Issue #54 – Refactoring Your Business for AI

Instead of bolting AI onto a broken business model, it’s time to redesign how your business operates for this new world

Dylan Anderson

Mar 29, 2026

Read Time: 10 minutes

The AI world is evolving at an unprecedented pace. Nobody can truly keep up.

…Yet, businesses pretend to in a very expensive and misguided way.

They buy shit they don’t need or know how to use!

“Which model is best?”
“What tool should we buy?”
“What agents can we deploy?”

Each of these questions misses the point; they are reactive queries focused on technology.

And this is why your company isn’t winning with AI. The business model, org structure, and operational systems at most companies were never designed for AI. Yet executives are bolting AI onto these things without thinking strategically about it all.

*All execs need to hear is: “Copilot comes free with your Microsoft subscription!” and they think they’ve done AI*

And you better know I have an issue with that.

So while over the past two articles we’ve covered what a business model is and why it matters for data, now it is time for the explosive thought experiment—how does the business model adapt in this AI world?

The State of AI & Its Bolt-On Problem

I am by nature a realist and a pragmatist:

I’m very logical in my approach to questions and answers
I prioritize data foundations over fancy tools/ tech because they often don’t deliver the value promised
I’ve repeatedly talked about how companies aren’t ready for AI

But let’s give credit where it’s due—AI has gotten way too good to ignore!

What started as companies spending way too much money building internal chatbots is now an environment where AI workflows accelerate productivity by 10x, prototype new tech without oversight, and fully replace software services and companies. Hell, I’ve revolutionized how I run my entire freelance consulting business with an AI-native business model for less than $100 per month!

But here are the problems. Most of the adoption is tactical and ad hoc. Most of the focus is technical, not strategic. Most companies aren’t a one-person shop and can’t rework their business model as I can.

Essentially, AI is not being added to companies in a sustainable way; it’s a bolt-on.

Companies are treating AI as complementary to their operations. They implement use case by use case, tool by tool; a business team identifies a pain point, buys or builds an AI tool, deploys it within the existing workflow, and declares success. Then another team does the same thing. And another. And another.

*Whoever has to govern/ own this is going to have their hands full…*

What you end up with is a fragmented collection of AI tools that don’t talk to each other, aren’t governed consistently, and don’t reflect any coherent vision for how the business should operate.

Sound familiar? This is exactly what happened with data tools 10–15 years ago. Companies bought Tableau, Hadoop, and a dozen other tools without a strategy, and spent years cleaning up the mess. And honestly, plenty still haven’t.

The core issue with this new AI craze is this: companies are leading with technology rather than strategy. They’re asking “what can AI do?” instead of “how does AI change how we operate?” Those are fundamentally different questions, and the first one leads you straight into bolting AI onto your legacy business model.

There are three symptoms that tell you an organization is stuck in bolt-on mode.

Siloed AI Initiatives – No coordination when it comes to AI tooling. Marketing, sales, finance, operations all have bought into the AI Kool-Aid from their existing technology vendors. With this comes individual procurement decisions, workflows, and ways of defining/ measuring success. Worst of all, there is no shared infrastructure, governance, or visibility across the organization into what AI is actually doing. This mimics the siloed data problem that most companies are still dealing with…
No Change to Workflows or Roles – AI should make work easier and reduce friction for business processes across functions. But instead of redefining those processes, AI is just added to whatever exists today. Therefore, people may use AI to do the same thing slightly faster. Shouldn’t we rethink whether the task should exist at all? Or be done in a completely different way? Or change roles to use AI in more ways than just as a chatbot? AI used to speed up bad processes is still bad processes. Also, without workflow change, it isn’t scalable.
Governance as an Afterthought – This symptom is starting to rear its head. AI governance often gets tacked on after use rather than designed into the system from the start. As you know, I think companies should take a proactive and value-led approach to governance. Without AI governance, companies create inconsistencies in how they operate, introduce risk into their operations, and foster distrust (and anxiety) among non-technical stakeholders who don’t understand how these tools work.

If you work in an organization without any of these three symptoms, congratulations! But truthfully, I haven’t seen any.

Introducing AI Systems Design

Do I have the answer about how to solve this and make AI work perfectly in a huge, multinational corporation?

Well, I don’t have a silver bullet (because those never exist). But I do have an approach.

If I were to put a rough label on what’s needed, I’d call it AI Systems Design. This isn’t about a new technology, creating a context layer (the newest buzzword), or AI literacy training.

No, this is about redesigning your business model, strategy and processes to align with a new way of working. And when I refer to ‘systems’, I’m talking about the full sociotechnical system that is designed in a coherent and independent way.

To be clear, this goes beyond systemization within any one domain. It isn’t enterprise architecture. It isn’t data or AI strategy. And it isn’t AI solution development or implementation.
Similar to the Data Ecosystem, AI Systems Design is a more holistic approach for how a business should function in an AI-embedded world. It’s about refactoring how the business works, not bolting on cool solutions to speed up something that is becoming outdated.

There are two components to this.

The first is a Refactored Business Model & Strategy.

The second is three underpinning layers that make up the sociotechnical system executing on the Refactored Business Model & Strategy (which we will outline in detail in the next article).

Refactoring the Business Model & Organizational Strategy

If your company's business model looks the same today as it did ten years ago, that is a problem. If it continues to look the same in two years, then you may be out of business soon.

Most companies think of their business model, data capabilities, and AI initiatives as three separate things. There may be some references to each other in strategy materials, but in practice, they operate independently.

And this separation is the root of the problem.

