The burning question on everyone’s mind is: Could AI be a game-changer for ITAM? The short answer is YES!

Since the 1960s, AI and machine learning (ML) have been pivotal in integrating data to derive meaningful insights. However, progress was limited due to computational constraints and the absence of large datasets.

In the following decades the technology experienced a resurgence with the advent of more powerful computers and the accumulation of substantial datasets. The 2000s witnessed further advancements, particularly in deep learning—a subfield of ML focused on neural networks with many layers.

Overall, the history of AI reflects a journey marked by continuous innovation, fueled by advancements in technology, data availability, and evolving societal needs and expectations. For any data heavy industry such as ITAM, it’s just going to be massive.

Before I dive deeper in some use cases, I’d like to shed some light on the terminology. Artificial Intelligence and Machine Learning are often used interchangeably, but they refer to different concepts and scopes within the field of computer science.

The goal of AI is to simulate natural intelligence to solve complex problems. It can enable machines to perform cognitive functions like those performed by humans. Machine Learning is a subset of AI focused specifically on developing algorithms and statistical models that allow computers to perform specific tasks without using explicit instructions. ML’s goal is more specific—improving the performance of a model on a task over time through learning from data.

AI applications can be as simple as a rule-based system for automated responses in a chatbot or as complex as an autonomous driving system that uses a combination of ML and other techniques to make decisions.

ML applications are typically focused on specific tasks like image recognition, language translation, or stock market prediction, where patterns in historical data are used to predict future outcomes or trends. Machine Learning relies on feeding large amounts of data into algorithms, allowing them to learn from the data’s patterns and characteristics. The more data these systems are exposed to, the better they can perform.

In the context of ITAM I believe that Machine Learning is the more interesting piece of AI, while the (generative) AI use case is more about something like typing in a message in a chatbot and getting something back that’s interesting and saves time. Think about a typical SaaS Management question such as “How many licenses do I own for Microsoft Office” and it’ll go, it’ll search, and it’ll say you have this many licenses. It’s a cool function that makes our lives easier but the overall impact on business is smaller. Strong use cases in ITAM are more around transparency, optimization, risk reduction and workflows.

So, when we talk about inventory management, ML can automate the process of discovery and categorization of all the SAM and HAM assets. It can ensure that the organization is up to date and has comprehensive IT visibility without somebody constantly happening to look at the data.

Imagine a big company having a very large data set of information and the ITAM tool must identify all the software assets. Machine Learning is going to be able to do that much faster than a human can. IT asset identification, normalization and classification are going to be huge ML algorithms because as the models can analyze data, they’re then able to improve accuracy.

That’s why we developed the EULA Recognition tool which provides a normalized summary of the relevant content like license metrics. Customers can do a comparison of different EULA versions and outline the relevant changes or can receive suggestions for required new catalog content.

We can start using ML for predictive analytics so it can cover the forecasting for future IT asset needs based on historical usage patterns, business growth projections, any technology trends, and it can anticipate demand on hardware, software, and cloud resources.

ML brings significant benefits to IT asset management by automating workflows and repetitive tasks, improving accuracy, providing actionable insights. It can also be used for enhancing security and compliance as well as optimizing resource utilization and costs.

One great example is our Invoice Recognition tool. Customers can upload invoices to create licenses within USU License Management to get rid of a rather manual process. New license invoices are sent to the Identification and Reconciliation Engine (IRE), automatically read in, the essential information such as manufacturer, product, price, quantity, etc. is recognized and prepared for import into the corresponding fields of the LIMA platform.

AI-driven optimization algorithms can analyze data on asset usage, performance, and costs to identify opportunities for consolidation, virtualization, turning machines on or off by right-sizing that IT infrastructure or eliminating unused or redundant assets. So, organizations can really start reducing operational expenses, improving resource efficiency and do IT optimization with just a few clicks within seconds.

Another great use case is IT security by identifying security vulnerabilities. Imaging a machine taught process that can scan through every piece of software and then identify the vulnerabilities based on another data set up, identifying which ones are in your systems, or which ones are being deployed.

Let’s consider a piece of software with a vulnerability score of seven and then someone must look at that and identify which software can remain or poses a risk. IT risks will become much less because the time of exposure becomes less so from a security standpoint, machine learning can help there effectively.

ML/AI offers many promising possibilities, but there will also be issues with the trustworthiness of data. AI can be trusted as a useful technology within your organization if used correctly.

There is the importance of “controlling the AI” by monitoring and checking data. We should not become too reliant on AI’s smart guidance and results, so you need to control the AI to control the results. That’s why we at USU are building our own AI models under a controlled environment. For IT asset management, the way we’re building it, the way we’re deploying it, the way we’re using it is also very good.

The key is to find specific task for AI. To improve AI’s reliability, it’s vital to train the AI for specific domains, such as SaaS Management, Cloud FinOps or Knowledge Management. This will involve contextualizing its capabilities within the boundaries of the training data the AI has been training on.

With a human in the loop and staying within the safeguarding boundaries established, AI can provide reliable answers, depending on how far you push AI as we are still in the early stages.

Get involved, get interested. Spend time relearning your AI interactions skills (we mastered asking Google questions to find answers) and establish an AI training program. Experiment and learn how to adapt AI to your ITAM processes. Determine how AI can best serve your organization and identify uses cases that can drive your digital transformation and business acceleration.

Treat AI like a tool. It’s not a replacement for strategies or people. AI doesn’t replace jobs. It’s a skill. People will still need to drive and manage AI, so it delivers the required business outcomes.

Don’t rush the process: While AI promises transformative potential, organizations must navigate its challenges and be aware about how to best use it. Be proactive and prioritize strategies that maximize business value and recognize that AI represents an ongoing learning journey rather than an immediate solution.

This is the time to embrace the possibilities that AI can bring for ITAM, even if uncertainties exist.