Technology in Collection: 21 Must-Know Buzzwords

By on May 24th, 2022 in Industry Insights, Machine Learning, Product and Technology

Your Guide to Key Terms for Today’s Debt Recovery Strategy

Reaching consumers today requires a more sophisticated process than simply dialing the phone or sending a generic email, especially when it comes to debt recovery and collection. But reviewing potential strategies can often leave you lost in a sea full of acronyms and buzzwords. Between terms like AI, machine learning, and data science, it can be difficult to keep up with the different definitions—and understand how they impact your business and bottom line.

To help keep this word salad straight, we’ve compiled a glossary of helpful terms, definitions, and examples to help differentiate them:

  • Accounts per employee (APE), account to collector ratio (ACR): The number of delinquent accounts that can be serviced by an individual recovery agent – often used to measure cost effectiveness.
  • Artificial Intelligence (AI): AI is a blanket term describing a range of computer science capabilities designed to perform tasks typically associated with human beings. Machine learning (ML) is a subset of AI. Through AI, processes like debt collection can become more efficient by developing better outreach and deployment strategies.
  • Big Data: This term means larger, more complex data sets . Big data can save collectors a lot of time by using many variables for analytics-based customer segmentation, insert, insert..
  • Coverage: The percentage of users for whom organizations have digital contact information, such as email addresses or phone numbers.
  • Customer Retention Rate: Measures the total number of customers that a company keeps over time. It’s usually a percentage of a company’s current customers and their loyalty over that time frame.
  • Data Science: ‍A cross-discipline combination of computer science, statistics, modeling, and AI that focuses on utilizing as much as it can from data-rich environments. Data science (which includes machine learning and AI) requires massive amounts of data from various sources (customer features such as debt information or engagement activity) in order to build the models to make intelligent business decisions.
  • Deep Learning (DL): A subset of machine learning. Deep learning controls many AI applications and services and improves automation, performing analytical tasks with human intervention.
  • Delinquency rate: The total dollars that are in delinquency (starting as soon as a borrower misses as a payment on a loan) as a percentage of total outstanding loans.
  • Deliverability: The percentage of digital messages that are actually reaching consumers (e.g., as opposed to ending up in email spam filters).
  • Digital engagement metrics: A range of KPIs that capture how effectively digital channels are reaching and engaging consumers.
  • Digital opt-in: The percentage of users who have indicated their preference to receive digital communications in a particular channel.
  • Efficiency: Measures a company’s ability to use its resources efficiently. These metrics or ratios are at times viewed as measures of management effectiveness.
  • Machine Learning (ML): Technology that uses algorithmic modeling techniques to observe patterns and trends, reassessing the best approach to achieve a goal, and adapting behavior accordingly. It continuously, automatically learns and improves at a massive scale as more data is observed. With the help of machine learning, companies can make sense of all their data and take on new approaches to debt collection processes from better customer experience to more efficient delinquent fund recovery.
  • Net loss rate: The total percentage of loan dollars that get charged off (written off as a loss).
  • Open rate, clickthrough rate: The percentage of users who are actually opening and clicking digital communications.
  • Predictive Analytics: Predicting outcomes is one specific application of machine learning. It allows companies to predict which accounts are more likely to pay sooner and allows them to better plan operations accordingly.
  • Promise to pay kept rate: The percentage of delinquent accounts that maintain a stated commitment to pay.
  • Promise to pay rate: The percentage of delinquent accounts that make a verbal or digital commitment to pay.
  • Right party contact rate: The rate at which a collections team is able to establish contact with the consumer associated with a delinquent account.
  • Roll rate: The percentage of delinquent dollars that “roll” from one delinquency bucket (e.g., 60 days past due) to the next (e.g., 90 days past due) over a given timeframe.
  • SMS: An acronym that stands for “Short Message Service” referring to text messages on cellular devices.

For more information on how to get started integrating innovative technologies into your debt recovery strategy, schedule a consultation today.

TrueAccord Announces Results Confirming Effectiveness of Digital-First Retain Product for Early-Stage Delinquencies

By on January 25th, 2022 in Company News, Machine Learning, Product and Technology

With more than 1 million consumer accounts now managed through the intelligent, client-branded product, results show 40% more effective than leading “call and collect” vendors

LENEXA, Kan., Jan. 25, 2022 — TrueAccord Corporation, a debt collection company offering machine learning-powered digital recovery solutions, today announced results following the September 2021 rollout of Retain, the client-branded product that addresses early-stage recovery challenges for organizations with customers with delinquent accounts. TrueAccord Retain is now being used by creditors to manage more than 1 million consumer accounts and has shown to be 40 percent more effective at repayment than traditional “call and collect” debt collection vendors. 

