Bias Index: Marcia Tal on how to cope with customer complaints

CEO of PositivityTech Marcia Tal reveals her thoughts on the current state of bias and discrimination in financial services

Bias Index: Marcia Tal on how to cope with customer complaints

PaySpace Magazine has reached out to PositivityTech’s CEO Marcia Tal to ask her about how the Bias Index helps financial institutions to pinpoint bias and discrimination within their customer complaints.

Could you give a short introduction to the Bias Index?

In banking, bias reveals itself in subtle ways. Customers may have a negative experience with a financial institution and find that discrimination is at the root of the problem. They may face discrimination due to their race, age, gender, sexual orientation, religion, military service, or citizenship.

In May of 2020, we released the Bias Index, a product of the PositivityTech® intelligent platform. It is an artificial intelligence-powered predictive model that identifies prejudice within customer and employee complaints and makes it possible for financial institutions to respond to systemic discrimination and repair their products or unjust practices.

With tensions due to social inequities in the spotlight, the PositivityTech platform can help financial institutions turn their customer voices into intelligence and proactively weed out systemic discrimination. We believe that negative input can have a positive impact and that together, we can flip the script on complaints and help institutions provide equal treatment to all of their customers.

How did you come up with the launch of the Bias Index?

Each one of us has faced unequal treatment at different points, and it is wrong that discrimination continues to play a role in banking. For example, last year, the media revealed that a Black customer and a Black employee at JPMorgan Chase were unable to gain access to the same opportunities as their white peers. In response, the bank implemented mandatory diversity training and said that it “would pay more attention to employee complaints.” Recent protests have highlighted issues of discriminatory practices within many industries, including banking. While banks are doing their best to weed out discrimination, it still rears its ugly head.

Customers share with us what we need to know, but they often do so indirectly. I created the Bias Index because I understand the value of customer voice data, I have expertise in the advanced analytics capabilities that are needed to extract this intelligence, and I am passionate about eliminating bias from our institutions. Bias is wrong and there is no place for it in our organizations’ decision-making processes. Financial institutions must address issues of bias in order to enforce fair banking practices, establish safe and sound lending practices, ensure compliance, and prevent lawsuits.

How do biases related to race, age, religion, and sexual orientation reveal themselves in customer complaints?

With the Bias Index, we can pinpoint particular products, policies, and practices that have a high correlation with bias. For example, we’ve discovered that products like mortgage and auto loans are four times more likely to have biased practices because these transactions are done face-to-face and feature a high risk of employees bringing bias into their decision-making. On the other hand, complaints about products with low face-to-face interactions, like virtual currency, debt collection, and credit reporting show much lower bias scores.

Has the ongoing COVID-19’s pandemic affected the level of customer complaints as well?

The emergence of the coronavirus health crisis has negatively impacted financial institutions’ ability to help their customers. Consumers are complaining loudly and are not afraid to publicly air their complaints. While personally identifiable information is masked by the Consumer Financial Protection Bureau (CFPB), customers typically do not agree to publicize their complaints. However, the number of published narratives has more than doubled since March 10th, revealing a behavioral shift. People are more frustrated and are now actively sharing their complaints.

Those complaints reflect a growing disappointment in banking institutions’ lack of contingency plans for a crisis like this, fears about the inability to make payments, deferred payments negatively affecting credit scores, and predatory lending practices.

Using PositivityTech, we have also found that complaints related to COVID-19 have a Bias Index that is 45% higher than non-COVID-19-related complaints.

How much time would it take to lower the level of discrimination in financial institutions? Do you have any expectations regarding this problem?

Bias is a systemic problem that has long plagued financial institutions, and the solution will come from the strategic intersection of human expertise and advanced AI.

With the Bias Index, we are bringing together untapped, unstructured data and powerful analytic capabilities to draw conclusions and solve complex problems. A leadership team that wants to solve discrimination in their institution understands the potential of the Bias Index’s findings, and what high-level financial decisions need to be made to support them.

It is difficult to put a timeline for this kind of necessary structural change. We are excited to be a catalyst for positive innovation.

How can AI be used to identify bias?

The Bias Index is an AI predictive model that digests the contextual reference of a complaint in order to reveal the root of the complaint. This allows financial institutions to focus on repairing unjust products or practices. Bias can be explicit, implicit, and suggested. Artificial intelligence makes it possible to pinpoint.

What are the steps banks can take to prevent systemic discrimination?

First, use your data. Start approaching your customer complaints as strategic data, and extracting intelligence in order to address the individual issues that need to be resolved. Customer complaints are powerful and predictive data points that can help institutions prevent high-risk issues. Consider how your customers’ language can be used to predict future issues, and recognize that unstructured data with expert analytical rigor and relevant business context can bring structure and solutions to complex issues.

Secondly, explore whether your employees represent the diversity of your customer base.

Empower employees to build task forces that address large and significant challenges, and then invite them to share recommendations and to implement those recommendations.

Finally, ensure that leadership embraces diversity and inclusion. This is what drives an organization’s culture.

Let’s imagine banks failed to implement these steps. What do you imagine would be the worst-case scenario?

There are endless worst-case scenarios that can happen to financial institutions that do not give their employee and customer voices the value and attention they deserve. In some instances, those worst-case scenarios have already happened and include lawsuits from dissatisfied employees and customers, negative media focus, regulatory enforcement actions, and permanent damage to the company’s reputation.

Marcia, congratulations on being selected as a speaker at Money20/20! Do you have any other plans regarding enhancing customers’ experiences in the fintech niche?

With the Bias Index, and PositivityTech more broadly, we are giving the voice of the customer the attention and power it deserves. Our goal is to invigorate organizations to see their customer complaints as a source of insight that can power success, instead of a drain on resources. Those insights can motivate structural change that enhance customer satisfaction with their financial institutions.

PositivityTech directly addresses key drivers of business risk, breaking down complaint language by parsing components in a multitude of sophisticated models. Those include the Lawsuit Model, which predicts the likelihood of a lawsuit, the Debt Collection Model, which predicts the likelihood of non-payment of debts owed, and the Keyword and Severity Score, which identifies severe complaints and future risks based on the language of customer narratives.

One particularly useful model for improving the customer experience is the Categorization Solution Model, which uses powerful analytic techniques to extract multiple issues nested within a single customer complaint and more successfully address the root of the problem. Complicated customer complaints are often wrongly manually categorized by financial institution employees and dealt with unsatisfactorily, resulting in frustration and loss of trust.

Customers tell us what we need to know in their complaint narratives, but often indirectly – and the issue is not always the issue. That is where PositivityTech steps in. Banks are under pressure to make their operations smarter and to find ways to accurately solve their customers’ issues. PositivityTech has made it possible for financial institutions to use their own data to pinpoint the right solutions for their customers – a win for both customers and banks.

SEE ALSO: