What Still Matters: The Problems Lean Six Sigma Was Built to Solve Have Not Gone Away

The case for Lean Six Sigma was never primarily about the tools. It was about the underlying operational problems that the tools were designed to address. And those problems have not changed with the arrival of AI. If anything, several of them have become more visible and more urgent.

Processes are still poorly understood

In most organisations, there is a persistent gap between how a process is documented and how it actually runs. Official procedures describe the intended path. Operational reality includes the workarounds, the informal shortcuts, the exceptions that accumulate over years of adaptation to changing demands. AI makes this gap more visible sooner — process mining tools, for example, can surface variation and deviation patterns in system data almost immediately. But making the gap visible is not the same as closing it. Understanding why the gap exists, which parts of it represent genuine waste versus necessary adaptation, and what to do about it requires the kind of disciplined, context-sensitive analysis that Lean Six Sigma provides.

Root cause analysis is still necessary

AI tools are increasingly capable of identifying that a problem exists and where in a process it tends to cluster. They are significantly less capable of determining why it exists. Root cause analysis — the structured discipline of separating symptoms from causes, testing hypotheses, and avoiding the trap of solving the wrong problem — remains a distinctly human responsibility. DMAIC and similar frameworks provide the thinking structure that prevents teams from jumping to solutions before they understand the problem. That structure does not become less important when pattern recognition is faster. It becomes more important, because the temptation to act on the first plausible explanation is stronger when data arrives quickly.

Improvement initiatives still struggle to scale

One of the consistent failure modes in process improvement is the initiative that produces good results in a pilot or a single team and then fails to propagate across the organisation. This is not a data problem. It is a people, governance, and change management problem. AI does not solve it. The disciplines of stakeholder engagement, structured learning cycles such as PDCA, and sustained improvement ownership remain as relevant as they were before AI tools existed.

Automation is still applied to poorly understood processes

The same principle that has always applied to automation applies to AI-powered automation: applying it to a process that is not clearly understood does not fix the process. It executes the confusion faster and at greater scale. Lean Six Sigma's insistence on understanding the current state before designing the future state is not a relic of a pre-AI world. It is a prerequisite for getting value from AI investment.

What Needs to Evolve: The Work of Applying the Methodology

What changes with AI is not what Lean Six Sigma is trying to do. What changes is the effort required to do it.

Several of the most time-consuming activities in a traditional Lean Six Sigma project — current-state process mapping, stakeholder interviews, data collection, baseline analysis, and project scoping — can now be supported significantly by AI tools. This does not mean they disappear. It means they take less time, require less coordination, and produce output faster.

Process discovery and current-state documentation

Building an accurate picture of how a process currently runs has traditionally required weeks of interviews, observation, and workshop facilitation before a single improvement could be designed. AI process mapping tools can compress this dramatically. Platforms like Famla capture structured process knowledge asynchronously from the people doing the work, process existing documentation, and generate current-state maps and Lean analysis automatically. The first-draft work that once consumed a large part of the Define and Measure phases of DMAIC can now be available within days.

Pattern identification and issue surfacing

Identifying recurring patterns of waste, delay, rework, and variation across a value stream used to require significant manual analysis. AI can surface these patterns from process data and structured input consistently and at scale. The practitioner's role shifts from performing this analysis to validating it, contextualising it, and determining which patterns are worth acting on given the organisation's priorities and constraints.

The shift in where time is spent

As a result of these changes, the constraint in a Lean Six Sigma project is shifting. It used to be access to data and analysis. Increasingly, the constraint is deciding what to act on, aligning the stakeholders needed to make change happen, and following through on implementation. AI accelerates the front end of the improvement cycle. The back end — decision-making, change leadership, and sustaining improvement — remains as demanding as it always was.

How AI Affects Each Phase of DMAIC

What Changes and What Does Not: A Clear Summary

An Evolving Practitioner Role

The shift in how Lean Six Sigma is practised has direct implications for what a Green Belt, Black Belt, or Operational Excellence leader is expected to contribute.

In a traditional project, a significant portion of a practitioner's time was spent organising interviews, facilitating workshops, consolidating inputs, and producing documentation. That work was necessary and valuable, but it was also the part of the role that was most resource-intensive and least differentiated. Anyone with enough time and patience could produce a process map. What required real expertise was deciding what the map revealed, how to interpret root cause evidence, and how to guide a team toward a solution that would actually stick.

AI reduces the time and effort required for the former. This concentrates the practitioner's contribution in the areas where expertise genuinely matters: interpreting findings in context, challenging surface-level explanations, asking the question that reframes the problem, and maintaining the discipline of structured improvement cycles when organisational pressure pushes toward quick fixes.

Practitioners who adapt to this shift will find that their expertise is more valuable, not less. The ability to guide a team through genuine problem-solving — as distinct from producing process artefacts — is exactly what AI cannot do, and exactly what organisations need when AI surfaces more improvement opportunities than they have capacity to address well.

Practitioners who do not adapt risk becoming process documentation facilitators in a world where process documentation is increasingly automated. The role is not disappearing. It is clarifying: less about producing analysis, more about leading the decisions that follow from it.