A manual, error-prone process standing in the way of a Lean culture

Value Driving is a core part of Nippon Paint’s Lean for Growth (LFG) culture, one that emphasizes efficiency and resourcefulness while still delivering quality outcomes. Certificates of Analysis (COAs) sit right at the center of that tension.

COAs are crucial documents for compliance, confirming that a given batch of product meets its required specifications. At Nippon Paint, they are created and managed by the procurement team, drawing on handwritten and printed lab data submitted from facilities and third-party labs across the business.

Before fileAI, that process was entirely manual. Information arrived in inconsistent formats, in both English and Japanese, with no standardization in how it was submitted. Procurement staff had to read, interpret, and re-key this data by hand for every COA, a repetitive process that was slow by nature and prone to error by design.

Nippon Paint’s key goals for fileAI were clear:

• Improve COA creation speed

• Customize each COA by product

• Close culture and language gaps

• Process handwritten English and Japanese documents

A Procurement AI Agent built to match, create, and publish COAs automatically

fileAI’s AI-first solution empowered Nippon Paint to digitize, categorize, and accurately input large volumes of handwritten and printed technical data and lab results, with efficiency built in at every step.

1. Import and extract, from any format

Structured invoices and unstructured COAs flow into the system via Box, whether they arrive as clean digital files or handwritten lab submissions. fileAI extracts the underlying data regardless of format or language, closing the gap between English-speaking HQ functions and Japanese-speaking facilities and labs.

2. Auto-matching COAs to invoices

The Procurement AI Agent automatically matches COAs to invoices on a many-to-many basis, reflecting the real-world complexity of a business where a single invoice can cover multiple COAs, and vice versa. Matched COAs are renamed using the extracted data itself, replacing inconsistent manual naming with a standardized convention.

3. Automated folder creation and publishing

Once matched, the agent auto-creates invoice folders in Box and exports the corresponding COAs directly into them. A daily report of any unmatched COAs is pushed to Box as well, so nothing is left unresolved or hidden in a backlog. Published COA updates then flow back out through Box, with audit trails built in for full transparency.

4. Smart approvals, built around real business requirements

Underpinning the workflow is a layer of smart approvals, ensuring that COAs move through the process in line with Nippon Paint’s actual business requirements rather than a one-size-fits-all rule set.

Measurable ROI

Every quantified outcome from the engagement, in one place.

From procurement to the next frontier of automation

Automating the COA process did more than solve a single operational bottleneck. It gave Nippon Paint’s management team a proof point for AI-driven transformation, and a template for extending it further. Having seen a complex, repetitive process automated successfully in procurement, the team has identified HR and Finance as potential next steps for applying the same approach.

That is the pattern fileAI enables: not a one-off fix, but a foundation that pays down manual work wherever it is standing in the way of a leaner, faster organization, one department at a time.