Software ate the world. Now AI is eating software—and it’s coming for digital transformation services. 

The next five years are likely to see relentless demand for digital transformation services. Fortunes will be made. But the exponential advancement of AI is going to result in disruptive forces that will thereafter deeply disrupt the industry. Many digital transformation firms will not survive. 

But services firms that can also bring to bear the expertise, organizational structure, and capitalization necessary to build durable products have an opportunity to create extraordinary and durable asset value. 

What is Digital Transformation?

Digital transformation (or “DX”) services are a large (approximately $375B in the US in 2023)¹ subset of the digital services industry that involve leveraging technology to improve business systems and processes. Some key areas of focus include: 

• Cloud Migration
• Legacy System Modernization
• Business Process Automation
• Data Analytics and Business Intelligence
• Sales Enablement
• Customer Experience Enhancement
• Internet of Things (IoT) Implementations
• Cybersecurity

Why Outsource Digital Transformation?

There are powerful reasons why it might make sense for organizations to outsource their digital transformation needs, which is why virtually every organization does so to some extent. Digital transformation is challenging and often involves temporary sprints that make sense to outsource to experts. 

Expertise: Most digital transformation requires specialized skills and knowledge that might not be available in-house, but can be provided by consultants.

Capacity: In-house IT teams may already be stretched thin with day-to-day operations, leaving little capacity for transformation projects. The cost of hiring and training a full-time team for a one-time project may also not make sense.

Speed: Consulting firms can often execute projects faster due to their experience, proven methodologies, and ability to dedicate full-time resources.

Objectivity and Risk: Outsiders may more easily challenge existing assumptions and processes that may be hard for insiders to tackle. This may also come into play for political and risk management purposes, where the ability to blame an outsider can be valuable. 

The Cons of Outsourcing

However, there are also powerful reasons not to outsource. 

Costs: Outsourcing generally involves higher costs than using in-house resources. In some cases, these cost differences can be substantial. 

Knowledge: Outsourcing means that potentially valuable expertise and knowledge remains outside the organization. This can reduce organizational adaptability and leave clients beholden to service providers. 

Control: Outsourcing typically means handing control over key parts of a project to an outside firm, which may have conflicting motivations and values. This loss of control can persist over time, particularly when proprietary knowledge is involved.

Security: Outsourcing typically requires sharing sensitive data and system access with an external party.  

Demand is Exploding

For the time being, outsourcing seems to be the consensus. DX service revenue in the US is expected to grow from $375B in 2023, to almost $1.1 trillion within five years—a CAGR of 23.9%.¹ 

The adoption of cloud computing and AI is driving much of the growth in this segment of the technology services industry. And given the exponential improvements in AI—and the implications for DX customers—these growth estimates may very well be an understatement.

This in turn is driving demand for data analytics and AI services, and leading to a shortage of talent in areas such as data science, cloud computing, and cybersecurity. All of this drives further demand for outsourced resources. 

And there is no reason to expect a slowdown any time soon. Margins and growth rates for digital transformation firms are likely to rise significantly in the coming years.

Planting the Seeds for Their Own Disruption

The eventual problem for digital transformation service providers is that bringing AI and cloud to clients may involve planting the seeds of their own disruption by delivering to their customers the capabilities required to eventually render themselves obsolete. It’s not that the need will go away, but rather AI will eventually change the decision calculus in favor of doing most digital transformation projects in-house.

Right now, there’s little risk of AI replacing most digital transformation tasks, much less entire jobs. Current models excel at many tasks, for example trouncing specialized PhDs in their fields on the GPQA test.² But they lack some chain of thought and reasoning skills required for broader use. And they have issues with data security, hallucinations, oversight, and bias, among others.

But AI is improving exponentially, and the scope of things it will be able to accomplish is increasing rapidly. It seems inevitable that AI will be applied to meaningful transformation tasks—likely even before AGI (Artificial General Intelligence) is achieved. And as AI grows in competence, it will begin to commodify the expertise of digital transformation experts. Once AI knows how to perform a particular task (or job), it opens the door for infinite and inexpensive³ workers to perform it at a scale. 

Suddenly the expertise, capacity, and speed arguments in favor of outsourcing will wane. Even the objectivity and risk arguments may come under threat from AI. Although it will be hard to blame AIs for mistakes, arguably they can be designed for greater objectivity than the humans that otherwise perform the same roles. 

When Will This Start?

It’s not possible to predict with any confidence when any of this might happen. But I’ve separately proposed a few scenarios that describe the most likely evolution of AI technological capabilities:

• Slow: Transformative change either never happens, or not within the next 10 years
• Moderate: Transformative change in the next 10 years
• Fast: Transformative change in the next 5 years (and maybe radical within 10)
• Radical: Radical change within the next 5 years

This refers to how quickly the technology evolves, not how long it takes to diffuse into the market thereafter. I’d argue that the Slow and Radical scenarios are unlikely (although Radical shouldn’t be counted out entirely). And the more I see, the more I believe we’re headed for the Fast scenario—transformative technological change in the next five years.

Software is particularly susceptible to AI disruption. We’re already seeing how much it can affect productivity, integration, user experiences, and more. Software development is an important part of most digital transformation projects. It’s also not a stretch to imagine how AI might encroach on—or replace—other important aspects of those projects. 

And while digital transformation is often quite complex, it’s also true that most projects are fundamentally about applying knowledge workers to repetitive technical tasks which AI will become increasingly capable of performing. 

