About Lean Path
We build AI demand forecasting for Tier-2 automotive supply chains in Japan.
Our focus
Lean Path was founded on a specific observation: the demand signal problem in automotive Tier-2 supply chains in Japan is not a data availability problem. Suppliers have ERP data, shipment records, and in many cases, electronic connections to their customers' systems. What they lack is a forecasting layer calibrated to how pull actually flows through a multi-tier automotive supply chain.
Most general-purpose forecasting tools are built for retail replenishment or B2B distribution. They apply smoothing and trend analysis to shipment history. In automotive supply chains, that approach misses the planning layer that actually drives orders: the assembly plant's daily production schedule and the supplier's own scheduling rules for each SKU family.
We built Lean Path to address that specific gap. The platform ingests both the customer's pull schedule and the supplier's ERP data, calibrates a forecast model to the dynamics of that specific supply chain relationship, and delivers weekly forecasts in a format the planning team can actually use.
How we started
Takeshi Yoshida spent several years working in supply chain operations roles at Tier-1 and Tier-2 automotive suppliers in the Chubu region before concluding that the forecasting problem was not being solved by the existing software options available to mid-size suppliers. The tools were either too expensive and complex for organizations without dedicated data engineering capacity, or too shallow to model the actual demand dynamics of a multi-tier automotive supply chain.
He connected with Kenji Tanaka, whose background was in applying machine learning to operations problems in manufacturing and logistics. Together they spent the first year doing the groundwork: mapping how demand actually propagates through automotive supply chain tiers, understanding why general-purpose forecasting tools fail on this specific problem, and running a series of manual pilots at design-partner plants to validate whether a signal-based approach could outperform the baseline.
Lean Path K.K. was incorporated in Tokyo in 2025. The company is backed by angel investors with backgrounds in industrial software and Japanese manufacturing operations. We are not pursuing venture funding at this stage. Our focus is on delivering measurable forecast accuracy improvements for a small number of design-partner customers before expanding the program.
The team
Takeshi Yoshida
CEO and Co-founder
Background in supply chain operations at Tier-1 and Tier-2 automotive manufacturers in the Chubu region. Focused on the demand signal gap between assembly plants and their supply base. Founded Lean Path to build the forecasting layer that the existing software market had not addressed for mid-size automotive suppliers in Japan.
Kenji Tanaka
CTO and Co-founder
Background in applied machine learning for operations problems in manufacturing and logistics. Built the initial forecasting models that became the basis for Lean Path's core platform. Responsible for the technical architecture, data pipeline design, and the calibration methodology that distinguishes steady-state pull from demand variability caused by model changeovers and seasonal factors.
Rika Hayashi
Head of Operations
Builds practical systems for kanban flow and cross-tier coordination between assembly plants and parts suppliers. Leads the design-partner engagement process from initial data connection through the structured pilot, weekly forecast delivery, and accuracy review. Manages the onboarding workflow that allows a small team to support multiple active design-partner sites simultaneously.
How we work
Small cohort, close support
We work with a deliberately small number of design-partner sites at any given time. That constraint exists because we think hands-on engagement during the pilot phase is where the real forecast calibration work happens. We do not try to onboard more sites than we can support well.
Measured accuracy, not promises
Every pilot starts with a defined accuracy baseline derived from the site's existing planning approach. At the end of the pilot, we compare Lean Path's forecast accuracy against that baseline across the SKU families in scope. We report the actual numbers, not a general claim about improvement.
Built for the planner, not the analyst
The planning teams we work with do not have dedicated data science staff. Lean Path delivers forecasts in formats the planning team already uses, with explanations of what is driving the forecast for each SKU family. The goal is a tool the planning team can act on directly, not a model output that requires interpretation.
Work with us
If your team is managing demand forecasting for automotive supply chains in Japan and you want to understand whether Lean Path could be a fit, we would like to hear from you.