How generative AIOps will revolutionize IT service delivery
Generative AI (GenAI) has burst into day-to-day life, from reshaping education and the arts, to changing the nature of work in organizations across every sector.
In the IT world, GenAI is helping tech experts reinvent how they develop software. It can write code, program and test applications, and automate development processes end-to-end.
What’s less clear is how it will improve the way IT services are delivered and managed for large enterprises. This convergence of AI and IT operations — called AIOps for short — is an area of opportunity that Atos has been exploring over the last year.
How AIOps is transforming IT Service Management
Since the early 1990s, IT Service Management (ITSM) has remained largely unchanged. Simply put, teams of skilled IT professionals monitor IT services 24/7, logging each issue or alert as an incident, then triaging them according to urgency, impact and severity. The most straightforward incidents are immediately resolved by the front-line team; the more complex ones are routed to the right expert team for fast and effective resolution.
In an AIOps scenario, an AI learning model controls a robotic process automation (RPA) engine, which intelligently automates the first three or four steps of the ITSM process. It effectively eliminates all the human effort required to gather, analyze and summarize the relevant data, prioritizes incidents, and triggers the automated and manual actions necessary for resolution.
It's arduous and time consuming to evaluate and act on hundreds of alerts per day, but AIOps can pinpoint in seconds what a human might take hours to find.
Massive time and resourcing benefits
This GenAI-powered "robot tech engineer" has potential in any area of ITSM, from incident management and change management to the orchestration of servers, applications and databases.
This means that using AIOps, a two-person team could operationally manage the same IT landscape as 10 people previously – from initial monitoring, through to ensuring human resolution of the most challenging or complicated incidents.
AIOps will dramatically accelerate every ITSM process, contributing to the ultimate goal of reducing time to repair (a key metric in ITSM). In some cases, it could even prevent problems from happening. The benefits are significant. Given that any outage is a potential disruption to the business, reducing the time to repair from eight hours to eight minutes is of huge value to enterprises and IT teams.
Harmony between people and tech
Crucially, this is not about taking human intelligence out of ITSM. Rather, it removes the burden of lower-value tasks so your talented IT specialists can focus on applying human insight and ingenuity where it can be best used.
One area where AIOps truly excels is in handling the vast volumes of data involved in ITSM. It can be arduous and time consuming to evaluate hundreds of alerts per day, to determine which have the biggest impacts and assess their precise root causes. Across thousands of IT assets and services, AIOps can pinpoint in seconds what it might take hours for a human to find.
Imagine, for example, an initial call between six IT experts who are all discussing a complex, high-priority incident. Also on the call is the AI tech engineer, ready to provide a comprehensive analysis of what caused the problem, what systems it is impacting, and the next best action. Everyone on the call can interrogate the robot in real time to quickly determine the best outcome for the business.
Mitigating the risk of AIOps
In discussions with enterprises, we have listened to their concerns about the perceived risks of handing processes over to machines. The same questions emerged in the early days of process automation in ITSM. The reality is that when an AIOps model can simulate a million ways to solve a problem and continuously self-learn, it becomes highly effective – and even faster than the most experienced human engineer.
However, any AI application is only as good as the datasets on which it is trained — and the learning models it creates from them. In this case, success depends on how you select and use data sources and how exhaustively the model is trained. A key factor is setting the right success parameters to ensure the right outcomes, and knowing at which points to transfer the process to a human.
The roadmap game-changing results with AIOps
The Atos R&D team has been working on an integrated AIOps offering, and based on our vision and progress so far, the team has a roadmap for the next two years. As with any transformative program, this won’t be a big bang. We are working on incremental proofs of concept, starting with straightforward incidents and building from there. At every step, the purpose is to support and empower human IT engineers to maximize the value they deliver to your enterprise.
GenAI-powered IT service management will be game-changer — especially in combination with capabilities such as Atos Bridge, which connects ITSM performance directly to business performance and orchestrates the necessary outcomes. Integrating AIOps will provide enterprises with next-generation IT — resilient, adaptive and optimized — delivering exactly what the business needs at every touchpoint.
Posted on: July 23, 2024
Willem van den Boogaard
Vice President Service Integration & Orchestration (SIO) Bridge
View details >