As I mentioned in my article on the future of enterprise software, Cognition Labs announced what they touted as the world’s first AI software engineer (“Devin”) in March 2024, claiming it “can learn how to use unfamiliar technologies.”⁴ There are legitimate questions about Devin’s real world capabilities and operating costs (e.g., high volume of expensive OpenAI API requests). Nonetheless Devin’s advances in chain of thought (COT) and tool utilization predict a future where semi-autonomous bots are capable of doing very complex programming tasks—and even jobs. Extending this to many DX tasks is not much of a stretch.

That’s why the Slow scenario seems unlikely in this industry. AI is improving at an exponential pace. AI that can do every task better than humans is probably less than ten years away, and could arrive much more quickly—possibly within a year or two.

That’s why I think that banking on the Slow scenario as a technology services executive is wishful thinking. The likelihood that AI doesn’t transform the technology services industry is about nil. Even the Moderate scenario seems quite unlikely—AI is moving quickly. 

My base case for the digital transformation services industry is Fast—transformative change in technology capabilities within the next five years. And for reasons I’ll discuss, I suspect the diffusion rate of that technology into the market once technology capabilities are achieved will likely be fairly quick as well. 

I estimate the industry has perhaps five years of extraordinary growth, followed by 5-10 years of deep disruption that will displace many incumbents, leaving a much smaller set of survivors who have built durable expertise and reusable assets. 

How This Will Likely Play Out

Competition for AI products and platforms will be fierce and come from multiple directions, fueled by an extraordinary amount of capital: 

• Cloud providers will add offerings to expand their footprint (e.g., AWS’ AI offerings)
• Software incumbents will integrate AI capabilities into their offerings
• AI infrastructure companies will strive to increase the scope of their models’ capabilities (e.g., OpenAI’s PDF reading capabilities)

That’s why an entire wave of AI product and wrapper companies will likely come and go. In terms of product offerings (not services), only companies with deep technology and / or sustainable market advantages are likely to thrive. 

But all of this will benefit digital transformation firms in the near term. Focused on battling each other, the product and infrastructure firms will need to partner with service providers to address burgeoning demand for implementation and integration. I’ve already heard confirmation of this from multiple players in the space.

There will be so much demand that digital transformation firms will at first focus on data infrastructure (including cloud migration and optimization), data architecture, and integration of third-party AI solutions (as opposed to bespoke development). Successful services firms will go on hiring sprees, expanding their teams and infrastructure. 

But over time demand will be so insatiable—and talent in such short supply—that DX services firms will have little choice but to automate their own activities. Most will use third-party software to do this because the expertise, incentive structures, and capitalization models for services firms generally conflict with building durable assets. 

This automation will result in greater scalability and higher margins, but not necessarily durable competitive advantage or asset value for these service providers. Those advantages will accrue to two groups: the AI product companies selling “pickaxes” to the digital transformation “miners,” and a much smaller set of digital transformation companies that have the expertise, organizational structure, and capitalization necessary to build durable products. 

I believe it is these hybrid organizations that will end up owning the future of digital transformation. The early versions of these automation products will be “bionic,” requiring substantial human contributions. By combining real-time insights from the market with client relationships, these firms will have an unfair advantage in building the best products and establishing market footholds—and long term advantages. 

AI deep technology and infrastructure companies will create extraordinary value. But so will these hybrid product-service companies. As these products improve in capabilities, they will erode the fundamental benefits of outsourcing. And it will be the mainstream digital transformation firms that will have sown the seeds of their own disruption by using automations to do their client work. Inevitably, clients will recognize that they can implement their own AI workers to achieve expertise, scalability, control, and cost efficiencies. 

As more firms succeed in these implementations, more of their peers will follow. I anticipate this will reach a tipping point when demand suddenly declines for many digital transformation firms. The larger and less agile firms are unlikely to survive this transition.

But wait, you say. This will merely unleash extraordinary demand and DX services firms will still be needed to meet that demand.

I think that’s true for now. But at some point—potentially fairly soon—automation tools (such as AI workers) will beg the question: what is the benefit of an outsourced DX firm? The expertise, capacity, speed, and objectivity arguments won’t make much sense any more. 

So what’s left? What and why might firms continue to outsource? Clients will probably consider continued outsourcing to achieve:

• Best-in-class DX automation products and tooling
• Management and coordination of particularly complex DX projects
• Special technology and industry expertise, although the scope of that will diminish over time

But clients won’t outsource the bulk of the work any more—that will go to AI agents and tools. This will fundamentally change the economics and the profile of successful DX firms. 

As I have written separately, the evolution of automation in the advertising industry may be an interesting case study for other industries. At first, ad agencies benefited from media buying automation and the need to help clients navigate the complex new media world. But over time, automation and the commodification of digital media destroyed their margins, forcing them to acquire new capabilities. Some died. Many others were consolidated—often on poor terms. And from what I can see, it’s only getting worse for the ad industry. Continued automation and in-sourcing seem likely to force further consolidation and some notable failures. 

I suspect the DX services industry will follow a similar trajectory. The more adaptable and product-forward firms will have the advantage. Most of the rest will either be acquired or won’t survive.

1. “Digital Transformation Market Size, Trends Report, 2023-2030”. Grand View Research
2. Rein, David, et al. GPQA: A Graduate-Level Google-Proof Q&A Benchmark. 20 Nov. 2023
3. The cost of effective AI agents may currently be prohibitive, but is likely to drop quickly
4. Cognition Labs Blog, "Introducing Devin, the first AI software engineer". 12 Mar. 2024