サービス開発部兼ドラクエ部のくればやしです。 今度、淡路島のニジゲンノモリというところにある[ドラゴンクエスト アイランドに遊びに行く予定です。アトラクションもさることながら、グッズコーナーに行くのも非常に楽しみです。 nijigennomori.com はじめに 前回の記事では、既存のLambda関数をAWS::Serverless::Function としてインポートする方法を解説しました。 blog.serverworks.co.jp 本記事では一歩進んで、API Gateway(REST API)をAWS::Serverless::Api としてインポートする方法を解説したいと思います…

On Monday, April 20, 2026, AWS held the invitation-only event “AWS Retail & CPG EXPO 2026 — Build the future with AI” at the AWS Tokyo office. We welcomed more than 100 guests from over 60 companies, and the event concluded on a high note. We localized AI agent demos for the Japanese market, providing an opportunity to see, touch, and experience them firsthand. A networking reception was also held from 6:00 PM, offering attendees a chance to connect with one another. For those who were unable to attend, we would like to share the highlights of the event through this blog post. The Future of the Industry, Transformed by AI Agents The age of AI is evolving into the age of AI agents. We are now in an era where the key question is how autonomously we can deploy AI agents and apply them to our business operations. Alongside the platforms that enable AI agents to operate safely, a critical theme has emerged: how to leverage them in the retail, consumer goods, and food service industries. However, in Japan, there is a prevailing reality of “we understand, but we can’t move forward.” This event was designed to help break through that barrier — giving attendees the chance to experience real, working AI agents and discover their next steps from the case studies of companies that are taking on the challenge. Exhibition Area — Experience Real AI Agents In the exhibition area, we localized demos announced at NRF 2026 for the Japanese market and presented 4 categories across 7 booths, aligned with retail, consumer goods, and food service industry workflows: product development & production planning, pricing strategy, supply chain, and in-store/online customer experience. A common concept across all demos was “Multi-Agent × Human-in-the-Loop.” Attendees experienced through live demos the design principle where AI agents collaborate autonomously while humans make the critical decisions. Detailed materials for each exhibit can be downloaded from the links in each exhibit description below. If any exhibit caught your interest, please feel free to contact your account representative to arrange a private demo. Merchandising — Product Development & Pricing Strategy Product development and pricing strategy are critical areas that determine a company’s competitiveness. We showcased two multi-agent demos. Luggage Lab: Attendees saw a demo where three agents — Researcher, Designer, and Planner — collaborate to accelerate product innovation. ( Luggage Lab detailed materials ) Luggage Lab — Agent Role Diagram Luggage Lab — Architecture Retail Pricing Agent: We presented a demo where three agents — Competitive Research, Demand Analysis, and Price Decision — collaborate to automate pricing strategy using generative AI. ( Retail Pricing Agent detailed materials ) Retail Pricing Agent — Process Flow Retail Pricing Agent — Architecture Supply Chain — Autonomous Response to Disruptions Supply chain disruptions are a constant and significant risk for the retail, consumer goods, and food service industries. Attendees experienced a scenario where multiple agents collaborate to gather information and propose countermeasures, while humans make the final decisions. Agentic Supply Chain: We exhibited a demo where four agents — Procurement, Inventory, Planning, and Logistics — collaborate, with AI agents autonomously responding to supply chain disruptions. ( Agentic Supply Chain detailed materials ) Agentic Supply Chain — Specialized Agent Team Structure Agentic Supply Chain — Architecture Omnichannel — Transforming the Customer Experience Seamlessly connecting digital and physical stores is a critical theme for the industry. We introduced three demos showcasing new customer experiences where AI provides personalized engagement for each individual. Smart Beauty: This demo uses image analysis to classify skin types into 16 detailed categories and automates improvement recommendations. AI replicates the expertise of beauty consultants. ( Smart Beauty detailed materials ) Smart Beauty — Skin Type + Personal Color Analysis Smart Beauty — Architecture Fix&Fab: Attendees saw a demo that generates DIY repair instructions from just a photo and description, and even provides shopping lists and expert referrals. ( Fix&Fab detailed materials ) Fix&Fab — Project Design from a Single Photo Fix&Fab — Architecture Retail AI Concierge: We presented a demo that delivers a seamless journey from e-commerce to physical stores — from product recommendations and inventory checks to event guides and visit planning. ( Retail AI Concierge detailed materials ) Retail AI Concierge — A New Customer Experience Retail AI Concierge — Architecture Product Innovation — Next-Generation Personalization Returns and sizing challenges are major themes across the industry, whether in e-commerce or physical stores. Bodd: We exhibited a demo featuring Bodd’s contactless body scanning technology that delivers precise measurements in just 60 seconds, enabling cross-brand size recommendations. Bodd — Precise Measurements via Body Scanning Bodd — Customer Journey If any exhibit caught your interest, please feel free to contact your account representative to arrange a private demo. Sessions — Learning from Companies Taking on the Challenge The sessions consisted of AWS opening sessions (2 sessions), customer case studies (4 sessions), and an AWS session (1 session). AWS Opening Sessions Keanu Nahm — Common Success Factors in AI Adoption Seen from a Global Perspective, and the Opportunities for Japan’s Retail & CPG Industry Keanu Nahm (Head of AWS Global Retail & CPG Business Development, Japan) drew on trends from NRF 2026 and ShopTalk 2026 to describe how the global retail and consumer goods industry is transitioning from a “year of experimentation” to a “year of implementation” with AI. He identified three conditions for AI to be adopted broadly and deeply: “ease of use (zero friction),” “trustworthiness,” and “becoming a habit,” and highlighted the