Lean Six Sigma and Automation: What to Fix Before You Automate
Automation has become one of the most compelling promises in operational improvement. Robotic Process Automation (RPA), workflow automation platforms, and AI-assisted processing all offer the possibility of dramatic productivity gains -- tasks completed in seconds that previously took hours, with no manual errors and no end-of-shift fatigue. The investment case is compelling, the vendor demonstrations are persuasive, and the executive enthusiasm is real.
And yet a significant proportion of automation initiatives fail to deliver their projected benefits. The reasons vary -- poor change management, inadequate testing, underestimated complexity -- but there is one root cause that is both extremely common and almost entirely preventable: organisations automate broken processes. They take a workflow that is inefficient, inconsistent, and riddled with workarounds and they make it run faster. The result is that errors arrive more quickly, inconsistencies are embedded at scale, and the problems that were manageable at human speed become systemic at machine speed.
Lean Six Sigma offers the disciplined framework for avoiding this trap. Its core insight -- that waste and variation must be eliminated before optimisation is applied -- translates directly into a sequencing principle for automation: fix the process first, then automate the fixed process.
The Lean Principle: Simplify and Standardise First
Lean thinking identifies seven categories of waste -- overproduction, waiting, transportation, over-processing, inventory, motion, and defects -- and its discipline is the systematic identification and elimination of these wastes before any attempt is made to speed up the process. The logic is straightforward: automating a step that should not exist at all is not an improvement, it is an investment in the wrong outcome.
Standardisation is the companion to waste elimination. A process that varies depending on who is doing it, which system version is installed, or what day of the week it is cannot be reliably automated. Automation depends on consistency -- the same inputs produce the same outputs through the same steps. Without standardisation, the automation must accommodate variation through exception-handling logic, which adds complexity, fragility, and maintenance overhead that erodes the efficiency gains the automation was supposed to produce.
The practical implication is that process mapping and simplification are prerequisites for successful automation -- not parallel workstreams, not follow-up activities, but prerequisites. The time spent eliminating unnecessary steps and standardising the remaining ones pays back directly in automation that is simpler to build, more reliable in operation, and easier to maintain.
The Sequence: Eliminate, Standardise, Automate
The correct sequencing of a Lean-informed automation programme follows three stages. The first stage is waste elimination: map the current state of the process, identify every step that does not add value from the customer's perspective, and redesign the process to remove those steps. This is the application of Lean's Value Stream Mapping discipline. The output is a leaner process -- fewer steps, shorter cycle time, reduced handoffs -- that is already more efficient than the original before any technology is involved.
The second stage is standardisation: document the redesigned process in enough detail that it can be performed consistently by anyone following the documentation. This includes defining the inputs the process requires, the decisions it involves and the rules governing those decisions, the outputs it produces, and the exception conditions it must handle. Six Sigma's process documentation tools -- particularly the SIPOC (Suppliers, Inputs, Process, Outputs, Customers) framework and detailed process procedures -- are directly applicable here.
Only at the third stage does automation enter the picture. With a lean, documented, standardised process in hand, the automation design is dramatically simpler: the process steps are clear, the decision rules are explicit, the exception conditions are defined, and the expected inputs and outputs are specified. The automation solution that emerges from this foundation is more reliable, faster to build, and easier to test than one applied to an undocumented, variable process.
What to Automate and What to Leave Alone
Not every process is a good candidate for automation, and Lean Six Sigma provides a useful lens for identifying which ones are. High-volume, repetitive, rule-based tasks with stable, well-defined inputs are the strongest candidates. Data entry, format conversion, report generation, invoice matching, compliance checking against fixed rules -- these are tasks where automation delivers the highest returns and where the risks are most manageable.
Judgment-intensive tasks are poor candidates for current automation technology. Processes that require contextual interpretation, stakeholder relationship management, creative problem-solving, or the exercise of professional discretion in novel situations involve forms of reasoning that automation cannot yet replicate reliably. Attempting to automate these processes produces systems that either require extensive human oversight -- defeating the purpose -- or make poor decisions at scale.
Low-volume and highly variable tasks present a different challenge: the return on automation investment is simply too low to justify the build and maintenance costs. A task performed twenty times per month with significant variation in each instance is unlikely to produce a positive business case for automation regardless of how well the process is designed.
LSS as the Foundation for Successful RPA
Organisations that have applied Lean Six Sigma disciplines before launching RPA programmes consistently report better outcomes than those that automate without this foundation. The reasons are predictable: the processes selected for automation are already clean and well-documented; the bots built on these processes are simpler and more stable; the exception rates are lower because the process variability has been reduced; and the maintenance burden is manageable because the underlying process logic is well-understood.
Equally important, the Lean Six Sigma discipline helps organisations avoid the sunk-cost trap of having invested in automating a bad process. The Value Stream Mapping exercise will often reveal that the process being considered for automation is actually a workaround for a system limitation or a policy that should simply be changed. Fixing the underlying problem eliminates both the automation need and the process waste simultaneously -- a far better outcome than automating the workaround and maintaining the bot indefinitely.
Automation is a powerful tool. Like any powerful tool, its value depends entirely on how well it is applied. Applied to a broken process, it accelerates failure. Applied to a lean, standardised, well-understood process, it delivers the efficiency gains its proponents promise.
XNM Consulting helps organisations apply Lean Six Sigma disciplines to process design and improvement -- including preparing processes for successful automation. Learn more about our strategic advisory services.