The Telecommunications Risk Management Association (TRMA) Conference was one of our most exciting events of the year so far, mainly because we had so much to look forward to leading up to it! Our company recently partnered with Equifax in cultivating a more comprehensive collections solution, and the official announcement of our union was released the day preceding the conference in which we were presenting with one of their representatives, Eric Foulk. Needless to say, we had our hands full with preparations for both occasions.
Scott Hensley, our VP of Client Solutions, worked tirelessly with Eric to put forth the most meticulously planned and insightful presentation they could on End to End Consumer Journey Solutions. Custom visualizations and expert analyses elevated their presentation to the highest level, creating a lasting impression on all attendees at TRMA. We’re so proud of the amazing job these two gentlemen did, that we can’t help but share their success over and over again.
As exciting as the opportunity to present at this year’s conference was, it wasn’t the only worthwhile event we got to experience. Hilarious icebreaker challenges involving props and a photo booth (yes, we’ve included photos below!), revered keynote speakers like Mark Kelly, and valuable discussions over the latest developments in the collection industry made up the rest of TRMA, and our Gwynn Group representatives that got to experience these events firsthand can’t wait to tell you all about it…
On March 18–20, 2019 I had the distinct pleasure of attending my first ever TRMA Spring Conference. Being a veteran of the third-party side of collections for more than 20 years, I was very interested in attending this conference to see what the prevailing themes were in the industry from a different perspective. The event was held at the El Conquistador Resort in Tuscon, AZ. Being from Buffalo, the beautiful warm weather was much appreciated, and the resort itself was simply gorgeous.
As I listened to our VP of Client Solutions, Scott Hensley, co-present with Eric Foulk from Equifax there was a question from an attendee that I found interesting. The issue surrounded the problem of AI bias in dynamic modeling and what can be done to avoid this problem in a system that uses machine learning (AI) to segment and prioritize inventory treatment strategies. This got me thinking. What is “AI bias”? What causes it? Moreover, how can it be avoided?
Machine learning bias, also known as algorithm bias or AI bias, is a phenomenon that occurs when an algorithm produces results that are systematically prejudiced due to erroneous assumptions in the machine learning process. Predictions from an algorithm are used to drive decisions on account segmentation treatments fairly often in the collections process. When you have a tool such as our offering, Kenekt, that literally can build and update your ideal collection strategy based NOT just on a score that you buy upon placement but on the results that your efforts to date produce, you have a powerful tool to drive efficiency in your operation. When we have AI bias, however, the resulting prioritization model can be skewed to leave essential points out of the algorithm, potentially causing significant pockets of high-value assets to be missed in your process. So, how do we avoid this from happening?
The thing with any algorithm or segmentation strategy is that it will only work as well as the data that you feed into it. There is no magic bullet in the industry that will take bad, limited or missing data and be able to give you an effective collection strategy. Our Kenekt application can use an infinite number of data points to drive your dynamic collections process through AI. The most important thing in assuring that we do not run into AI bias is to make sure we include as many data points as possible, so long as those data points are relevant to the task at hand. These points can and likely will vary widely by the type of business being worked, delinquency level, balance size, and so forth. It is important to go through past data results at a granular level to fully understand the trends that present themselves, so we can intelligently and adequately determine which data points are most effective for use at your company, by business line. The pre-implementation discovery phase of the process is critical in doing this. A rush through without fully understanding each of the factors at play would be a big mistake.
As we stand today, new technology continues to reshape our industry. However, no matter how technologically advanced we become, sound and intelligent reasoning and smart analysis by people is required to ensure our shiny new tools work to our intended benefit.
Gwynn’s third time attending TRMA was a resounding success. The backdrop for the conference was the “El Conquistador” in Tucson, Arizona, a beautiful location for a conference, family vacation, or to get away from it all. The conference was very well attended by both active members and business affiliates, and Gwynn’s very own Scott Hensley presented with Eric Foulk from Equifax on Collections — Customer Journey End to End. The presentation generated excellent questions from the attendees, producing even further dialogue during the Mexican reception later that evening.
Although our collaborative presentation with Equifax was one of the most memorable moments of TRMA, the conference kicked-off with Tucson’s very own Mark Kelly as the keynote speaker. An amazing and accomplished astronaut, Mark spoke about overcoming challenges, setting goals, and always professionally questioning superiors. Mark was very adamant and open about his rule of questioning decisions if another solution or idea works better. He stated that when preparing for a mission, his opening discussions with his team include that his decisions need to be challenged — this attribute of a highly accomplished leader resonated with me to my core. A great leader brings out the best in their people and motivates them to be the best person/employee they can be. Like the folks on Mark’s team, the employees of the Gwynn Group also have a true leader, Hugh Gwynn. Although our industries differ, Hugh demonstrates the same qualities of which Mark spoke.
Hearing from a few expert speakers throughout the conference wasn’t the only thing happening at TRMA. We were also able to witness presentations over relevant topics within the collection industry. “How carriers are answering the call” was a fascinating presentation about fraudsters and how companies are utilizing voice biometrics to offset the losses that are compounding. Due to large volumes of calls, the contact center has become extremely vulnerable to fraudsters around the world. Once a voice has been recorded, conversational biometrics and physical characteristics become prevalent to apprehending the fraudsters. Conversational biometrics is a combination of vocabulary used, grammar, and the sentence structure used by the fraudsters in the calls. Physical characteristics are all unique to each human being, very similar to a fingerprint. Physical characteristics include vocal cords, the space in your throat, and the formulation and position of your teeth. Fraudsters will always try to stay one step ahead of technology, but they can’t change the physical characteristics they were blessed with.
All in all, we can say with certainty that this year’s TRMA conference was an exciting event for Gwynn Group. The location was fantastic, the speakers were inspirational, and the presentations were well-informed and insightful. The events of the conference were only topped by the opportunities we had to network with our peers and pick the brains of industry experts about the landscape and direction of the collection industry. The best part? We get to go back and do it all over again next year!