, but it was fragmented, inconsistently maintained and hard to turn into useful analysis. Calligo’s bet was that customers cared more about the business outcome analytics produced than about owning the underlying infrastructure. Generative AI only magnifies that opportunity: an enterprise may hold millions of documents, CRM records, customer interactions, product records, service histories and financial transactions, but simply connecting an LLM to that pile doesn’t produce a reliable AI system. The information still has to be identified, classified, cleaned, permissioned, contextualized and maintained.

The lesson for services businesses: move beyond managing customer infrastructure and toward managing the customer’s information environment. That’s much closer to the source of value AI is trying to unlock.

Six Building Blocks of the AI-Era Services Company

Taken together, these transactions sketch a useful framework for value creation in services companies:

• EastBanc — the value of customer access. Long-term relationships create the opening to identify and implement AI use cases.
• MBS Source — the value of proprietary data. Unique information differentiates otherwise widely available AI technology.
• Insight Engine — the value of vertical data and domain context, turning generic AI into commercially useful intelligence.
• Cognits — the value of a talent engine, as AI reduces demand for undifferentiated coding capacity while raising the value of scarce technical talent.
• Teknion — the value of data readiness. Enterprise AI can’t scale until the data underneath it is integrated, governed and reliable.
• Decisive Data — the opportunity to turn prepared data into ongoing intelligence and managed outcomes.

Together, they map much of the AI implementation stack:

Customer Relationship → Enterprise Data → Proprietary/Alternative Data → Specialized Talent → AI Implementation → Managed Intelligence

A services business doesn’t need to own every layer. But companies that own several of them are likely to be worth substantially more than companies whose principal asset is a pool of billable labor.

M&A Is Already Moving Toward These Capabilities

The broader M&A market backs this up. EY reports that cloud, data and analytics, cybersecurity and managed-services providers led IT Services deal activity in H1 2026, with AI pushing buyers toward companies with greater scale, sector depth and ecosystem strength.

Large strategic transactions show the same pattern: buyers are pursuing combinations that join AI with data, vertical expertise, engineering, cloud infrastructure and recurring operations — not AI capability on its own.

The contrast with software is notable. S&P Global reports that PE- and VC-backed application-software transaction volume fell 21% in 2025, to 3,665 deals from 4,638 in 2024, as investors question how AI will affect the growth and defensibility of software companies acquired before generative AI reshaped the competitive landscape.

That leaves an unusual M&A environment: investors questioning the durability of some software businesses while actively hunting for companies that can help enterprises implement AI and transform how technology gets used.

The Investment Priorities for Services CEOs

For services-company CEOs, the response should go well beyond buying AI tools for employees. AI productivity matters, but on its own it may simply let customers buy fewer hours. The bigger opportunity is to use this transition to build assets that raise the company’s strategic value — across five areas:

1. Deepen enterprise customer relationships. Move from project vendor to strategic partner. The better a provider understands a customer’s systems, data and workflows, the better positioned it is to spot and implement AI opportunities.
2. Build a data strategy. Identify what proprietary information the company already has or can access, what vertical datasets could be acquired, and how customer data expertise becomes a service offering.
3. Invest aggressively in data readiness. Data engineering, integration, quality, governance and security aren’t ancillary to enterprise AI — they’re prerequisites.
4. Transform the talent engine. Recruit for AI, data, cybersecurity and vertical expertise instead of simply expanding generic engineering capacity.
5. Productize expertise. Every engagement should generate reusable IP — connectors, agents, workflow libraries, industry models, governance tools, implementation frameworks — so the company does more for each customer without adding headcount proportionately.

And finally, convert implementation into recurring managed services. Building an enterprise AI application is a project; governing its data, monitoring its output, maintaining integrations, controlling security and continuously optimizing performance is a long-term relationship. KPMG’s finding that more than 90% of executives already see managed services as important to agentic AI delivery suggests customers are heading in exactly this direction.

From Labor Arbitrage to Intelligence Arbitrage

For decades, IT Services created value largely through labor arbitrage — finding qualified people in lower-cost markets and selling their time to higher-cost enterprise customers. AI won’t kill that model overnight, but it’s steadily undermining its basic unit of value: the billable hour.

The next generation of high-value services companies will practice something closer to intelligence arbitrage. They’ll know more about a customer’s systems because of long-term relationships. They’ll understand the customer’s data better than outsiders do. They’ll combine that enterprise knowledge with proprietary and alternative datasets, recruit scarce specialists who can deploy new technology, capture what those specialists learn as reusable IP, and increasingly sell managed capabilities and outcomes rather than hours.

The enterprise AI implementation gap is what makes this shift possible. Models are advancing faster than most enterprises can reorganize their technology, data and operations around them, and that gap won’t close simply because another generation of software arrives. Enterprises need partners who can do the hard work between the model and the business.

That’s where we believe a substantial share of value creation in IT Services and Tech-Enabled Services is heading. For CEOs and investors, the goal isn’t just to make today’s services business more efficient with AI — it’s to use AI to build tomorrow’s more valuable one: deeper enterprise relationships, differentiated data, scarce talent, proprietary IP, and recurring responsibility for the AI-enabled systems customers increasingly depend on.