Insight • Marc Schmitt

Lean First, AI Second: Why You Shouldn’t Rush Into AI-Driven Lean Management

Free expert overview by Marc Schmitt • Premium deep dive available after login

Free expert overview by Marc Schmitt

Why Lean Comes Before AI in Process Improvement

Many organizations are excited to use artificial intelligence (AI) to improve their operations. However, AI should not be the first step. Instead, it works best when applied to processes that are already stable, simple, and well-understood. This is where lean management comes in.

What Is Lean Management?

Lean management is a method focused on removing waste, improving flow, and making work predictable. It helps organizations fix broken processes by simplifying steps, reducing delays, and standardizing how work is done. This creates a strong foundation that AI can build upon.

Why Not Start With AI?

Applying AI to a messy or chaotic process can make problems worse. AI tends to amplify whatever it is given. If the process is unstable or the data is poor, AI may produce errors, cause rework, or create confusion. Lean fixes these issues first, so AI can truly add value.

Building a Lean Foundation

Step 1: Understand and Map the Process

Begin by clearly defining the problem and the boundaries of the process. Observe how work is done in reality, not just how it should be done. Use tools like process maps to visualize every step, handoff, and decision point.

Step 2: Measure Baseline Performance

Collect data on key metrics such as cycle time, errors, and rework. This baseline helps track improvements and shows where waste exists.

Step 3: Remove Waste and Simplify

Eliminate unnecessary steps, reduce handoffs, and error-proof inputs. Simplify decision rules so they are clear and explicit.

Step 4: Standardize the Process

Create clear instructions, roles, and training materials. Standard work ensures everyone follows the same best practice, making the process predictable and stable.

Ensuring Data Quality

Good data is essential for AI. Define data fields clearly, standardize how data is entered, and validate inputs to catch errors early. Replace free-text entries with structured fields where possible. Monitor data quality continuously to maintain trust in AI outputs.

Choosing the Right Process for AI

Not every process is ready for AI. Focus on high-volume or high-pain processes with measurable outputs and clear ownership. Avoid one-off or highly variable processes until they are simplified and stable.

Introducing AI Carefully

Only after the process and data are ready should AI be considered. Classify tasks into three groups: fix the process first, automate clear rules second, and apply AI for complex tasks last. Always keep humans in the loop to review and override AI decisions when needed.

Pilot and Scale

Start with small pilots that have clear success metrics and fallback manual options. Monitor performance closely and scale only after proven value. Update standard work and governance to maintain control and continuous improvement.

Summary

AI is a powerful tool, but it works best on a solid foundation. By applying lean principles first to fix and standardize processes and ensure data quality, organizations can harness AI to improve efficiency, accuracy, and decision-making sustainably.

Key steps

Establish a Lean Process Foundation
Begin by thoroughly mapping, simplifying, and standardizing your current processes. Use direct observation to capture how work truly happens, identify waste, and measure baseline performance. This stable foundation ensures that processes are predictable and repeatable, which is essential before introducing AI. Without this, AI risks amplifying existing inefficiencies rather than improving outcomes.
Ensure Data Readiness and Quality
Focus on defining and standardizing data fields, validating inputs, and creating feedback loops to maintain data integrity. Reliable, consistent data is critical for AI to function correctly and avoid errors. Establish a data dictionary and monitor data quality continuously to build trust in AI-driven decisions.
Build a Cross-Functional Team with Clear Roles
Assemble a small, focused team including a process owner, frontline experts, lean facilitator, data specialist, and AI/automation expert. Clear role definitions and collaboration ensure that process improvements and AI integration are managed effectively, grounded in operational realities and technical feasibility.
Select Suitable Processes for AI Integration
Choose processes that are high volume or have significant pain points, span multiple teams, and have measurable outputs. Avoid one-off or highly unique processes without clear ownership. This selection ensures that AI efforts target areas with the greatest potential impact and measurable results.
Use Readiness Checklists Before Applying AI
Apply process and data readiness checklists as gates to confirm stability, waste reduction, control mechanisms, and data quality. Only proceed with AI implementation if these criteria are met, ensuring AI acts as a force multiplier rather than amplifying chaos.
Pilot AI with Human Oversight and Scale Gradually
Conduct controlled pilots with fallback manual processes and clear success metrics. Monitor performance closely, including error rates and adoption. Scale AI only after proven value, updating standard work and establishing governance for ongoing monitoring and continuous improvement.

FAQ

Why is it essential to establish a lean process foundation before applying AI in process management?

A lean process foundation ensures processes are stable, simplified, and standardized, removing waste and variability. Without this, AI amplifies chaos and errors. Lean creates a reliable baseline and improves data quality, enabling AI to act as a force multiplier rather than magnifying inefficiencies.

What roles are critical in the cross-functional team for integrating lean and AI, and why?

Key roles include the process owner (accountability), frontline SMEs (practical insights), lean facilitator (guides improvements), data/systems expert (ensures data quality), and AI specialist (implements AI). This team ensures alignment between operations and technology for successful integration.

How should organizations select processes suitable for AI integration?

Select high-volume or high-pain processes with measurable outputs and clear ownership. Avoid one-off, unclear, or highly unique processes initially. Focus on stable, cross-team processes where AI can deliver measurable impact and improvements.

What is the purpose of the Process Readiness (Gate A) and Data Readiness (Gate B) checklists before applying AI?

These checklists confirm that the process is stable, standardized, and waste-reduced (Gate A), and that data is consistent, validated, and traceable (Gate B). Passing both gates ensures AI is applied on a solid foundation, reducing risk of failure.

How should AI pilots be designed and managed to ensure safe and effective implementation?

Design pilots with narrow scope, clear success metrics, and fallback manual processes. Include human-in-the-loop review to manage risk. Monitor accuracy, errors, adoption, and trust. Scale only after proven value and establish governance for ongoing control.

Marc Schmitt