Working in the fast-moving world of artificial intelligence (AI), we are well-versed in tackling ethical debates head on. Long ago, we outlined our five core principles of AI. We pledged that our application will always be Beneficial, Ethical, Explainable, Relevant and Secure. Our BEERS strategy has defined how we have developed our natural language processing because we believe ethical AI has the greatest benefits for businesses, customers, and society.
We’re helping millions of consumers have automated, intelligent conversations with organisations across numerous industries. Whether its renewing telco contracts or insurance policies, rescheduling deliveries or making changes to banking services, the principles are the same.
We use proactive conversational AI to reach out before customers identify there is a problem. Next, we present solutions to make them feel in control. Finally, we keep conversations going using the channels that they are most comfortable with, and language tailored to their preferences.
Underpinning all of this is a commitment to ethical AI, with a focus on explainability and data protection.
However, it’s not just current services that proactive AI can help with. We’ve seen success in using our AI to increase the effectiveness of debt collection too. We have generated a 350% improvement in recovery rate (compared to bailiff collections) and a 42% reduction in time to collect debt for one client by using ethical AI methods.
Here, we’ve outlined our four guiding principles, which show how we apply AI in debt recovery cases.
Fairness and integrity
Debt collectors should follow moral principles of fairness and honesty. They should avoid treating debtors differently based on race, gender, or any other attribute. They should also act in a way that does not cause distress or fear. Sadly, this isn’t always the case. Four in 10 people in problem debt say they have been left “fearing for their lives” due to the aggressive tactics of debt collectors.
But the dial is slowly shifting in how much people trust AI to make fair and calm decisions. Forty-eight per cent of Americans say the positives of greater AI adoption outweigh the negatives. Americans want AI to play a greater role in decision making in an array of industries, such as technology (66%), manufacturing (61%), logistics (58%) and retail (52%).
In the workplace, 54% of people are more trusting of robots than their managers. One in 4 say machines are better at working with unbiased information.
We believe the best application of AI is one where humans and computers are working to their strengths.
Within debt collection, this could be AI proactively reaching out to debtors to inform them of a missed payment. AI can then collect necessary information about someone’s financial situation before suggesting options for repayment, which an agent can sense-check. By using AI that is explainable, the end user can see the methodology behind why recommendations have been made. The language used by the tool can be controlled, ensuring that no one receives unfair treatment.
Debtors should feel in control of the conversation and be able to agree realistic payment plans. This is especially important for those who are facing many different debts or priorities in other areas of their lives.
Creditors who use digital tools to offer flexibility are more likely to see a return in what they are owed because these give the debtor control over their situation. A recent example is the UK Government’s Time to Pay facility. This tool enables people to spread the cost of their tax bill into monthly payments. Since April 6th last year, 21,600 people who were unable to pay their bill in full have set up a payment plan – an increase of 3,900 on the year before.
Organisations should consider going the extra step and proactively communicating with debtors using AI. This way, creditors can reach out and help debtors create payment plans that are right for them once the first payment has been missed. This ensures that debtors feel more in control of their situation, which in turn increases response rates.
Money is a daily worry for many Americans. Sixty-five per cent say money is a “significant source of stress”, while the average American worries about debt six times a day. More than half of Americans say they do not feel comfortable talking about their finances. This is a significant barrier for debt collectors who need to encourage debtors to have these conversations. Those who owe money can feel strong-armed into opening up and be left feeling judged, vulnerable, and misunderstood.
Here, AI once again presents a solution. A Penn State University study found people trust giving data to machines rather than people because they do not gossip and are not prone to human error. After all, machines are less likely to leave their work computer unlocked when they go to lunch.
Debtors should always feel respected. This is important in the tone that you take in your communications. Our conversational AI never uses inflammatory or accusatory language even when customers’ messages do.
This is one of the many benefits of using AI to have these challenging conversations. Our programme’s language can remain cool and professional in the face of angry or upset customers. AI can suggest objective solutions or redirect customers to a call centre if debtors do need to speak to someone. It’s worth noting in our case study that 55% of our debt recovery is done digitally, with no agent required. We can achieve this because our methods draw on the experience of millions of similar conversations, ensuring the debtor is treated with dignity and respect.
By adhering to these four principles of fairness, control, data responsibility and respect, we are creating a better experience in debt recovery for both those who owe money and the creditors they owe. We are proving it is possible to increase debt recovery rates in a more efficient and more ethical way.