opportunities this presents for Japan’s retail and consumer goods industry. (Session materials are available for download here .) Kempei Igarashi — Experience How AI Agents Are Transforming the Retail Frontline, and Find Your Next Step from the Case Studies of Companies Taking on the Challenge Kempei Igarashi (Head of Industry Solutions Division, AWS) shared AWS’s perspective on the current state of AI agents and the trend of AI agents evolving from “tools” to “colleagues,” and provided an overview of the event’s exhibitions and sessions. (Session materials are available for download here .) Customer Case Study Sessions KOSÉ Corporation — KOSÉ’s Challenge: Company-Wide Deployment of Generative AI — Building a Framework for Frontline Adoption Haruka Yokoyama, Generative AI Promotion Leader, DX Promotion Section, Information Management Department Miku Kaneda, Generative AI Promotion, Infrastructure Development Section, Information Management Department KOSÉ Corporation presented their case study on company-wide deployment of generative AI, driven by the awareness that “even a great system doesn’t mean everyone can use it effectively.” Led by the Information Management Department, they simultaneously advanced system development and user support/education. They created an environment that encourages voluntary adoption through intuitive UI design, easy model selection, and gamification-driven engagement. With the support of AWS’s Prototyping team, they built the platform in just one month and shared how generative AI has become a “common language” across the organization. asken, Inc. — The Walls “asken” Overcame in Developing New Features Starting from Vibe Coding Takuya Ito, Senior Product Manager, AX Promotion Department, Product Development Division Yoshihiro Iwama, Senior Tech Lead, Product Development Department, Product Development Division asken, Inc., the company behind the meal management app “asken,” presented on a new co-creation process between product managers and engineers. Product managers use Vibe Coding to create “working PRDs (prototypes)” and run user validation cycles on the AWS-based experimentation platform “asken Lab.” However, they also shared the “painful lesson” of development costs tripling when they tried to use the working PRD directly in production. From that failure, they established a refinement process leveraging AI-powered reverse engineering, arriving at a “design of sequence and boundaries” that allows product managers and engineers to maximize their respective expertise. GOLDWIN Inc. — AI: We Understand, But Can’t Move Forward — How to Bridge the Gap Between the Frontline and Management Takahiro Suemitsu, Senior Expert, Corporate Planning Division GOLDWIN Inc. addressed the problem many companies face: “AI gets stuck at search and summarization.” They identified five barriers preventing AI from reaching implementation (ambiguous management expectations, lack of bandwidth on the frontline, absence of success metrics, etc.) and visualized the “invisible gap” between management and the frontline. As a breakthrough, they advocated the approach of “showing something that works rather than explaining with a proposal,” and introduced a case where they built a prototype in just one day. The message that resonated most was: “The phase of explaining and waiting for understanding is over. Let’s make 2026 the year we embed AI into our operations.” CAINZ Corporation — The Fitting Room Experience Realized in Next-Generation Stores Takehiko Suga, General Manager, CX Management Department, Information Systems Division, CAINZ Corporation Tsuyoshi Mukai, Manager, E-Business Engineering Section, Digital Engineering Department, Digital Transformation Division, AsiaQuest Inc. CAINZ Corporation and AsiaQuest Inc. presented their “CAINZ Fitting Room” initiative, addressing the in-store purchasing challenge of “not being able to try before you buy” using generative AI. While they had previously attempted to visualize room coordination, reproducing textures had been a challenge. With advances in generative AI, they solved this using image generation powered by Amazon Bedrock. Furniture and curtains can now be virtually placed in room images, with realistic textures and shadow reflections. The experience is currently available on in-store touch panels, and future plans include allowing customers to check arrangements using photos of their own rooms and AI-generated coordination recommendations. This initiative was also featured in Mynavi TECH+ . AWS Session Koji Matsumoto — Transforming Advertising & Marketing in the Agentic AI Era Koji Matsumoto (Principal Business Development Manager, Strategic Business Development Division, AWS) explained three “tectonic shifts”: the rapid growth of retail media, the rise of Agentic AI, and the gap between AI investment and implementation. He introduced the latest trends in the democratization of advertising and marketing through Agentic AI — including Amazon Ads’ Creative Agent and Unified Campaign Manager — along with three recommended actions. (Session materials are available for download here .) Thank You for Attending Thank you to everyone who attended AWS Retail & CPG EXPO 2026. AI agents are beginning to take a central role in business operations. The world is already moving. Your company’s transformation starts with today’s “this could work.” Start small, and AWS will be right there with you. For those who were unable to attend, please feel free to reach out to your AWS representative. We look forward to seeing AI agents come to life at your workplace. Event Information Event: AWS Retail & CPG EXPO 2026 — Build the Future with AI Format: Invitation-only Date: Monday, April 20, 2026, 1:00 PM – 6:00 PM JST Venue: AWS Tokyo Office Attendees: More than 100 guests from over 60 companies Organizer: Amazon Web Services Japan G.K. This blog post was translated by Tomo Yamashita, Solutions Architect at AWS Japan. The original post is available here .

1. はじめに 本稿では、Kiro を用いた仕様駆動開発の検証を通じて、API 開発における生産性と品質の変化を評価した事例を紹介します。 実際の開発案件で実装された領域に対して、設計から試験工程までを Kiro で再実装し、従来開発と比較するという試みになっています。 2. 自己紹介 株式会社 NTT データ ソリューション事業本部 C&D 事業部に所属しており、パブリッククラウド領域の案件に取り組んでいます。 本稿の取り組みのような 生成 AI を活用した開発にご興味をお持ちいただけた方は、お気軽にお問い合わせください。 3. Kiro における仕様駆動開発 一般