Building our own systems in practice
Our current work includes migrating email functions on a VPS, setting up a sending relay, and rebuilding our company website. We use AI to inspect the existing configuration, identify the functions to preserve and build the new environment.
Email operation involves routing, sender authentication, permitted access and certificate renewal. We test sending and receiving on the new setup, checking both the configuration and its results.
Work quickly with AI. Verify through records and checks.
Our use of AI continues after the initial build. We research technical information, inspect configurations, apply changes and check the results. Moving quickly through that cycle also supports ongoing management and security work.
In this project, we checked access controls and email authentication and ran a simulated certificate renewal. We have also set up daily log collection and retention, and are developing operational checks for update status, known vulnerabilities and suspicious access. We value connecting implementation with records and verification.
This website was built with AI too
We use AI for the site's positioning, writing, structure, design and implementation. Requests such as a more off-road RC illustration or CAD imagery instead of product photos become part of the design through an ongoing conversation.
Product descriptions come from our product register, with reference imagery rendered from CAD files. We are organizing information on physical products through our retail partners in Japan and planned digital-file sales exclusively to customers residing outside Japan. We continue to update this website through conversation with AI.
Choose a direction, then test it in use
AI carries out the practical work of research, configuration, migration and production. We set the direction based on practical experience, then review the work records and actual behavior. People decide what to preserve, what to change and when to publish, and verify the results.
AI also helped structure and draft this article. We want to apply what we learn in our own operations to workflow improvement and making—and share the process along the way.