> If you haven’t read part I, first read “ITSM and Artificial Intelligence – a Good Marriage?”.
Automation and AI
In the first part of this series, we discussed the increasing complexity of IT environments and the need for automation to alleviate the burden on IT staff. While automation can significantly reduce repetitive work, it’s not the only solution. Manual triggers and business rules have their limitations, especially as IT environments become more diversified.
What is AI?
According to Wikipedia, the general definition of AI is: “Any intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and other animals. In computer science, AI research is defined as the study of ‘intelligent agents’: any device that perceives its environment and takes actions that maximise its chance of successfully achieving its goals.”
In simpler terms, AI is the capability of a system to evolve knowledge based on training with known data constantly. It cannot recognise anything without training and self-learns based on known data. Compared to business rules, the most significant advantage is:
The Power of AI
This capability is powerful because, based on the proper learning of the algorithm, it can start concluding similar situations without an exact match, as is the case with business rules. This means fewer algorithms can achieve the same effect as 40,000 business rules, with a probability percentage that can be influenced by tuning the algorithm. Based on available data (volume and quality), there are multiple ways to train the algorithm, called Machine Learning. The more data and higher rate will lead to higher probability percentages and, thus, more effective algorithms.
AI in Various Industries
AI has recently been used in various industries to become more effective and proactive. Think of airport service bots that help with traveller queries, the elevator industry that predicts maintenance to save on unplanned downtime, the car industry’s quality improvements before shipping products to clients, and more. There are still many opportunities in IT to elevate service quality and proactivity.
Use of AI in ITSM
Let’s look at opportunities to improve service and reduce the effort needed. It’s important to understand that we might need to rethink some processes to maximise the investment’s value. If we are not prepared to adjust some methods, the result of adding new AI technology might be akin to the following formula:
Old business process + investment in new technology = More expensive old business process
Maturity of AI implementation
To better explain this bold statement, we need to understand the maturity levels of AI implementation:
- Informative: The AI algorithm supports staff with information and conclusions. Staff needs to act on the event.
- Reactive: The AI algorithm detects issues or registered service requests and triggers automation to fix the issue.
- Predictive: The AI algorithm predicts what will happen next and signals this to staff for intervention.
- Preventive: The AI algorithm will prevent issues and outages by automatically detecting and resolving possible upcoming issues.
The technology is ready to perform at the above level 4. But are we, as human beings, prepared for this as well? I’ve seen many situations in various industries where letting go with proper reporting is difficult.
This situation is often underestimated and eventually leads to the result of the above equation: More expensive old business processes. Therefore, any implementation of AI should be closely monitored to determine if the implementation goal is the best possible goal or just a suboptimal improvement.
Data, Data, Data
The value of data is usually more than what we think it is. There are many relationships between different types of data that are most of the time not used. Think about monitoring data, service request data, problem and known error data, financial data, project progress data, maintenance data, historical data, supplier-related data, etc. Putting relationships between these data types for different processes leads to better insights into the strategies, performance, and proactivity. It will maximise the value of our data to its full potential.
There are, however, some barriers we need to overcome. One of the most significant barriers is the siloed infrastructure. As stated in the beginning, the number of platforms in infrastructures has grown substantially.
Not all platforms are accessible when it comes to integration and data exchange. This slows down the progression toward digitisation in IT, but there are ways of handling this barrier. Two of the most common are integration platforms and immense data structures.
Practical Pointers for Implementing AI in ITSM
- Service Level Management: Regarding ITSM, stability and predictability are paramount. Traditional service level management typically involves looking back and evaluating the past, determining whether it was a good or bad month. However, the integration of data and AI presents new possibilities. Analysing historical data makes it relatively straightforward to perform trend analyses on critical resources such as people, storage, network, and processing power. This allows for a better understanding of fluctuations in capacity needs. Further, introducing algorithms enhances predictability, enabling organisations to foresee changes in service demand.
- Service Request and Incident Management: A key area where AI can be leveraged is automating the triage processes for service requests and incidents. Rather than relying on manual intervention, organisations can develop routines that automatically triage incoming requests and incidents. Additionally, AI can monitor system data for anomalies, a basic form of anomaly detection. Automation can be triggered when trained algorithms identify known patterns or deviations that resemble these patterns. Alternatively, staff can be alerted based on probabilities, reducing response times and improving efficiency.
- Problem Management: Problem management, often overshadowed by immediate incident resolution, can benefit from incorporating AI. AI streamlines this process by automating the identification of recurring incidents and subsequently categorising them as problems within the system. It can further search for root causes, assign probability rates, and find potential solutions within internal or external knowledge bases. This automated approach provides problem management teams with actionable insights, including the most impactful problems, their root causes, and suggested solutions. This streamlines decision-making and accelerates the resolution process.
These are just a few examples of how AI can be practically implemented in ITSM, enhancing the effectiveness of staff and improving service quality.
Conclusion
Adding AI in ITSM is a good marriage. It can keep repetitive work away from staff and raise the service level. Meanwhile, it can improve the communication between IT and other departments/units. With limited investment, significant gains can be achieved. It should be approached on a step-by-step basis with a straightforward architectural setup from the beginning to ensure alignment of solutions. Adopting AI in the processes will impact the operations and might even lead to resistance. A step-by-step approach can overcome resistance by achieving quick, noticeable results and improving staff well-being.
