AMBL to Fully Support Autonomous System Operation with AI Agents

AMBL Inc. (Headquarters: Minato-ku, Tokyo; President and CEO: Masahiro Mori; hereinafter referred to as "AMBL"), with its vision to "Drive Japan's AX from the Ground Up," will fully launch its operation automation support service. This service enables AI to autonomously handle the entire process of system operation, including "detection, root cause analysis, correction, and sharing," by integrating generative AI with observability. By combining this with AMBL's strength in AI-driven development capabilities, we will not only accelerate incident response but also provide end-to-end support for building a system where customers can internalize their operations.

Background

Currently, interest in observability is rapidly increasing in the Japanese market.

According to a survey by Fuji Chimera Research Institute, the Japanese observability market is expected to expand from 94.6 billion yen in 2024 to 113.8 billion yen in 2028 (※).

With the advancement of generative AI utilization and cloud-native adoption, the complexity of system operations is increasing, making the assurance of reliability and operational efficiency a critical management issue.

On the other hand, many companies are stuck in a "passive operation" mode where failures are detected through user reports, and a "person-dependent" system where operations rely on specific experts.

As a result, MTTR (Mean Time To Recovery) is prolonged, leading to a high-cost structure. The risk of engineer burnout and turnover, as well as opportunity loss due to service downtime, are also serious challenges.

Against this backdrop of changing market conditions, the practical application of AI in system operations has finally entered its implementation phase.

The transition to a new operational model, where AI agents autonomously handle operational tasks, allowing humans to focus on essential decision-making, is becoming a realistic theme for Japanese companies.

In February 2026, AMBL, along with its parent company Dirbato Inc. (Headquarters: Minato-ku, Tokyo; President and CEO: Yasuhide Kanayama), entered into a partnership agreement with Grafana Labs (Headquarters: New York, USA; Co-founder and CEO: Raj Dutt). This partnership marked the full-scale launch of our unique AX support services, which combine the establishment of an observability foundation with AI utilization, allowing us to further address customer challenges.

※ For details on Fuji Chimera Research Institute's "Software Business New Market 2025 Edition" (published July 2025), please visit https://www.fcr.co.jp/report/251q06.htm.

Solution Overview

AMBL's operation automation support service integrates monitoring of metrics, logs, and traces with the inference power of generative AI, transforming incident response from "reactive and person-dependent" to "predictive and AI-driven." Human intervention is limited to the final stages of "approval and deployment," significantly reducing end-to-end response times. Specifically, it consists of the following four functions:

1 Predictive Detection (Detection): Integrates monitoring of metrics, logs, and traces to instantly detect anomalies and issue alerts.

2 Root Cause Analysis (Analysis): Generative AI comprehensively analyzes error logs and GitHub commit history. It identifies the root cause by correlating past changes with failures.

3 Automatic Correction (Fixing): Based on the identified cause, AI automatically generates corrected code and test code, presenting them as Pull Requests on GitHub.

4 Collaboration and Sharing (Collaboration): Automatically generates dedicated channels in Slack/Teams to share real-time information from detection to proposed fixes with relevant parties.

The key feature is that these are not merely the introduction of operational tools. By combining them with AMBL's expertise in AI-driven development, we can accompany customers in building an internal system that they can manage themselves.

Expected Effects

By implementing this support, it is expected that MTTR (Mean Time To Recovery) can be reduced by up to approximately 70% through the reduction of log collection and investigation time. Furthermore, by having AI complete a certain level of initial response and situation assessment even in the absence of experts, we can resolve person-dependency and shift engineers from "investigation and recovery work" to "decision-making and approval" tasks. Additionally, by reallocating resources to high-value-added tasks such as recurrence prevention and service improvement, operations will shift from "defense" to "offense."

Future Outlook

AMBL will support Japanese companies' AX in areas such as AI-driven operational decision support, incident response automation, and cost optimization.

Specifically, we will streamline operational tasks that have traditionally been person-dependent and time-consuming using AI. By combining this with AMBL's strength in AI-driven development capabilities, we will provide end-to-end support for building internal systems that customers can manage themselves.

By accumulating these initiatives and support achievements, we aim to contribute to Japan's AX.