Guotai Junan’s AIOps Success: Inside the Award‑Winning Intelligent Operations Platform

Intelligent Operations (AIOps) applies artificial intelligence technologies such as machine learning to IT operations problems, enhancing and partially replacing core IT operational functions. Gartner describes AIOps as extracting and analyzing ever‑growing volumes, varieties, and velocities of IT data in a loosely coupled, scalable manner to support IT operations management products.

On December 26, 2022, the China Academy of Information and Communications Technology (CAICT) announced the latest batch of AIOps standards assessment results. Guotai Junan Securities’ “Digital Intelligent Operations Service Platform” passed the CAICT “Cloud Computing Intelligent Operations (AIOps) Capability Maturity Model – Part 2: System and Tool Technical Requirements” assessment, earning an “Excellent” rating for the anomaly detection module, marking the firm as the first securities company to achieve this level.

Interview Highlights

Q: Please introduce yourself, your company, and the evaluated project.

Yu Feng, CIO of Guotai Junan, described the firm’s long‑term leadership in comprehensive financial services and its “SMART investment bank” vision, emphasizing digital transformation, data‑driven services, and the development of a unified digital platform for wealth management, trading, and research.

Q: What considerations led you to participate in the AIOps evaluation?

Yu Feng explained that the CAICT AIOps standards provide a complete, professional, and authoritative evaluation framework, helping the company accurately assess its AIOps maturity and guide continuous improvement, while also allowing benchmarking against industry peers.

Q: What changes has the evaluation brought to your enterprise and team?

The assessment reinforced the company’s digital transformation, focusing on growth, efficiency, experience, and security, and guided the data center to build an integrated intelligent operations “brain” that leverages data and algorithms for real‑time anomaly detection and decision‑making.

Q: What are the next steps for AIOps development?

Feng Yixin outlined three trends: strengthening data governance to ensure high‑quality data for AI; tailoring AIOps scenarios to address specific pain points such as alarm reduction and log clustering; and platformizing AIOps to create a reusable, composable intelligent operations foundation.

Technical Overview

The platform, launched in 2016, follows a “data‑driven, intelligent, automated” architecture, covering monitoring, management, and control of core business systems. It addresses challenges of rapid business iteration, massive distributed micro‑service environments, and mixed legacy and modern architectures.

The anomaly detection module, powered by machine learning, provides real‑time analysis of metric and log data, achieving over 80% detection accuracy and establishing a closed loop of algorithm development, delivery, evaluation, and optimization.

AIOps Maturity Model

The CAICT AIOps Capability Maturity Model, co‑created with leading internet and industry players, defines eight assessment modules: anomaly detection, fault prediction, alarm convergence, root‑cause analysis, fault self‑healing, fault prevention, capacity prediction, and knowledge‑base construction.