How to separate hype from help when bringing AI into your Operations Center

How to separate hype from help when bringing AI into your Operations Center
  • Homepage
  • >
  • All News
  • >
  • How to separate hype from help when bringing AI into your Operations Center

You’ve probably heard about all of the incredible promises AI makes, from autonomous infrastructure to incident resolution. All the stories of horror that have accompanied them, including how AI would replace operations teams and create better results than humans. In the real world the hype can often take the focus off the help that AI offers to overwhelmed, stressed teams. AI in the Operations Center doesn’t replace humans, it enhances them. AI can enhance monitoring, response, and maintenance workflows in a practical way. This article clarifies the differences between buzzwords and business benefits.

1. Hype: AI Replaces Entire Ops Teams. Help: AI Enhances Human Response Times.

Imagine yourself totally hands-off in terms of monitoring your system. Many vendors will claim “zero touch” operations. You dream about AI constantly monitoring your infrastructure, finding anomalies and correcting them instantly. You only need to receive the AI summary the next morning. Even the most sophisticated models struggle to provide the context and nuance necessary for this performance. AI is able to reduce dramatically the time required to analyze alerts, detect them, and then react. AI is used by observability platforms such as Dynatrace and New Relic to provide real-time application metrics. Operations teams can focus their attention on the root cause and fix in seconds, instead of having to sort through thousands alerts. This collaboration between AI and the human judgement can help you improve your mean resolution time (MTTR).

2. Help: AI detects problems before they escalate.

You’re right, it sounds like magic. It’s a myth that “predictive AI” can predict problems before they occur and automatically schedule repairs. You should expect your systems to run smoothly and without surprises. However, in reality, to be able to predict, you need a large dataset of high quality and trained models. The models need to be refined over time and the majority of environments are too dynamic to make blanket predictions.

Even so, the addition of AI to Operations Centers with an emphasis on predictive capabilities is extremely useful in certain use cases. IBM Watson AIOps, for example, can identify early signs of app crashes by analyzing patterns from historical data. Google Cloud Active Assist uses AI to detect misconfigured policy and underutilized resource. So, they can be caught before they affect availability. These AI systems won’t solve all of your issues, but will allow you to intervene sooner, reducing downtime, and keeping your service robust.

3. AI fixes security holes automatically. AI-Assisted Software for Patch Management Predicts Priority. AI-Assisted Patch Management Software Predicts Priority

A common claim is AI will detect vulnerabilities automatically and immediately deploy the correct patch. There will be no need for human supervision. It’s obvious that the truth is much more complex. The context of patching, including compatibility, risk, and timing are all important factors. AI cannot yet weigh the full range of business consequences. AI has transformed patch management into a smart, useful process. AI-assisted platforms such as Automax and Threatlocker rely on machine learning in order to assess the importance and severity of newly discovered vulnerabilities. Your specific environment will be taken into account, as well as CVSS scores and exploit activity. Also, they’ll look at the history of your patches. They’ll then recommend which patches to prioritize. This will allow your team to focus more on closing the high-risk holes faster.

4. AI Chatbots can handle every ticket. AI Triage Bots Increase Workflow Efficiency. Many people believe chatbots will be the future of customer service. These chatbots can handle tier-1 customer service, ticket routing, escalated issues, or even resolve them with little human interaction. Chatbots are unable to understand complex problems. This dramatizes the fact that they lack the understanding necessary for them. You know that context is key if you work in operations. However, your chatbot does not. Chatbots are useful for triage.

Moveworks, Aisera and other tools can translate natural-language tickets or alerts. They can then pull up relevant knowledge base articles and send the issue on to your team. If a user sends in a ticket stating “VPN is down in Chicago”, the bot will tag the message correctly, add it to monitoring alerts and log the incident. Now, your operations team will spend less time on the resolution.

5. Hype: AI Guarantees Zero Downtime. AI enables smarter self-healing infrastructure. The marketing slogan “zero time” is gaining popularity. The AI system is so sophisticated that it promises to eliminate all outages. Even hyperscalers such as AWS and Azure are not immune to disruptions. They are usually caused by bugs that occur rarely, dependencies in certain regions, or cascading problems. Yes. Yes. It cannot, however, eliminate all downtime.

Self-healing AI infrastructure is far more powerful and achievable. Kubernetes-based platforms that use tools such as Karpenter, or OpenShift’s AI autoscaling feature can react immediately. The system can detect spikes in usage, failures of nodes, or container crashes without human involvement. These systems, which are based on policy-based automaton, can also restart services and reallocate work loads. Even failed deployments can be automatically rolled back. Although they are not perfect, the software can help your team save hours of work.

Bring AI to the Operations Center does not mean that your team is replaced. It means they are given superpowers. AI will not replace your Operations Center. It can, however, change your team’s way of working for the better. Focus on use cases that have measurable results. When you allow AI to augment and improve human expertise, it shines. Asking questions like “What is the specific pain this solution solves?” or “Can we integrate it with our existing tools?” and “Can it reduce the time to action?” will help you separate hype from real value.

View Article Source

Share Article
Facebook
LinkedIn
X
Ascot marketing agency says small businesses can win Black Friday week without cutting prices
Harley Street-trained psychotherapist moves his practice to Poland to treat British expats online
DoorDash to buy Instacart in $17bn deal, its largest takeover to date
Secret Link