L2L Execution AI isn’t just for summarizing shifts—it’s a powerful tool for driving long-term Continuous Improvement (CI). By using specific prompts, you can transform a mountain of dispatch data into a structured CI Opportunity Register. This guide shows you how to define your inputs, structure your outputs, and save your prompts to create a repeatable improvement process.
Defining the Input
To get high-quality insights, you must give the AI clear boundaries. In your prompt, use plain English to define exactly what data the AI should "look" at:
The Window: Specify the time frame (e.g., "the last 30 days") or the volume (e.g., "the last 100 dispatches").
The Location: Identify the specific Plant, Line, or Machine Group.
The Context: Briefly state what "success" looks like for this area (e.g., "Our goal is to reduce MTTR on the Packaging Line").
Formatting the Output
Don't just ask the AI to "find problems." Tell it exactly how you want that information presented. For a Kaizen-ready report, ask the AI to:
Build a Table: Request a "Ranked CI Opportunity Register" in table format.
Identify Drivers: Ask for the "Top 5 Systemic Drivers of Downtime."
Prioritize: Instruct the AI to rank opportunities based on frequency and total duration of downtime.
Sustainability
The most effective prompts are the ones that get used every day.
Save as Custom Prompts: Once you’ve dialed in a prompt that generates a great CI Register, save it within L2L as a Custom Prompt.
Standardize: Share these prompts with other teams or sites. This ensures that every shift lead and CI engineer is looking at data through the same lens, creating a standardized "Golden Prompt" for the entire organization.
Actioning on the Data
By targeting a specific set of records—such as the last 100 dispatches—and asking for a prioritized list of issues, you move from "reacting to dispatches" to "solving systemic problems." Execution AI does the heavy lifting of data analysis in seconds, leaving your team with more time to focus on the actual Kaizen work.
Validate the Opportunity Register: Review the top findings with your team to "sanity-check" the AI's data and select one or two high-impact opportunities to pursue.
Launch with Clear Ownership: Use the template to spin up a formal project, assigning an owner, defining the scope, and setting target metrics. This can easily be done using L2L's Kaizen module.
Using the Kaizen module
Convert Opportunity to Kaizen: Seamlessly bridge the gap between AI analysis and action by opening the L2L Kaizen module to create a targeted event.
Pre-fill the Scope: Use the direct recommendations from your Execution AI output to populate the Kaizen scope, ensuring alignment with data-driven insights.
Target Dispatches: Tag the specific dispatch reasons the Kaizen targets to maintain a digital thread between the improvement and the original problem.
Build the Kaizen Workspace: Establish a centralized hub to run the Plan-Do-Check-Act (PDCA) cycle with your frontline team.
Drive Accountability: Assign specific tasks, owners, and due dates within the module to ensure progress doesn't stall.
Document Evidence: Attach photos, logs, and real-time updates directly to the workspace for full visibility.
Track the Result: Quantify your success by comparing performance metrics from before and after the Kaizen implementation.
Monitor Metrics: Measure the direct impact on disruption frequency and total downtime minutes.
Calculate ROI: Convert "hours saved" into "dollars saved" to provide leadership with a clear picture of the project's value.