I’ve helped multiple companies redesign their business models for the data-driven world. For example, a logistics provider that wanted to track data using their assets and sell insights back to customers. Or a media and events company that realized that to properly connect audiences, vendors, and other customers, they needed to be more data-driven, which changed how they approached all three.

Unfortunately, most companies don’t think like this, and it takes an outside perspective to get beyond the hump of “this is how things have always been done.” The positive part about this AI hype? Well, now these discussions are being forced to take place. And that is where the AI Systems Design framework comes in.

*This is a rough outline of how companies need to rethink how they operate in today’s AI-driven world*

There are three components of it. The foundational business model. The data enabler. And the Strategic AI Lever.

Here’s how to think about it.

The Foundational Business Model

You need to start here. You need to figure out what your business does, how it makes money, and its competitive advantage. In the end, these are the bedrock questions of why your company exists and how it survives. Unfortunately, we are seeing many companies start with an AI product and build from there. They won’t exist soon as the funding runs out.

And the key thing to remember is that this foundation doesn’t go away. Even the AI-forward companies like OpenAI, Claude, Google, etc., need revenue-generating subscribers to fund their operations. And for that, they need real customers. And to get those, they need business functions such as sales, marketing, and operations, etc.

No amount of AI changes that fact. AI can help with this, but it doesn’t replace it!

So to be clear, the business foundations come first (especially for an organization that has operated for years already). They always have and they always will.

AI as a Strategic Lever

Let’s skip data for now and go to the other side, where you have AI. AI’s role in this equation is to help refactor how you think about your foundational business model. It needs to be a question of how your operations, revenue-generating activities and competitive differentiation change by using AI? Right now, companies are mostly focused on AI productivity, efficiencies, or cool-sounding use cases. These are related to the refactoring components that AI should bring, but it’s not usually done in a symbiotic way.

In the refactored model, AI isn’t a side project. It’s a strategic lever that is woven into how the business creates, delivers, and captures value. I apologize for all the buzzwords here, but in practical terms, this means: “How does AI change our value proposition, our revenue mechanism, or our differentiator?”

For example, a subscription business might use AI to automate churn prediction and customer communications. But refactoring with AI might mean redesigning how it personalizes the customer journey, thereby changing the value proposition itself. Instead of the classic customer outreach process, an AI-enabled enterprise might build digital twin profiles for each customer, with its model learning each customer's communication preferences and targeting them at an optimal time period separate from others (i.e., going beyond segmentation). That is a lot of work, and sounds daunting, but AI has changed the game and should allow companies to do that.

And when AI changes the business model or processes to that degree, rather than just supporting it, that’s refactoring. That is where the next wave of disruption lies.

Data as the Enabler

And then there’s data. Sitting right in the middle, connecting both sides.

As we’ve covered before, data enables the foundations. Your data strategy, data model, technology decisions, and governance framework should all be shaped by the business model, processes and that line of thinking. Data makes the foundational business model more effective, measurable, and efficient.

That mindset should not disappear!

Because data will continue to enable the foundational business model. And it enables AI! We are seeing this a lot, but without well-governed, well-modelled data, AI doesn’t work. Data provides the path for AI to follow, and if it isn’t aligned with your business model, processes, and context, it will fail. The data layer is what connects “this is how our business works” to “this is how AI can enhance how our business works.”

Therefore, in the refactored business model, data serves as the active bridge between the two. It doesn’t just sit beneath the business, continuing to enable operations as before; it connects the foundational model to the AI-augmented model and enables the evolution between the two.

Companies aren’t thinking like this.
Executives aren’t thinking like this.
Even data leaders aren’t thinking like this.

But it is time we start thinking like this, because the technology has gotten to a point where we can, and should!

What “Refactored” Actually Means

So what does a refactored business model and strategy actually look like?

Essentially, it is a forward-looking business model & strategy, incorporating what exists and what has worked for years, evolving with the role AI should play in the future, all augmented and enabled by data.

In this instance, AI is not a bolt-on tool or technology; it integrates into and changes the business processes and workflows within how people deliver organizational value. And data cannot be an ad hoc, siloed exercise that delivers value on occasion; it needs to flow strongly into the AI-refactored business model.

More practically:

Your business fundamentals define what you do and how you make money
AI helps you enhance, automate, or completely redesign parts of those operational business processes
And data flows into that, both enabling the organizational understanding and value delivery, while improving its autonomous activities

Tackling this in an integrated way will be the next transformational thought experiment leaders need to consider. For example, when they are building their 2- or 5-year strategic plan, this needs to be front and centre.

But saying an organization needs to refactor their business model & strategy isn’t enough. You also need the “how.” That’s where the AI Sociotechnical Operational System comes in.

Next Week’s Introduction to the Sociotechnical Operating System

The second component of AI Systems Design is the operating system built underneath the refactored business model & strategy. I’ve called this the AI Sociotechnical Operating System.

This is a three-layer system:

Technology at the top
Process & People in the middle
Governance, Trust & Culture at the base

I’ve written about the People, Process, Technology & Data framework, but this is a build on that thinking to specifically address how these factors need to work together in an AI-driven world. Each component is designed to work cohesively and be framed with a view into the future, rather than what exists today. Moreover, as you may imagine, this can’t be built with standalone capabilities bolted together.

We’ll break down each layer in detail in the next article. But here’s the key idea: AI technology is only one component of it, and it’s probably the easiest part, given what is constantly being delivered to market. It is the process redesign and cultural trust that will determine whether companies succeed or fail. And right now, almost nobody is investing in those two layers.

Discussion about this post

Ready for more?