TrueAccord Retain, which harnesses digital technology and machine learning to deliver a personalized, effective early-stage recovery strategy, significantly outperformed three traditional “call and collect” agencies across several of an anonymous client’s portfolios. Relative to the best-performing “call and collect” vendor for each product portfolio, TrueAccord Retain drove a 24 percent improvement in roll rate, a 28 percent improvement in early-stage gross flow through rate and a 40 percent improvement in late-stage gross flow through rate*.  

“With more than 1 million consumer accounts now being managed through Retain, we’re able to see the robust results of the product on improving early-stage delinquencies for our clients,” said Mark Ravanesi, CEO of TrueAccord Corp. “The results of our client’s evaluation were unambiguous: Retain’s machine learning-powered, digital-first approach resonated with consumers and drove significant growth for the early-stage recovery business. With a lingering worker shortage, especially in the call center space, we expect these performance numbers to continue to grow as more consumers are brought into the Retain ecosystem in 2022.”

Powered by TrueAccord’s industry-leading tech stack, key benefits of Retain include a simple, intuitive and effortless-to-use digital platform leading to great user experience, constant A/B testing and optimization to reduce friction and boost conversion rate, infinite scalability, and second-to-none channel deliverability. Retain implements e-commerce-based innovations like the focus on digital experience and outreach, machine learning-based personalization, and deliverability at massive scale for early-stage use. 

To learn more about TrueAccord and its digital-first recovery solutions, visit www.TrueAccord.com and follow @TrueAccord on Twitter and LinkedIn.

*This data comes from an anonymous client’s evaluation of performance of different delinquency  approaches  side-by-side. The client randomly assigned credit and retail card accounts to TrueAccord Retain and the other vendors. Key success metrics included roll rate, or the percentage of dollars that became progressively delinquent, and gross flow through rate, or the percentage of dollars that flowed from one delinquency category across multiple subsequent categories.

About TrueAccord

TrueAccord is the intelligent, digital-first collection and recovery company that leaders across industries trust to drive breakthrough results while delivering a superior consumer experience. TrueAccord pioneered the industry’s only adaptive intelligence: a patented machine learning engine, powered by engagement data from over 16 million consumer journeys, that dynamically personalizes every facet of the consumer experience – from channel to message to plan type and more – in real-time. Combined with code-based compliance and a self-serve digital experience, TrueAccord delivers liquidation and recovery rates 50-80% higher than industry benchmarks. The TrueAccord product suite includes Retain, an early-stage recovery solution, and Recover, a full-service debt collection platform.

How TrueAccord Embraces Machine Learning to Create Positive Consumer Experiences in Debt Collections

By on December 23rd, 2021 in Industry Insights, Machine Learning, Product and Technology
TrueAccord Blog

By Laura Marino

TrueAccord’s Chief Product Officer, Laura Marino, was recently featured in the New Standard in Debt Collection panel as part of the Beyond Digital: The Next Era in Collections summit. As a civil engineer turned product management executive, Laura has a unique viewpoint on the evolution of machine learning in software across a variety of industries. In this blog post, Laura shares her perspective on machine learning at TrueAccord and in collections, in general.

At TrueAccord, we know that consumers prefer digital channels and self-service. We also know that just providing the digital channels is not enough. To truly engage with consumers we need to help them throughout the journey. This is where machine learning comes in.

What is machine learning?

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. In the context of collections, and specifically in the context of our consumer-centric approach to collections, machine learning is a wonderful tool to personalize the experience for each consumer, effectively engage with each of them, and ultimately help resolve their debt.

There has been so much hype around machine learning, but often companies that claim to do ML are really using fixed rules or heuristics (if a consumer does X, then do Y) without including any of the automatic learning and improvement. Or they may be using ML for a very specific, very limited scope – like automating some consumer support responses. The reason that leveraging ML is so difficult for something as complex as collections and recovery is that it requires a lot of expertise in data science and behavioral science, it requires a lot of user research, and it requires a lot of data.  This is not something that a company can decide to start doing overnight as an add-on.

How does TrueAccord apply machine learning to debt collection?

