A Lean 5.0 Decision-Support System Toward Sustainable Continuous Improvement: Integrating Attribution Analytics, Yokoten and Learning from Evidence

Lean 5.0 extends continuous improvement beyond efficiency toward human-centricity, sustainability, and resilience, but company performance systems rarely measure these dimensions. This study proposes a Lean 5.0 Decision-Support System (L5.0-DSS) that starts from the Key Performance Indicators (KPIs) a company already has. It ranks activities by their below-target KPI gaps without weights, traces each loss to the activity where it originates through Generative AI text coding and expert validation, uses Yokoten to transfer internal practices, and interprets the results along the three Lean 5.0 dimensions. In a temperature-controlled fresh-food warehouse, nine projects were implemented; service failures were observed to fall by up to one percentage point, and productivity to rise by 53 and 71 cases per labour hour in Consolidation Staging and Picking. For human-centricity, practitioners reported less manual handling and less dependence on experienced pickers, with Friday peaks covered by cross-trained staff instead of overtime. For sustainability, less product damage suggests lower food waste and rework. For resilience, cross-training and synchronized inbound flows made demand peaks easier to absorb. These effects were not measured directly; the case therefore identified new KPIs, such as the damage rate and the overtime ratio, that existing records allow in the next cycle.