Kessel Run Applies User Analytics, AI to Improve Applications

User analytics is key to helping Kessel Run inform their product decision-making processes.

According to Gregory Horvath, Kessel Run Product Manager for Enterprise AI, AI-driven user analytics helps Kessel Run gather data on user behaviors and demographics with the intent to tighten the feedback loop between product teams and end users.

Horvath said user data can then be leveraged to overcome cognitive bias in decision-making, get away from making subjective decisions and inform them with data.

There are three inflection areas where data is most commonly used at Kessel Run.

“We have to determine adoption. Is our software reaching target audience and is it being used? Usability is also used to identity bottlenecks and potential workarounds that user is taking not intended in our workflow,” Horvath said during the 2022 DOD Digital and AI Symposium Wednesday.

Security is also intrinsic to a good user experience, he added.

“You have to determine if users are relying on outdated software,” he said. “You also have to look at what can be done to get users to migrate to modern software and stop them from using anything that is vulnerable.”

Brian Austin, Kessel Run Tech Lead with Dragon, said Kessel Run relies heavily on UserAle, which is part of Apache Flagon, to handle user analytics.

“It’s a foundation that allows customizations that enable us to analyze activity and report it back as discreet pieces of information and trends,” Austin said at the DOD Digital & AI Symposium Wednesday. “It tracks DOM and HTML events. It’s also flexible enough to add pieces that would surface higher levels of understanding.”

UserAle provided some major benefits to Kessel Run including an easy, consistent implementation process.

“It’s able to scrape up all of the user events and collect them in a way that is useful for us to find quick pieces of information. It gives us an immediate idea of what is most important,” Austin said.

Kessel Run was also able to add friendly identifiers to individual elements as well as group activities.

“These trend lines may be familiar to many people with from the idea of cyber reliability engineering. We’ve tried to mimic that effort to focus on users, as user reliability engineering in what ways have users changed and how can we quickly identify those trends from user activity,” Austin said.

Kessel Run applied a back-end integration with its ID log-in system, which allowed them to understand who was logging in and who wasn’t.

Austin said UserAle provided a granular look at user activity, which enabled Kessel Run to analyze the activity and draw meaningful insights to improve user experience.

“We’re able to instrument visit events and additional custom logs. We’re able to build around points of time which allows examination of when people start and stop certain types of activities,” Austin said. “We built out time-based user activity patterns at the alias level as well as the page level using domain names or URLs, and what we got from that was the ability to see a set a patterns that helped us analyze and monitor the workflows that form the core of Kessel Run developer applications.”