TrueAccord is leveraging machine learning and behavioral science throughout the entire journey, from initial engagement all the way to resolution. We were built specifically around the hypothesis that focusing on machine learning-driven, digital-first experiences was the way to transform debt collections. We have been doing this since 2013, and we have orders of magnitude more data than anyone else. Just to give you an idea: we send millions of emails per day, and hundreds of thousands of text messages per week and our ML engine learns from every open, every click, every action on our website, and every interaction with our call center agents. Because of all of this, we have something that is very hard for anyone to imitate.

Unlike traditional collections, we do not use demographic data like age, zip code, or creditworthiness to personalize the experience. Instead, we use engagement data about how the consumer responds at every step in the process.  

We have handled debts for over 24 million consumers and we have collected data about each individual interaction with those consumers. That wealth of data, combined with our ongoing user research is behind the ability of Heartbeat (our fully automated and reactive decision engine) to personalize the experience for each consumer.  We’ve seen this data-driven machine learning customer-centric approach lead to increased customer satisfaction, better repayment rates, and lower complaint rates.

Machine learning is used to personalize and optimize every step of the customer journey. The first thing we need to do is to effectively engage with the consumer.  For that we have several models: 

  • Cadence optimizer: determines the right cadence to communicate with each consumer about their debt. Specifically, it determines which day to send the next communication. We don’t have a fixed rule that says “send an email every x days.” Our decision engine decides it dynamically based on the type of debt, the consumer behavior, and where they are in the process. 
  • Send time optimizer: determines when during that day, communication should go out. A working mother who is busy with her kids in the morning and in the evening is more likely to check her messages in the middle of the day during her lunch break. A construction worker has a very early start to their day, may prefer to check messages at the end of the day.  We want our consumers to receive our communications during their preferred times so that they are at the top of their inbox and not buried under 50 other emails. Reaching people at the right time of day has a big impact. Due to our send time optimizer, we saw a 23% increase in liquidation for certain types of debts. 
  • Email content rater: we also want to make sure that the tone of our communication is one that will best resonate with a specific consumer. For each piece of content we send out, our content team has created multiple versions with different voices, ranging from very empathetic to more ‘to the point’ because different people respond to different styles. Heartbeat chooses which one to send based on what it has learned from the behavior of each consumer. 

After engaging the consumer with the right cadence, timing, and content we want to make sure that they commit to a payment plan and stick to it until their debt is resolved. For that, we have machine learning models that determine the best combination of discount and length of payment plans to offer to each consumer. The options that the consumer sees when they get to the payment plan page are tailored to them based on what Heartbeat believes will work best. The consumers can build their own plan but, if we can proactively offer options that work, we make it easier.

We also have a ‘payment plan breakage model’ that helps us identify consumers who are at risk of not making a payment so that we can proactively reach out to them and give them options. With this we were able to increase the resolution rate among customers at risk by 35%.

What do customers think about TrueAccord’s model?

We have a lot of very positive feedback from our consumers which I attribute very much to our machine learning capabilities. It is one of the things that I think is so exciting for everybody who works at TrueAccord. We consistently get messages saying, “Thank you for making it so easy. Thank you for allowing me to do it via digital channels without having to talk to anybody.” And then when people call with questions, our call center knows that they’re there to help. People definitely respond very positively to the approach we’re taking to collections.

This content originally appeared as part of the Beyond Digital: The Next Era in Collections summit. Watch the entire summit here

Building a Digital-First, Third-Party Collections Solution with Snap Finance

By on August 13th, 2021 in Industry Insights, Machine Learning, Product and Technology

In 2019, Todd Johnsen, Snap Finance’s Senior Manager of Collections Vendors, was charged with doing something that had never been done at Snap before: developing a third-party collections program. According to Johnsen, “At that time, the only recovery program for charge-off accounts was a call-and-collect settlement at tax season. I knew we could do better, but we’d have to start from scratch.” Johnsen also found a large amount of backlog accounts that had never been worked by a collections agency, as well as the need for a forward flow third-party recovery program.

Johnsen and team surveyed their options: they looked at both traditional agencies (predominantly making outbound calls) and digital-first collections solution providers, like TrueAccord. Johnsen was particularly interested in how digital-first providers like TrueAccord used machine learning to optimize their relationships with consumers via digital channels like email, SMS, and push notifications. 

“My thought process was — we work with subprime consumers who may have bad associations with debt collection,” said Johnsen. “This audience may have already had experiences with incessant collection phone calls, and they are used to avoiding them. I wanted to find an agency that was doing things differently. I knew that TrueAccord was using technology and digital channels in a way that other providers weren’t.”

