Learning objective: By the end of this article, you will understand why organisations should improve their processes before using AI or automation, and how process clarity, ownership, measurement, and continuous improvement help technology create real operational value.
AI creates value when the process is ready
AI is now part of almost every serious business conversation. Leaders are asking how they can use it to work faster, reduce cost, support teams, improve decisions, and serve customers better. The opportunity is real. AI can help organisations analyse information, automate routine work, identify patterns, and support better decision-making. But there is an important step that often gets missed: before organisations automate a process, they should first make sure the process works properly. The same applies when they digitise a workflow or introduce AI into daily operations. If the work is unclear, inconsistent, or poorly managed, technology will only make the problem move faster.
This is where Operational Excellence becomes especially important. AI can support better operations, but it does not replace the need for good operations. A weak process does not become strong just because it has been automated. A confusing approval flow does not become clear just because a digital tool now moves it from one person to another. Poor data does not become reliable just because it is shown on a dashboard. Good technology needs good processes underneath it.
Improve the process before you automate it
A useful way to think about this is simple: improve the process before you automate it. This does not mean every process has to be perfect before technology is introduced. That would slow progress too much. It means organisations should understand the process well enough to know what they are improving, what they are automating, and what outcome they expect to achieve.
This distinction matters because many organisations are investing heavily in AI but still struggling to turn that investment into operational value. McKinsey’s 2025 workplace AI report puts it clearly: “Almost all companies invest in AI, but just 1 percent believe they are at maturity.” The same report says that while 92 percent of companies planned to increase AI investment over the next three years, only 1 percent of leaders described their companies as mature, meaning AI was fully integrated into workflows and producing substantial business outcomes.
That gap is important. It shows that buying tools, launching pilots, and talking about AI transformation is not the same as building the capability to use AI well. Many organisations are not short of technology. They are short of process discipline, ownership, measurement, and practical routines that help improvements stick.
What AI-ready Operational Excellence means
AI-ready Operational Excellence means the organisation has prepared its operations well enough for AI and automation to help. The processes are not necessarily perfect, but they are visible. People understand how the work flows. Roles are clear. The organisation knows where delays happen, where errors are created, what data is available, and what decisions need support. In this kind of environment, AI has something useful to build on.
Start with process clarity
Process clarity comes first. Before automating a process, teams should understand how the process starts, who is involved, what decisions are made, what information is needed, what exceptions occur, and how the process ends. This sounds basic, but many operational problems live in these details. One team follows the process one way. Another team follows it differently. Some approvals are formal, while others happen through messages or informal conversations. Some data is recorded, while some remains in people’s heads.
When this kind of process is automated too early, the organisation often locks in the confusion. The tool may make the process look more modern, but the underlying work is still unclear. People still disagree about what should happen. Exceptions still take too long. Handoffs still break down. The only difference is that the broken process now moves through a system.
Ownership makes improvement accountable
This is why process ownership matters. Every important process should have someone accountable for how it performs. That person does not need to do every task in the process, but they should understand the process, monitor it, support improvement, and help decide when automation makes sense. Without ownership, process problems become everyone’s frustration and nobody’s responsibility.
Measurement proves whether improvement worked
Measurement is just as important. If an organisation does not measure process performance, it becomes difficult to prove whether AI or automation has improved anything. Did the cycle time reduce? Did errors decrease? Did teams spend less time chasing information? Did customer response time improve? Did the process become easier for employees to follow? These questions need clear measures. Without them, technology success becomes a matter of opinion.
Better data starts with better processes
Data quality also depends on process quality. AI needs useful data, and useful data usually comes from work that is performed and recorded consistently. If teams use different definitions, skip steps, enter incomplete information, or record decisions in different places, the data becomes harder to trust. AI may still produce outputs, but those outputs will reflect the quality of the information underneath.
This does not mean organisations should delay AI until everything is perfectly standardised. That is not realistic. It means they should standardise the parts of the process that matter most. Core steps, decision rules, handoffs, data points, and responsibilities need enough consistency for technology to support the work properly. Flexibility can still exist, especially where human judgement is needed, but the process should not depend on guesswork.
Improve before you digitise
Continuous improvement should also come before automation. Before digitising a workflow, teams should look for obvious waste. Are there duplicate approvals? Are people entering the same information more than once? Are requests waiting because nobody knows who should act next? Are teams asking for information that is no longer needed? Are exceptions handled manually because rules are unclear? Removing these issues first makes automation more valuable.
