SGB Humangest Holding, a major Italian recruitment agency, is looking for an IT manager for its client in Rome.
We know little about the client company – the job description calls it “an important client company.” But essentially, that’s not really important. The problems they’re trying to solve are so universal that I could write this text with my eyes closed, just hearing “looking for an IT manager.”
Let’s look at their pain points. The company needs someone who will guarantee “operational continuity, security, efficiency, and reliability” of the entire IT infrastructure. This person must coordinate the team, manage the service desk, network specialists, monitor servers, backups, firewalls, antiviruses, respond to incidents, and even “promote a culture and awareness of cybersecurity.” Essentially, they are looking for an Atlas to shoulder the entire IT world of this company. Reliable, experienced, with 5-8 years of experience, who will be a strategist, a firefighter, and a psychologist for their team all at once. Classic. I’ve seen it hundreds of times. It’s a cry for stability and control in a world that’s becoming increasingly complex.
But what if I told you that most of this headache is just data? Huge streams of data that a human simply cannot process in real time. And instead of looking for a hero who will try to keep up with this avalanche, you can build a system that will do it for them. An AI-based system.
Imagine such a dialogue in the meeting room of that very Roman company.
Skeptic: We need a person. A person with experience who can sense where things might break. Who will look at the team and understand who is burnt out. A machine won’t do that.
Optimist (that’s me): You’re right, intuition is a great thing. But intuition is often just experience based on analyzing past mistakes. Modern AIOps platforms (Artificial Intelligence for IT Operations) do the same, but at speeds inaccessible to humans. Take Dynatrace or Datadog, for example. They don’t just show a red light when a server goes down. They analyze thousands of metrics per second – from CPU load to database response time – and say: “Look, in 2 hours, this service will most likely stop responding because there’s an anomaly in the request pattern.” This isn’t a replacement for intuition; it’s an amplification of it by an order of magnitude. A human manager learns about a problem when angry users call them. An AI system prevents the call itself.
Skeptic: Okay, infrastructure is clear. What about security? Incidents? Here, a cool head is needed to make a decision.
Optimist: Exactly. And nothing is colder or faster than an algorithm. SOAR (Security Orchestration, Automation, and Response) systems, such as Palo Alto Cortex XSOAR or Splunk SOAR, integrate with all your defense systems. When a firewall detects suspicious activity, the system doesn’t just send an alert to a manager who might be sipping their espresso. It instantly launches a scenario: checks the IP address against threat databases, analyzes traffic, if the threat is confirmed – it isolates the infected machine from the network, blocks the user, and only then generates a report for a human. All of this happens in fractions of a second. Your manager will only have time to blink. We remove human error and reaction time from the most critical moments.
Skeptic: What about the team? The service desk? Who will sort through requests, prioritize them?
Optimist: AI already does this perfectly. Modern Service Desk platforms, like Jira Service Management, use NLP (Natural Language Processing) to analyze the text of a user’s request. The system itself understands the core of the problem, determines its criticality, and assigns the right specialist. Moreover, simple requests like “I forgot my password” or “how to set up a printer” are resolved via a chatbot that provides the necessary instructions from the knowledge base. Your L1 support team is freed from routine and can focus on truly complex tasks. And the manager receives not a subjective assessment of “we’re swamped with work,” but clear analytics: what problems occur most often, where our “bottleneck” is, what knowledge the team lacks.
How to implement this and overcome distrust? No need for a revolution. Start small.
1. Pilot project. Take one area. For example, monitoring. Deploy an AIOps solution to observe a system that is not the most critical, but important. Collect data for a month. Show the team and management how many potential problems the system predicted.
2. Automation with confirmation. Configure the SOAR system not for automatic blocking, but for creating “recommendations for action.” Let a human press the “Approve” button initially. When the team sees that in 99% of cases, the AI’s recommendations are correct and save hours of work, they will ask to enable full automation themselves.
3. Transparency. Modern AIs are not black boxes. They show based on what data and anomalies a decision was made. This educates the team and removes the fear of a “machine uprising.”
And how to check if AI works better than a human? Very simply. With numbers.
Before implementation, you had certain metrics; afterward, they should be different.
– Mean Time To Detect (MTTD). For AI, it approaches zero.
– Mean Time To Resolve (MTTR). Automation reduces it significantly, if not by orders of magnitude.
– Number of incidents affecting the business. This will decrease manifold because problems will be resolved at a preventive level.
– Internal user satisfaction (CSAT). When their requests are resolved faster and more efficiently, they become happier.
In the end, the Roman company will get not just a “manager” who tries to keep everything in their head and put out fires. It will get a digital nervous system for its IT, which works 24/7, doesn’t get tired, doesn’t go on vacation, and provides the remaining team specialists with crystal-clear data for strategic decision-making. And the budget saved on the manager’s salary can be invested in these very specialists and the further development of the system. And that, in my opinion, is a much wiser investment in the future.
Источник: https://www.linkedin.com/jobs/view/4413667800/