While Johnsen was curious about working with a digital-first agency like TrueAccord, he wasn’t ready to go all-in immediately. The Snap Finance team decided to engage both a traditional agency and TrueAccord and compare the results. In evaluating the competing partners, key considerations included liquidation rate performance, security and compliance, and optimization efficiency. The result? TrueAccord delivered better results across all measured parameters.

“The reality of the results really knocked me out,” said Johnsen. “What we saw was almost 25-35% better performance on the accounts that we placed with TrueAccord, compared to the accounts we placed with traditional agencies. It was a real eye-opener. In fact, TrueAccord is number one in every tier I have them in. We’ve seen nothing but huge benefits as a result of that individual, digital-first interaction that TrueAccord tailors to each consumer.”

To learn more about TrueAccord’s work with Snap Finance, read the full case study.

Beyond Coding: Using AI to Improve the Healthcare Revenue Cycle

By on July 8th, 2021 in Industry Insights, Machine Learning
TrueAccord Blog

Generally, when talking about artificial intelligence (AI) in regards to medical collections, we hear about how it has automated the once-painstaking process of medical coding for billing. But why stop there? With all of its capabilities, AI has much more impressive and patient-facing applications when used to improve customer experience, especially in the healthcare industry which is increasingly digital-first and self-serve. In this post, we’ll explore how AI and machine learning can supercharge the healthcare revenue cycle by catering to consumer preferences, turning billing and collections into a seamless, efficient experience for both patients and providers.

But first: why is it necessary—and even urgent—to improve healthcare revenue management? The answer is patient expectations. Patients now expect the same type of personalized, easy-to-use experience they’ve grown accustomed to receiving from other industries, including banking, airline and retail industries. Patients are now “digital-first” and look for an end-to-end experience that allows them to handle medical-related issues on their own, often from their mobile devices. Patients can already schedule appointments, request prescription refills, receive test results, and even contact their healthcare provider directly through digital platforms. The application of digitization through AI and machine learning to other touchpoints in the patient journey, all the way through billing and collections, can improve customer experience and thereby their overall interactions and relationships with their healthcare providers.

First, digitization powered by AI and machine learning can replace manual and paper processes to speed up the recovery timeline. A 2020 report by InstaMed, a J.P. Morgan company, found that patient collections take more than a month for 63% of healthcare providers. This figure isn’t surprising when 81% of providers still leverage paper and manual processes for collections, while 75% of consumers want to receive eStatements for medical bills. The traditional method of collections does not align with consumer preferences, with more than half (54%) of consumers surveyed saying they prefer electronic communications (emails, text messages, in-app messages and live chats) for medical bills. And a majority of consumers (65%) preferred paying those medical bills digitally as well – whether online through their doctor’s or health plan’s website, their bank’s bill-pay portal or mobile apps – instead of manually. Using AI and machine learning to match the consumer’s communication and payment preferences can drastically improve the time needed to engage and collect from patients.

Second, AI-powered systems can personalize the billing and collections process and offer intuitive payment solutions for patients to achieve the best possible recovery rates. According to the InstaMed report, collecting patient financial responsibility in a timely manner was especially challenging for large patient balances, with 49% of surveyed providers reporting that they cannot collect bills of more than $400 in 30 days. Especially with multiple billers on different payment cycles, it can be difficult for a patient to set up a payment plan with terms they can successfully meet. AI can improve this experience by identifying the most efficient time, place and manner to communicate with a patient about their financial responsibility and go a step further in presenting personalized, affordable payment options. 

Third, AI can be used to interface directly with clients where they are and minimize the need for waiting on hold for the next available representative, creating a more seamless, humane process and a better customer experience. AI-enabled chatbots can answer basic questions, while automation can help provide information on why claims were denied and other status updates. Empathetic customer service is important in the healthcare industry and customized customer self-service can reduce frustration for the patient and the number of service agents needed for the provider.

At TrueAccord, we use AI and machine learning to build digital debt collection solutions for billers that put customers first. By implementing behavioral analytics to predict consumer communication preferences and machine learning to create smart, intuitive processes that increase likelihood of patient repayment, TrueAccord products stay a step ahead to ensure a successful revenue cycle where both patients and providers win. To safeguard personal patient information, TrueAccord’s policies and procedures are designed to comply with all HIPAA-related requirements (Health Insurance Portability and Accountability Act), including documenting the use of protected health information (PHI) and the physical, technical, and administrative safeguards implemented to protect PHI. Learn more about how we use AI and machine learning to provide a personalized collections experience at scale here.