Process intelligence still needs human judgement
This is also where process intelligence can help. Process mining, workflow analytics, and operational dashboards can show how work actually moves through the organisation. They can reveal bottlenecks, variation, rework, and delays. But these tools still need human judgement. A dashboard can show that a process is slow. It cannot always explain why the process was designed that way, which controls are necessary, or what trade-offs the organisation should make.
McKinsey’s 2025 State of AI survey also reinforces this point. It found that most organisations were still experimenting or piloting AI, while nearly two-thirds had not yet begun scaling AI across the enterprise. The same survey found that workflow redesign is one of the strongest factors linked to meaningful business impact from AI, and that high-performing organisations are much more likely to redesign workflows rather than simply add AI on top of existing work.
That is the real lesson for operations leaders. AI works best when it changes how work gets done, not when it simply sits beside the existing process. If the existing process is slow, unclear, or poorly owned, AI may create some local improvements, but it will struggle to create enterprise-level value. If the process is understood and improved first, AI can help scale a better way of working.
People understand the real work
The human side matters the most. People need to understand why AI is being introduced, how it will help them, and what will change in their daily work. They also need clear instructions, training, and confidence. When people are involved in improving the process before automation, they are more likely to trust the solution. They can point out real problems, explain practical exceptions, and help design a workflow that fits the work.
This is especially important because many process problems are not visible from a management report. They are visible to the people doing the work every day. They know which fields are confusing, which approvals create delays, which reports nobody uses, and which workarounds keep the process running. A good Operational Excellence approach brings that knowledge into the improvement process before technology is applied.
A practical example: approval workflows
A simple example makes this clearer. Imagine an organisation wants to automate an internal approval process. The goal is to reduce delays and make approvals easier to track. That sounds sensible. But before automation starts, the organisation should clarify the process. What types of requests exist? Who can approve each type? What information is required? Which requests need escalation? What happens when information is missing? What is the expected turnaround time? Who owns the process performance?
Once those questions are answered, automation becomes much more useful. The system can route requests correctly. It can ask for the right information at the start. It can remind the right person at the right time. It can show where delays happen. It can create better data for future improvement. The technology is now supporting a better process, not just digitising a poor one.
A simple AI-readiness checklist
A practical AI-readiness check for operations can be built around seven questions:
Is the process clearly defined?
Does the process have an accountable owner?
Are the key steps, handoffs, and decisions understood?
Are performance measures in place?
Is the process data reliable enough to support decisions?
Have obvious delays, duplication, and waste been removed?
Are the people involved prepared to adopt the new way of working?
These questions are simple, but they are powerful. They help organisations avoid rushing into automation for the sake of appearing modern. They also help leaders focus on value. The goal is not to “use AI” as an activity. The goal is to improve performance, reduce friction, support people, and create better operational outcomes.
AI increases the value of process work
BCG’s 2025 research on AI value makes a similar point from a performance perspective. It found that future-built companies are pulling ahead because they are not only investing in AI, but also building stronger people and technology capabilities, tracking value, and reinvesting gains. BCG also reported that future-built companies expected twice the revenue increase and 40 percent greater cost reductions by 2028 than lagging companies in the areas where they apply AI.
For Operational Excellence and process improvement professionals, this is a major opportunity. AI does not reduce the importance of process work. It increases it. Organisations need people who can understand operations, improve workflows, define ownership, measure performance, and help technology fit the real work. The more AI becomes part of operations, the more valuable strong process thinking becomes.
Operational Excellence is the foundation
Operational Excellence gives organisations the structure to do this properly. It helps them understand how work is done, where performance is weak, what should be improved, and how improvements should be sustained. It also helps avoid the common pattern where a new tool is introduced, excitement rises for a short period, and then teams slowly return to old habits because the management system around the change was never strong enough.
This is also where a structured Operational Excellence Management System can help. PATH OEMS™ is built around the idea that organisations need a practical system for knowing where to start, what to improve, how to align people, and how to hold the gains. That kind of structure becomes even more useful when organisations are preparing for AI or automation. It keeps the focus on the work, not just the technology.
Final takeaway: improve first, automate second
The message is not that organisations should move slowly. The message is that they should move intelligently. AI and automation can create real value when they are applied to processes that are clear, owned, measured, and continuously improved. When the work is strong, technology can make it stronger. When the work is unclear, technology often exposes the weakness.
So before automating a process, improve it. Before digitising a workflow, understand it. Before applying AI to operations, make sure the process has enough discipline to benefit from it. That is the foundation of AI-ready Operational Excellence — and it is one of the most practical ways organisations can turn technology investment into real operational value.
Leave a Reply