The Build vs. Buy in Enterprise AI

We’re at a turning point where many enterprises are faced with a major challenge, which is to understand how to implement AI from the outer to the inner layer of their organizations.

Indeed, since 2022, we’ve seen a major explosion of executive roles tied to technological implementations.

A key reminder here, the executive roles will be critical to help redefine the overall company’s business model strategy, as it goes through “Incumbent Paradox:”

The incumbent paradox gives you a good news for the short-term, but a very bad one, for the long-run.

Indeed, in the first phase where the Incumbent in a sector will experience a distribution advantage, it will find solace in just implementing AI, without thinking too much about it and doing it only on the less strategic stuff.

That is tied to the distribution advantage of the incumbent.

In short, take any company, add an agentuc AI customer support layer, independently on whether that’s strategic or not, and you get a massive saving, now what?

Yet, and that’s the key take, this distribution advantage won’t last too long, and within a decade, the company will need to understand how to completely redefine its business model.

From that perspective, most enterprise businesses will have a plethora of innovation projects in their pipelines, with the major quandary: to build or to buy?

---

This is part of an Enterprise AI series of (possibly) daily pieces to tackle many of the day-to-day challenges you might face as a professional, executive, founder, or investor in the current AI landscape.

All these pieces are freely available to you. If you find the piece isn’t enough to help, you can contact me once you join in as a Founding Member.

---

In this specific piece, I’m tackling a very hard topic, that for those that have been working in the enterprise space for long enough know to be one of these issues that sound easy in theory, but extremely hard in practice.

As usual, faced with a complex real-world issue, we got to use a mental model, an heuristic, to help us drive change.

The Enterprise AI Adoption Matrix

For the sake of the Enterprise AI Adoption — as explored in the growing gap between AI tools and AI strategy — Matrix we’ll look at two key core perspectives:

• Strategic Importance: The extent to which AI adoption will pervate the company’s business model to its core (value proposition, key customers, and distribution).
• And the Technical Complexity: The degree of technical expertise, combined with infrastructure required to develop, deploy, and implement an AI project.

Let’s look at each of these two elements.

Strategic Importance

The strategic importance here needs to be addressed based on what elements of a business model it touches.

For instance, if you’re working only on the distribution side (e.g. you’re a retailer, and you start implementing AI agents for customer support), then it’ll be less strategic as a project, therefore more prone to fall into the “low strategic importance.”

When instead, you’re prioritizing your innovation pipeline based a project which might touch the core of your business model (value proposition, revenue model, or core IP), then the strategic importance becomes high.

As you move from low to high strategic importance, you’ll move to an innovation pipeline that might prioritize the build vs. the buy.

And yet, that’s not always as simple.

Technical Complexity

When it comes to technical complexity, here we got to look at several factors.

Of course, if you’re a large enterprise business, internally you might have all the competences needed to tackle even the most coplex projects.

And yet, even there, it won’t be easy to assess which project to bring in house, wich one not.

And when it comes to technical complexity, it’s not just about the technical expertise to build an MVP, but actually the technical expertise to build a solution that can be sent in production.

In addition, once the solution has been sent in production, is your team able to maintain it, without making the organizational structure — as explored in the new organizational architecture for the AI era — too complex?

Are you sure that it will make sense from an infrustructural cost?

As that will need to factor in, not only the consumption costs to run the application, but also all the potential security issues, compliance issues, and internal politics that come with it.

For that reason we want to classify the complexity that comes with a technical project, falling under three types of complexity:

• Base Complexity.
• Operational Complexity.
• And Organizational Complexity.

The more the technical implementation will move from base to organizational complexity, the higher will be the hidden costs that come with the implementation.

Strategic Importance vs. Technical Complexity

Once, we have mapped out these two factors, we can start building the innovation pipeline, to make sure to priortize the projects based on buy vs. build.

Just like on a surfboard, you want to adjust your position, based on where you are, and the type of wave you’re catching, you’ll find your sweet spot in the matrix, based on a case by case.

The Enterprise AI Innovation Pipeline

Again, the key point here is to have a general prioritization framework in place.

And yet, there might be ecxceptions.

For isntace, in general you might think that a high strategic project should always come into the “build area.”

However, take the case of being very late to the game, you’re trying to catch up, yet, even if a project has low complexity, it still might come with a lot of hiddend costs and delayed innovation pipelines.

Thus, a “temporary buy” might work out, to get quick, and cheap lessons leanred via a vendor, and use these also as an assessment on the future strategic implementation.

Possible Outcomes

In the plethora of the possible outcomes that might come, by balancing strategic importance vs. technical complexity, a key take is really to address the worst case scenario.

The Worst Case Test

On the strategic side, the worst case scenario, needs to be addressed on what it means to lose control of the core, as you move closer to the main value prop of the product/service.

On the technical complexity side, the worst case scenario is when the whole project might turn into an organizational mess.

Making sure an enterprise AI project is assessed against strategic impact, makes it an easier one to justify from an organizational change perspective.

