Skip to content

Published 30 Jan 2025· via Microsoft· Customer story

Woven by Toyota: Multi-agent Azure OpenAI workflow to auto-fix MISRA compliance errors

Toyota · Automotive · Japan

Headline result

The source does not publish a single comparable figure. Reported results are listed below.

Summary

Woven by Toyota, part of the Toyota Group, used Azure OpenAI Service to automate MISRA code-compliance fixes for embedded C/C++ software in autonomous driving and ADAS development.

The team built a multi-agent workflow with Coder, Reviewer, and Evaluator agents to generate fixes, review them, and provide reasoning and certainty for engineers.

The implementation integrated Azure App Service, Azure Cosmos DB, AutoGen, and GitHub Enterprise CI/CD, and was tested on sample code and in-house code.

The problem

  1. 01MISRA compliance for embedded C/C++ software required large amounts of manual effort.
  2. 02Engineers needed to learn hundreds of pages of standards while static-analysis error volumes were extremely high.
  3. 03The team needed faster, more reliable code correction for automotive software development.

What they built

  1. 01Woven by Toyota first tested Azure OpenAI Service GPT-4o to automatically fix MISRA compliance errors in sample and in-house code.
  2. 02The project evolved into a multi-agent system with a Coder agent that fixes code, a Reviewer agent that critiques and iterates on the corrections, and an Evaluator agent that produces reasons and certainty factors.
  3. 03The solution used Azure OpenAI Service reasoning and coding models, AutoGen for orchestration, Azure App Service, Azure Cosmos DB, and GitHub Enterprise integration to support development workflows.

Reported results

  1. 01The automatic correction rate improved from about 50 percent to about 80 percent with the o1 model.
  2. 02In a proof-of-concept on in-house code, the system achieved a 97.1 percent code generation success rate.
  3. 03The project automatically corrected 81.5 percent of MISRA compliance errors.
  • Automated MISRA compliance error correction
  • Built multi-agent code-fixing workflow
  • Enabled faster code correction for engineers

Source

Microsoft

Customer story · 30 Jan 2025

Read the original(opens in a new tab)

How this deployment compares

Computed from the library · data 10 Nov 2026

Toyota has 5 deployment records in the library, from 2023 to 2026. It is one of 26 legal workflow automation deployments, 1 of them in Automotive. Like 63% of records, its source gives no figure.

Toyota in the library

5 deployment records from Toyota, 2023–2026

  1. Sept 2023Toyota revolutionizes predictive maintenance models with AIPredictive maintenance
  2. Nov 2024Toyota Revolutionizes Vehicle Design Process with Multi-Agent AIAutomotive operations automation
  3. Jan 2025Toyota, BMW, and CarMax accelerate automotive innovation and transformationAutomotive operations automation
  4. Jan 2025This record · Legal workflow automation
  5. May 2026Toyota Motors North America Modernizes Predictive Maintenance with AWSPredictive maintenance

Followed upA newer Toyota record from 2026 is in the library, so the AI program continued after this announcement.

How Toyota uses AI: company page

Same pattern, whole library

One of 26 legal workflow automation deployments

  • Automotive accounts for 1 of them (4%), ranking #7 of 11 industries.
  • 7 records of this pattern came before it; 14 have been published since.
  • Only 6 of the 26 put a number on the result (23%).
Provider mix
  • Microsoft65%
  • AWS19%
  • GCP12%
  • Alibaba4%
Most-used technology
  • Azure OpenAI 5
  • Copilot 4
  • Microsoft 365 Copilot 4
  • Amazon Bedrock 4
  • Microsoft 365 3
  • Copilot for Microsoft 365 3

Outlined: used in this record.

Shares describe this library, not the whole market.

Legal workflow automation analysis

Results against peers

No figure here. What peers report

Like 63% of records in the library, the source gives no figure. Peers that do quantify similar claims report:

  • Risk, reliability & safety13 figures, all industries · middle half 60%–90%70%median reduction
  • Time & speed283 figures, all industries · middle half 30%–76%50%median reduction

Peer figures are as reported by their sources; percentage claims only.

How we checked this record

Evidence tier: supported

A credible source with a named customer, but not the organization's own account.

Named organization
Toyota
Who published the source
The organization itself
Deployment status
The source does not say
Quantified result
The source reports no figure
Summary checked against source
4 / 5 agreement

Last re-checked 13 Jun 2026.

This record is evidence for

Insight · industryAutomotive dossierOne of 96 Automotive deployments. See which patterns dominate there and what they report.Insight · patternLegal workflow automation26 deployments share this pattern. See where it is used and how it is built.Trend · momentumBuilding pattern7 new deployments in the last six months, against 2 in the six before.

Same pattern, other organizations

All legal workflow automation deployments
  1. 22 Jul 2026
    Professional Services

    eSignGlobal launches eSignGlobal 2.0 with Qwen via Alibaba Cloud Model Studio for contract lifecycle management

    eSignGlobal · Legal workflow automation · Alibaba

    eSignGlobal, based in Hong Kong, launched eSignGlobal 2.0 as an intelligent Contract Lifecycle Management (CLM) system built with Alibaba Cloud Model Studio and Qwen models. The platform adds AI Translator, AI Risk Assessment, AI Summarizer, AI Comparator, and AI Draft Generator capabilities to support multilingual, cross-border contract workflows. The company also integrates Alibaba Cloud Elasticsearch and other Alibaba Cloud services into its e-signature SaaS platform and API offerings.

    3 min
    draft generation time
  2. 13 Jun 2026
    Healthcare

    NSF accelerates regulated medical audits using agentic AI with Azure Document Intelligence and Azure OpenAI

    NSF · Legal workflow automation · Microsoft

    NSF, a nonprofit scientific and regulatory auditing organization, needed to organize, verify, summarize, and synthesize tens of thousands of documents across country-specific rules for medical audits. The solution uses Azure Document Intelligence, Azure OpenAI, Azure Model Context Protocol (MCP) tools and servers, Azure Blob Storage, Azure Python SDK, Azure Cosmos DB, Microsoft Entra ID, and Azure RBAC to automate document validation, structured sorting, version tracking, and summary drafting. Staff review and refine the AI-generated summaries, while the workflow remains inside Azure cloud controls and private tenant security. The implementation reduced audit turnaround time from 4–6 weeks to about 2 weeks and was reported to deliver near-perfect accuracy for the proof of concept.

    AgenticPoC

  3. 3 Jun 2026
    Professional Services

    ALPHAWAVE: Vertex AI-based proprietary AI agents automating browser workflows powered by Veo/Flow

    ALPHAWAVE · Legal workflow automation · GCP

    ALPHAWAVE is an Amsterdam-based AI-native scale-up and consultancy with presence in New York. It builds proprietary AI agents on Google Cloud infrastructure and Vertex AI to automate repetitive browser-based tasks such as reporting and format scaling. It also uses Veo and Flow to create photorealistic, fast AI video assets for brand campaigns and internal training and awareness content.

    Agentic

  4. 3 Jun 2026
    Professional Services

    Mantel case study | Google Cloud

    Mantel · Legal workflow automation · GCP

    Mantel, a tech consultancy and Google Cloud Partner, uses Google Workspace and Gemini Enterprise to improve consultant efficiency and client collaboration. The company built a Gemini Enterprise agent to automate multi-step Statements of Work (SOW) preparation by integrating Salesforce, Slides, Docs, and Confluence. Mantel also uses Gemini Enterprise to summarize project history and NotebookLM to accelerate research-paper production, and it uses the Google Agent Development Kit to build custom agentic automation for HR requests with human-in-the-loop review.

    Agentic