Kyoei Joho wins AWS AI Services Competency for production AI

Kyoei Joho's certification comes as many Japanese companies have tried generative AI but struggle to move beyond pilots. The company says authentication, access control, cost and answer quality are common blockers, and it has built internal standards covering readiness assessment, KPI design, model selection and re-evaluation, security and privacy, responsible AI, production builds and operations runbooks. AWS reviewed those standards together with four customer cases, which is what the competency recognizes.

The case details show what "production" means in practice. At Kajima's Technical Research Institute, RagChat runs on Amazon Bedrock Knowledge Bases and Amazon Aurora PostgreSQL with pgvector, and the company keeps monthly reports and periodic answer-quality reviews after launch. At Yakuji Keizai Sha, the translation team switched foundation models and expanded a glossary after mistranslations appeared, and the improvement history was recorded. Kyoei Joho also helped an equipment construction firm move a GenU-based environment from trial to company-wide deployment with Microsoft Entra ID single sign-on and department-level access control, and shortened corporate environment setup for a care-business SaaS from 4–6 hours to within 30 minutes.

The certification is Kyoei Joho's second AWS competency, following the AWS SMB Competency obtained in December 2025. President Kiyoshi Takeda said the company will deliver generative AI to mid-sized and small companies at the same security and operational quality as large enterprises. The company plans regular seminars and industry-specific best-practice posts on its cloud5.jp blog.