“If the only place your process lives is in someone’s head, it’s invisible to Artificial Intelligence (AI). And unscalable to everyone else.”
The sentiment behind that line comes up all too often in client conversations. There is a growing appetite to “do something with AI,” but very few organisations are genuinely ready.
In my experience, the real gains come when AI is paired with something most businesses still struggle with: clearly documented, well-owned processes, supported by reliable, structured data.
The temptation is to buy a tool and expect instant value. AI might impress with its insights, but without understanding how your business actually runs, it cannot improve efficiency, scale reliably and effectively reduce risks. That means understanding both your data and your processes is essential if you want AI to do more than generate interesting insights.
Why AI Struggles Without a Map
Documenting process is often seen as admin, something done for audits or training manuals. But at its core, it creates shared understanding: who does what, when, with what input, and what happens next. It sets the foundation for measuring improvements, highlights bottlenecks, and helps ensure systems support how the business actually works.
Data plays a similar role. It captures what has happened and gives signals on what might. But without a structured process context, data can mislead. A dashboard might show slow turnaround times, but without understanding how the work flows, where it hands off and who owns it, you risk drawing the wrong conclusions and fixing the wrong thing.
Process defines how work is done; data provides the visibility to measure, govern and improve it. Together, they form the foundation AI needs to understand the context, draw the right conclusions and make decisions or recommendations that deliver real value. That said, meaningful impact still depends on people who trust it, systems that support it and a culture willing to change.
If you’re serious about using AI to drive real change and not just analyse the past, you need both: reliable data to learn from and clear processes to act within. Anything else, and you risk leaving potential untapped, drawing the wrong conclusions and solving the wrong problems.
Why Systems Reflect the Chaos
I often hear the same thing:
“The system doesn’t quite work, so we just do this bit manually.”
Usually, the problem is not the system. It is the process, the system was built around. When that process is undocumented or unclear, the system mirrors the mess. Short-term fixes become permanent features. One team’s workaround becomes another’s best practice, handed down like a bad habit.
The result is complexity, friction and inconsistent data. If the process is messy, so is the data it generates. And any AI trying to use it will be working with a partial picture.
Shadow IT: A Process and Data Problem in Disguise
In one organisation, I found three separate scheduling apps. Each was built by a different region to solve the same issue. None were integrated. All were paid for. Each had slightly different logic and outputs.
This kind of duplication happens when teams are left to solve their own pain points without a shared understanding of the process or shared ownership of the data. It feels like agility, but usually means:
- The process is not defined
- The data is fragmented
When process is clear and data is structured, technology supports it. When they are not, tech becomes a patch, not a solution. And beneath all of this sits the thorny issue of governance: complex, critical and best explored in its own right.
Where AI Fits in
AI will still uncover patterns and opportunities in your data. It can flag issues, spot trends, summarise content and recommend improvements, even in messy environments. But if you want it to improve how work actually gets done, not just report on it, process becomes critical. AI cannot optimise what it cannot see, and it cannot effectively learn from poor-quality or inconsistent data.
When your processes are documented, your data flows cleanly and ownership is clear, AI becomes more than a dashboard. It becomes a tool that supports delivery and scale. You still get value from data alone. But when it is tied to process, the benefit multiplies.
This is Not About Slowing Down Innovation
This is not about red tape or writing dusty SOPs. It is about moving faster with more confidence. Clear processes and reliable data help teams test, improve and scale change without constantly starting from scratch. They also ensure systems and AI tools are built around how the business should work, not just what was quickest to configure.
If you are introducing AI or upgrading a key system, good data and clear process are not bureaucracy. They’re how you make sure what comes next actually works.
Where to Start
You don’t need a full transformation programme to get going. The best place to start is often where things already feel clunky: where the process is unclear, the data unreliable, or the handovers messy.
Start by:
- Mapping what is really happening
- Identifying the key steps, decisions and handovers
- Reviewing the data being captured and how it’s being used
- Clarifying who owns the process and who owns the data
This doesn’t instantly improve the process, but it exposes where the real problems are, builds shared understanding, and gives you a clear, grounded starting point for improvement, smarter automation and meaningful use of AI.
Final Thought
AI might be the shiny new tool, but it only delivers real results when it’s built on something solid. Documenting how work actually happens, and making sure the data behind it is clean, structured and owned, is how you turn hype into impact. If you’re serious about using AI to do more than generate insight, start by giving it something it can work with. Because when process and data are in shape, AI won’t just support the business, it will drive it forward.
By Matt Lord
Matt Lord is a Skewb Partner. To speak with him about this topic in more detail, you can connect with him on LinkedIn.