Pitfalls on the Path to Digital Debt Collection

By on August 25th, 2020 in Industry Insights, Machine Learning, Product and Technology, User Experience

Banks are accelerating their adoption of new digital debt collection tools in anticipation of a “tidal wave of consumer debt issues” when government stimulus programs end and financial institutions stop offering forbearance and loan deferral options.

That’s the premise of a new article in American Banker highlighting a variety of technology-powered strategies banks are using to make debt resolution more automated, conversational, and empathetic. These approaches range from the convenient (more flexible self-service payment options) to the high-tech (robotic process automation). 

The American Banker article highlights promising signs of progress, particularly for industry players that have not always been known for digital adoption. KeyBank, for example, is in the process of rolling out a self-service digital payment portal designed to offer banking customers privacy and flexibility in resolving payments. And Alabama-based Regions is implementing digital messaging and intelligent interactive voice response (IVR).

At the same time, the article shines a light on the massive challenges facing any financial institution looking to implement intelligent digital debt collection at scale. Here are three common hurdles on the path to digital debt collection maturity – and why they matter:

Challenge #1: “One-size-fits-all” approaches

The challenge: In its overview of Regions, the article makes reference to a single conciliatory messaging tone used in all outreach to delinquent customers. 

Why it matters: Consumers differ vastly in their preferences and responsiveness to digital touchpoints. For example, one consumer might respond to a friendly message delivered by SMS, while another might respond best to a straightforward message delivered by email. As a result, a one-size fits all approach falls short of realizing the potential – in both performance uplift and customer experience – of true one-to-one personalization.  

The TrueAccord approach: HeartBeat, TrueAccord’s patented machine learning platform, mines through tens of millions of data points to optimize digital outreach on the individual level within a programmed set of compliance rules  – and continues learning the more data it analyzes.  

Challenge #2: Narrow, channel-specific use of machine learning 

The challenge: Another challenge that banks face in scaling their use of intelligence – including artificial intelligence (AI) – is the limited deployment of algorithms and optimization within a single kind of channel, such as in a call center environment. The article profiles a collections and business process outsourcing company, for example, that developed an AI-based virtual assistant that can handle most inbound phone calls.

Why it matters: Machine learning and artificial intelligence (AI) are powerful tools for restoring intimacy and relevance to customer relationships at scale. At their most useful, these tools should be deployed to personalize the customer’s full experience with a bank – not just the limited interaction on one channel. 

TrueAccord’s Approach: HeartBeat captures a continuous data feedback loop and optimizes for each customer touchpoint across a variety of digital channels, ensuring that each customer is being reached on the channel that is most relevant for her.  

Challenge #3: Building a truly comprehensive and flexible self-serve portal

The challenge: Constructing a digital portal that drives consumer adoption and usage takes major work. To truly match the convenience of online banking, digital tools must also allow consumers to adjust the length and installment amount on a payment plan, defer a payment, dispute all or a portion of their debt, apply for a hardship pause on their debt, and much more.

Why it matters: Research suggests that customers want to be able to self-serve. But doing so requires the full, flexible range of interaction options that would be available to them through traditional analog channels.

TrueAccord’s Approach: Through a robust and flexible digital platform, TrueAccord offers  a best-in-class self-serve experience: over 95% of users resolve their accounts without ever directly communicating with an agent. 

Ultimately, digital debt collection technologies offer banks the ability to build lasting relationships with their customers. As Kimberly Snipes, consumer chief information officer at KeyBank puts it in the American Banker article: “We want our customers to say, I hate that I had that situation, but I felt like my bank was working with me, not against me.”

Being aware of the challenges on the path to digital debt collection – and having a plan in place to address them proactively – can help financial institutions ensure that they’re set up for long-term success. 

About TrueAccord

TrueAccord is reinventing the relationship between creditors and lenders with a machine learning-driven, digital approach to debt collection. Our technology personalizes outreach to each customer across digital channels, continuously optimizing for performance while delivering a customer experience that builds long-term brand loyalty. Schedule a demo today to learn more. 

How is machine learning driven by experimentation?

By on March 6th, 2020 in Machine Learning, Product and Technology
microscopes in a laboratory

Building scalable technology requires constant evaluation and improvement. Experimenting is defined by trying new things and creating effective changes that help teams to make informed decisions around product development. Trying new things creates momentum, and organizations that are driven by experimentation turn that momentum into growth.

Machine learning and artificial intelligence support large-scale, concurrent experimentation that helps these technologies to improve upon themselves. With the right tools in place, you can test a variety of scenarios simultaneously.

