Using AI to Optimize Licensed Product Use For 25,000 Team Members

Today’s modern businesses use a whole host of online tools to help them achieve their goals. Whether it’s a design tool, CRM platform, or project management software, these licensed products are essential to most companies’ functioning.

However, not all employees use these tools equally. While some may use a certain platform every day, others may need it only once or twice a year. This results in the costs spent on product licenses going to waste when they go unused, which can amount to hundreds of thousands of dollars per year for large companies.

In fact, spending on software assets is growing yearly, with almost a third of that spending going to waste, according to the 2022 State of IT Asset Management Report.

intive recently partnered with a leading gaming company to help it tackle this problem of wasted licensing costs. Here’s how we did it.

Wasted Spending on Unused Licenses

Our client is a global leader in the gaming industry with over 25,000 employees and subcontractors helping to build their world-class games. With such a large employee base, the client faces extremely high licensing costs for over 500 tools – from Adobe Suite to Jira, to Miro and Workday. While these tools are essential and valuable to those who need them, many licensed products were going unused resulting in unnecessary costs.

Simply taking every user off every licensed product until they needed access again would create a huge disruption to workflows and productivity. We partnered with the company on a Proof of Concept (PoC) which used AI/ML to examine disparate data sets to predict which employees need the licensed products being provided to them.

Leveraging AI to Predict Licensing Needs

To accurately predict which licenses are needed, we would need to put a large data set, including 2.2 million logins/month to licensed tools and a host of metadata, through the Extraction, Transformation, and Loading (ETL) processes and properly sanitize the data for our pipelines.

After extensive feature engineering on high data volume, we set a baseline with a simple Machine Learning algorithm, and in the next step, we improved our results by using a more sophisticated ML model. Finally, we deployed the model as an endpoint on AWS.

Eight Weeks and $300,000 Saved – And This Is Just the Start

In just eight weeks, intive was able to deliver the PoC.

The PoC is fully functioning and has already been used to complete the analysis and re-negotiation of a single tool, which on its own will produce annual cost savings of over $300,000. We plan to continue the partnership and advance the PoC into a complete product with UX design and data visualization which will be used to analyze the most common and costly licensed technologies used by our client.

Services We Used

Solution Design, Data Engineering, EDA (Exploratory Data Analysis), Data Science, Machine Learning, and integration into AWS environment.

Our Tech Stack

Python, scikit-learn, PyTorch, MLFlow, Optuna, Docker, AWS (Glue, Sagemaker Notebooks, Lambda, API Gateway)

Do you need support in building AI-powered optimizations that save you time and resources? Get in touch today and find out how we can help.


¿Cómo podemos ayudarte?
Hablemos.

Contactanos

¿Te ves trabajando acá? ¡Genial!

Unite al equipo