Which connects to the last point.

The Tipping Point: When The Technical Turns Into Transformational

At a certain point, a project becomes so strategically critical that building it in-house demands a complete organizational overhaul.

This shift moves beyond technical implementation, touching the core of the company’s operations and requiring significant changes to its business model.

It transforms the organization from within, aligning the project’s success with a broader strategic renewal.

When a project reaches this stage, it is no longer optional—it becomes an inevitable step toward maintaining competitiveness and long-term growth.

This highlights the distinction between merely addressing technical needs and embracing a transformational approach that reshapes the company’s future trajectory.

As we go through the AI era, we’re seeing this sort of transofrmation across many of the incumbents companies, which have realized, if they want to survive and thrive, passed this phase, that required a whole re-org.

Recap: In This Issue!

The Enterprise AI Adoption Challenge

• Enterprises face a critical challenge: How to implement AI from the outer to the inner layers of their organization, addressing both tactical and strategic priorities.
• Since 2022, there has been a surge in executive roles tied to technological implementation, reflecting the growing importance of leadership in navigating AI integration.

The Incumbent Paradox

• Short-Term Advantage: Incumbent companies benefit initially from a distribution advantage by applying AI to non-strategic areas (e.g., customer support) for cost savings.
  • Example: Adding generative AI customer support layers that deliver immediate benefits.
• Long-Term Risk: This distribution advantage erodes over time. Within a decade, companies must redefine their business models to remain competitive, requiring deeper strategic integration of AI.

The Build vs. Buy Quandary

• Enterprises must decide whether to build or buy AI solutions as part of their innovation pipeline.
• Build vs. Buy decisions should balance:
  1. Strategic Importance: How critical the AI project is to the core business model.
  2. Technical Complexity: The expertise, infrastructure, and organizational capacity required to implement and maintain the solution.

Key Framework: The Enterprise AI Adoption Matrix

The Enterprise AI Adoption Matrix evaluates projects along two dimensions:

1. Strategic Importance:
   • Low Strategic Importance: Projects that focus on peripheral activities (e.g., distribution or customer support).
   • High Strategic Importance: Projects targeting core business components like value proposition, revenue models, or intellectual property.
2. Technical Complexity:
   • Base Complexity: Building simple MVPs or foundational AI features.
   • Operational Complexity: Deploying AI solutions that are production-ready.
   • Organizational Complexity: Maintaining and scaling AI solutions without creating unnecessary overhead.

Strategic Importance vs. Technical Complexity

By mapping these dimensions, enterprises can prioritize AI projects and determine whether to build or buy.

General Guidelines

• Low Strategic Importance & Low Complexity: Buy off-the-shelf solutions for cost-effectiveness and quick deployment.
• High Strategic Importance & High Complexity: Build in-house to retain control and foster innovation.
• Hybrid Scenarios: Use a mix of building and buying for medium-priority projects or when internal expertise is limited.

Exceptions to the Framework

• A “temporary buy” might be a viable option for high-priority projects when a company is late to the game and needs to catch up quickly.
  • Example: Leverage vendor solutions for rapid learning and assessment, with plans to transition to an in-house build later.

Addressing the Worst-Case Scenarios

Strategic Risks

• The worst-case scenario on the strategic side is losing control of your core business value proposition due to over-reliance on external solutions.

Technical Risks

• On the technical side, the worst-case scenario involves an AI project becoming an organizational mess, with spiraling infrastructure costs, compliance issues, and internal inefficiencies.

The Enterprise AI Innovation Pipeline

• A systematic framework to prioritize projects based on their position in the matrix, ensuring alignment with business objectives.
• The pipeline allows companies to balance short-term wins with long-term strategic goals, avoiding pitfalls like delayed innovation or hidden costs.

The Tipping Point: Technical to Transformational

• When a project becomes strategically critical, it requires in-house development and often a full organizational overhaul.
• AI integration shifts from being a technical solution to a transformational force driving long-term growth.

Conclusion: Building Resilience in AI Adoption

• The Enterprise AI Adoption Matrix provides a structured framework to manage the build vs. buy decision, balancing strategic importance and technical complexity.
• Enterprises must:
  • Continuously evaluate the strategic and technical impact of AI projects.
  • Adapt to unique circumstances while prioritizing innovation pipelines.
• At a tipping point, AI adoption shifts from being purely technical to transformational:
  • Strategically critical projects often require in-house development.
  • These projects may necessitate a complete organizational overhaul, impacting business models and operations.
• Addressing worst-case scenarios helps safeguard against:
  • Losing control of core business value due to over-reliance on external solutions.
  • Creating unmanageable operational challenges, including cost overruns and inefficiencies.

With massive ♥️ Gennaro Cuofano, The Business Engineer

---

This is part of an Enterprise AI series of (possibly) daily pieces to tackle many of the day-to-day challenges you might face as a professional, executive, founder, or investor in the current AI landscape.

All these pieces are freely available to you. If you find the piece isn’t enough to help, you can contact me once you join in as a Founding Member.