For example, we use our systems to track changes in the collection process and better understand how our digital collections efforts can be improved. Since digital-first channels offer thorough tracking and analysis, including real-time tracking on our website, we can learn in short cycles and continuously improve our product. 

This kind of frequent experimentation helps to avoid making product development decisions based on untested hunches. Instead, you can test your instincts, measure them carefully, and invest energy where it matters.

Machine learning drives the experimentation engine

Aggregating historical data and processing it using machine learning algorithms and artificial intelligence helps you to understand their effectiveness. Regardless of how intelligent your learning algorithms may be, waiting to test and expand your knowledge base before marching blindly ahead can make or break the success of your product.

To launch an experiment, we follow these steps: 

  1. Start with a hypothesis that you want to test
  2. Assign a dedicated team to manage the experiment
  3. Monitor the performance of the test as it is guided by machine learning
  4. Iterate

B2B companies can benefit from partnering directly with clients to customize experiments for their unique product lines in order to make experimentation-based optimization an ongoing process for both new and existing business. Keep in mind that the goal of product optimization is not always jumping to the finish line. 

Understanding how your product works ultimately offers you and your customers more value, but it’s easy to become distracted by positive outcomes. Effective, scalable products require intentional design; if you’ve accomplished a goal, but the path there was accidental, taking a few steps back to review that progress and test it can help you to get a clearer picture and grow the way you want. 

Below are two sample experiments we conducted to optimize our machine learning algorithms. 

Experiment #1: Aligning Payments to Income

Issue

The number one reason payment plans fail is consumers don’t have enough money on their card or in their bank account. 

Hypothesis

If you align debt payments with paydays, consumers are more likely to have funds available, and payment plan breakage is reduced. 

Experiment

We tested three scenarios: a control, one where we defaulted to payments on Fridays, and one where consumers used a date-picker to align with their payments with their payday. After testing and analysis, we determined that the date-picker approach was the most effective as measured by decreased payment plan breakage without negatively impacting conversion rates.

By understanding which payment plan system was the most effective, we were able to provide our AI content that offered these plans as options to more consumers and integrate the knowledge back into our systems and track those improvements at a larger scale!

Experiment #2: Longer payment plans can re-engage consumers

Issue

Customers dropped off their payment plans and stopped replying to our communications.

Hypothesis

Customers can be enticed to sign up for a new plan if offered longer payment plan terms. 

Experiment

We identified a select group of non-responsive consumers that had broken from their payment plans and sent them additional text messages and emails. These additional messages offered longer payment plan terms than the plans they broke off from.

Ultimately, we found that offering longer payment plans, even with reference to the consumer’s specific life situations didn’t lead to an increase in sign-ups. The offers that we sent had high open and click rates but did not convert. This indicated that we were on the right track but needed to iterate and come up with another hypothesis to test.

This experiment was especially important because it illustrates that not every hypothesis is proven to be correct, and that’s okay! Experimentation processes take time, and the more information you can gather, the better your results will be in the future.

We’re able to simultaneously update our product and continue experimenting, thanks to algorithms called contextual or multi-armed bandits. Here’s what you need to know about these algorithms and how they help!

Building the newest, most innovative products feels exciting, but building without carefully determined direction can be reckless and dangerous. By regularly evaluating the effectiveness of machine learning algorithms, you can make conscious updates that lead to scalable change, and experimentation paves the way for consistent product improvement.

5 ways debt collection uses machine learning and artificial intelligence

By on February 28th, 2020 in Machine Learning

Machine learning algorithms are playing a key role in the collections industry’s technological growth. Companies are working to integrate artificial intelligence and machine learning into their strategies in response to changing regulations and evolving consumer preferences. These processes can look dramatically different from business to business!

Some technologies are being applied to optimize traditional call and collect strategies while others are building digital-first outreach platforms. Understanding how these algorithms are working for the industry can provide insight into the future of collections. 

Business intelligence and analytics

Business intelligence platforms are the foundation for the future of collections. They not only help companies understand how to best reach their existing accounts using traditional collections strategies but also integrate into other digital tools to create powerful automated systems. 

These algorithms process large sets of data such as call times, call effectiveness, the value of certain accounts, collections rates, and many other variables. By analyzing this information, teams can optimize their outreach strategies by focusing on accounts that are more likely to be collected on, understand what times of day or channels work the best, and even determine what language to use in conversation with specific subsets of accounts. 

