As per Boston Consulting Group, since 2008, Financial services firms have spent about $321 Billion in conduct/mis-selling related issues. Regtech firms have focused on improving efficiencies of compliance processes within banks. However, mis-selling products and services is a behavioural problem, and a harder nut to crack for Regtech firms. In this post, I have attempted to define some of these behaviours, the outcomes, and how a Regtech app using Artificial intelligence could catch these behaviours by plugging into the right sources of data.
Regtech firms focusing on mis-selling need to define a methodology and the data sources to develop AI and machine learning to proactively address mis-selling. Some of the key data elements needed to identify mis-selling in my opinion are,
- Product Life cycle
- Marketing and Sales
- Customer/Social Media Sentiment
- Employee compensation
- Operations and Technology Spend
Product Life Cycle data is critical to understand the process to develop and approve a product. It shows the involvement of senior management in the process, and the governance involved in getting the product out to the market. Data around complexity scores of a product will be useful for compliance teams to understand if there is enough support for both employees and customers to understand the product.
Marketing and Sales data is critical to ensure that right amount of money is spent in marketing the product and the sales commissions are aligned to the firms strategy. Sales data is also critical to analyse a sudden spike in sales of a product. Data can show if it was because of the new sales manager, a tweak to the product or just plain old mis-selling triggered by some year end target.
Customer complaints and social media sentiments are required to understand if a product or a service sold is keeping the customer happy. Also, in the open data world, product usage information could give firms a good view of, if a customer is using them.
Employee compensation often is directly related to aggressive sales done by sales people at banks. Combine this data, with sales of products, usage of products by customers and even complaints from customers, you get the view of how a particular bonus structure drove an employee to sell a product to a consumer when he or she didn’t need it.
An AI algorithm that can have this data can spot regularly occurring trends such as the above, and even alert senior management when they approve a particular product, or agree to a compensation structure. Of course, many firms already use social media sentiment to spot product issues, but that is just one side of the story.
The root of the problem is within Financial services firms where their strategy often doesn’t align with their culture. AI can spot if their product strategy and employee compensation are genuinely aligned with their “Values”. AI could spot mis-selling based on social media sentiments, identify them even before it gets to social media, but a even better state could be to identify patterns that instigate mis-selling behaviour even before they occur.
In proactively managing mis-selling, banks can not only save fines they have paid to regulators but also cut down on the £5 Billion claims market. I believe, a well analysed framework that identifies mis-selling issues and reports on them to the regulators would help all parties involved. It will most definitely save billions for banks. Regtech firms, are you listening?
Arunkumar Krishnakumar is a Fintech thought-leader and an investor.
Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.