Portfolio evaluation and exchange

By adding a clear scoring system to business analytics tools, teams can share their portfolios in an online marketplace with other creditors and debt buyers in order to buy, sell, and even outsource debts as needed.

While debt marketplaces are not new, real-time scoring updates and activity insights provide a dynamic, cloud-based view into a fluctuating market. 

Human-like contact center agents

As companies evaluate their data and optimize their outreach, they can also integrate digital agents to interact with consumers over the phone. Artificial intelligence software can be used to create human-like voices and personalized experiences for consumers.

These platforms can operate at scale more easily than sprawling call centers but still rely on a traditional call and collect model that consumers are shying away from. As consumer preferences shift toward digital channels, more machine learning tools can help to optimize for an omnichannel experience.  

Digital collections platforms

Digital collections software is able to optimize performance data and leverage it using a diverse, multi-channel communication approach. Phone calls may be included as part of a larger strategy, but these platforms are primarily built around modern consumer channels including email, SMS, push notifications, and direct drop voicemails.

Contextual bandit algorithms take channel selection to a level beyond traditional A/B testing. Even if 10% of your consumers prefer one message type to another, it’s important to understand all of your audience’s preferences.

Digital channels integrate seamlessly with decision making algorithms and can optimize communications in ways that call systems cannot. For example, digital channels like email can reach consumers outside of hours typically limited by the TCPA. 

25% of TrueAccord’s consumers access their accounts outside of the 9am to 9pm when traditional agencies cannot legally reach them.

Digital debt collection agencies

Each of these implementations of machine learning help to build a more personalized, more focused, and more forward thinking debt collecting experience for both consumers and creditors. One consistent factor that does limit their effectiveness is the need to build them into existing systems or alter processes at scale. 

A collection agency that bears the consumer in mind and has a machine learning-driven, digital-first strategy removes this hurdle and enables a full-service, easy to use experience for both companies and consumers. With these technologies built into a team rather than a product or service, digital debt collection agencies can provide the services outlined above alongside a dedicated infrastructure and a team of technology experts. 

Choosing the right tools and support for your company’s collection efforts is more important now than ever before, and understanding the options that are available can help you to future-proof your strategy before it’s too late.

Still have questions? Our team is happy to help make sense of what a digital-first collections agency can do. Set up some time to chat!

Multi-armed bandit models and machine learning

By on February 19th, 2020 in Machine Learning

The term “multi-armed bandit” in machine learning comes from a problem in the world of probability theory. In a multi-armed bandit problem, you have a limited amount of resources to spend and must maximize your gains. You can divide those resources across multiple pathways or channels, you do not know the outcome of each path, but you may learn more about which is performing better over time.

The name is drawn from the one-armed bandit—slot machines—and comes from the idea that a gambler will attempt to maximize their gains by either trying different slot machines or staying where they are.

How do multi-armed bandits fit into machine learning?

Applying this hypothetical problem to a machine-learning model involves using an algorithm to process performance data over time and optimize for better gains as it learns what is successful and what is not. 

A commonly used model that follows this type of structure is an A/B/n test or split test where a single variable is isolated and directly compared. While A/B testing can be used for any number of experiments and tests, in a consumer-facing world, it is frequently used to determine the impact and effectiveness of a message.

You can test elements like the content of a message, the timing of its delivery, and any number of other elements in competition with an alternative, measure them, and compare the results. These tests are designed to determine the optimal version of a message, but once that perfect message is crafted and set, you’re stuck with your “perfect” message until you decide to test again.

Email deliverability plays a key role in effective digital communications. Check out our tips for building a scalable email infrastructure.

Anyone that works directly with customers or clients knows that there is no such thing as a perfect, one-size-fits-all solution. Message A, when pitted against Message B may perform better overall, but there is someone in your audience that may still prefer Message B.

Testing different facets of your communication in context with specific subsets of your audience can lead to higher engagement and more dynamic outreach. Figure 1 below outlines how a multi-armed bandit approach can optimize for the right content at the right time for the right audience rather than committing to a single option.

Rather than entirely discarding Message A, the bandit algorithm recognizes that roughly 10% of people still prefer it to other options. Using this more fluid model is also more efficient because you don’t have to wait for a clear winner to emerge, and as you gather more relevant data, they become more potent.

Multi-armed bandits and digital debt collection

Collections continues to expand its digital footprint, and combining more in-depth data tracking with an omni-channel communication strategy, teams can clearly understand what’s working and what isn’t. Adapting a bandit algorithm to machine learning-powered digital debt collection provides endless opportunity to craft a better consumer experience. 

Following from Figure 1, digital collections strategies can determine which messaging is right for which consumer. Sorting this data in context can mean distinguishing groups based on the size or the age of the debt and determining which message is the most appropriate. In a fully connected omni-channel strategy, the bandit can take a step back and determine which channel is the most effective for each account and then determine messaging.

These decisions take time and thousands upon thousands of data points to get “right,” but the wonder of a contextual multi-armed bandit algorithm is that it doesn’t stop learning after making the right choice. It makes the right choice, at the right time, for the right people, and you can reach your consumers the way they want to be reached.

TrueAccord is optimizing how our multi-armed bandit algorithms create the ideal consumer experience. Come learn more about how we collect better!

Supervised vs. unsupervised machine learning

By on February 5th, 2020 in Machine Learning

Machine learning is a powerful tool that many companies can use to their advantage. The ability to have algorithms make decisions based on large scale sets of data enables teams to build efficient, scalable tools. Some of these algorithms require frequent monitoring and management from data scientists in order to get up to speed and continue learning. Others are able to operate and learn on their own in order to generate new information to act on! 

Supervised and unsupervised machine learning algorithms both have their time and place. Let’s discuss a few examples, the difference between the two, and how they can be used together to create a powerful, AI-driven strategy for your company!

Supervised Machine Learning

Supervised learning algorithms are trained over time based on foundational data. This data will provide certain features as data points that will teach the algorithm how to generate the correct predictions. Figure 1, below, provides an example of a binary classifier and a set of data about cats and dogs that will teach the algorithm how to identify one or the other!

These models function best in situations in which there is an expected, intentionally designed output. In the example above, the expected output is that the algorithm can properly separate cats from dogs. In digital debt collection, it may be separating accounts that will be easy to collect on from ones that are more difficult. 

Classification vs. Regression

The models above are both examples of a supervised learning model that is seeking classification, but supervised learning can also be used to build regression models. The key difference between the two is that in a regression model the output is a numerical value rather than categorical.

A regression-based model may use input features such as income and whether or not they have children to accurately predict a person’s age. When using a regression based model in combination with consumer data, you can even segment demographics for communication and marketing. 

For full transparency we want to state that TrueAccord does not use its customer demographic data for these purposes. This is strictly an example.

With proper supervision, these models will become more accurate over time, and the data scientists building them can adjust them as business needs change. Whether you are gathering data using a regressor or a classifier, it is dependent upon the data scientists to build the most effective inputs in order to get the “correct” output.

Unsupervised Machine Learning

While supervised models require careful curation in building proper features that will lead to the “correct” output, unsupervised models can take large sets of unlabeled data and identify patterns without aid. The output variables (e.g. dog or cat) are never specified because it is now the algorithm’s job to process and sort the data based on similarities that it can identify. Using this method, you can learn things about your data that you didn’t even know!

Clustering vs. Association

Just as supervised models have primary methods for training their output data as either classification or regression models, unsupervised models can be trained using clusters or associations. Clustering algorithms gather data into groups based on like-features that exist in the data set. 

If you have thousands upon thousands of customer accounts in your system, a clustering algorithm can learn using the customer data and form them into distinct (but unlabeled) groups. Once it has assigned these clusters, data scientists can review the output data and make inferences such as:

  • This cluster is all of the accounts that have not yet established a payment plan
  • This cluster is all of the users that started signing up for a payment plan but didn’t finish the process

This new data set then provides the foundation for a new outreach strategy!

Building the infrastructure to process this data is the hardest part. Learn more about how TrueAccord is laying the foundation for scalable machine learning systems!

Association algorithms are the other end of unsupervised learning algorithms. Associations take the idea of grouping random data points one step further and can make inferences based on the data available. Continuing on from our account creation example, an association-based model can identify two data points and draw conclusions based on the patterns it finds. One such pattern may be:

A person that signed up for an account the first time they opened an email is more likely to pay off their balance.

The algorithm recognizes that multiple steps in a customer’s journey creates another data point. Because association algorithms are still unsupervised, a team of scientists will be responsible for labeling the output data, but the algorithm can outline previously unnoticed patterns.

The power of teamwork

By leveraging both supervised and unsupervised machine-learning algorithms, you can make decisions based on previously unfathomable scales of data. While they cannot necessarily be used to substitute one another, they can be used to create a perpetually improving cycle. Using unsupervised models to extract meaningful information from large data sets and building new supervised models to further hone your data creates more opportunities than ever before.