Join the Community

21,754
Expert opinions
43,842
Total members
451
New members (last 30 days)
188
New opinions (last 30 days)
28,612
Total comments

Object reference not set to an instance of an object.

1 comment

AI/ML solutions can be used to help companies assess the needs of their customers more effectively using behavioral microtargeting; address those needs through personal and relevant offerings; and significantly increase their conversion rates for their offerings (e.g. consumer loans). 

Developing data solutions for financial institutions is somewhat unique in that although they typically have one of the most valuable collections of data, they are also some of the most strictly regulated businesses for data privacy. So when developing a solution for them, you always have to navigate three layers of privacy and security regulations: local, regional (EU), and the bank’s internal policies. 

This means that any solutions need to excel in accurately predicting what their clients are in the market for (i.e. their next-best offer), while still operating under rather tight constraints on security and data privacy.

The truth is that these restrictions aren’t as limiting as most banks imagine. Modern data practices allow you to retain most of the value from your data without violating any data privacy laws (e.g. the GDPR in Europe).

DATA ANONYMIZATION TECHNIQUES 

The key is to use a variety of data anonymization techniques that protect the privacy of your users or clients. Wherever possible you should use non-identifying proxies in place of sensitive and protected data. For example, you can add “noise” to location data, because exact coordinates are not needed – a general location or neighborhood is generally enough. There are equivalent proxies for many of the other protected fields that can be used to generate extremely useful, yet mostly synthetic, data.

Banks can use an anonymized version of both transactional and client data to anonymously sort clients into highly specific groups based on behavioral patterns such as spending habits, mobility, leisure activities, social background, family, household relationships, and friends.

This approach allows them to compute very accurate predictions about their clients' propensity to be in the market for various financial products, based on their similarities to others in their extremely microsegmented cohort. Sorting at this level of granularity is only possible with the assistance of AI. For example, if one member in this extremely narrow cohort is in the market for a consumer loan, then there is a very high likelihood that other members of this microsegment would be as well.

The final result is that with a well-trained model, any bank can now calculate each of their customer’s propensity-to-buy for each of the financial products that the bank is offering.

INTEGRATION & CAMPAIGN MANAGEMENT

For this kind of model to be useful to a bank’s sales and marketing reps, it should be incorporated directly into the marketing and campaign management tools that they already use every day. As an example, representatives would be able to instantly sort for the top X clients who are most likely to be in the market for whatever product they are creating a campaign for.

If you’ve never run marketing campaigns before, this is a major timesaver and significantly more accurate than spending hours flipping back and forth between Excel tabs trying to do ad hoc analysis on who to add to your campaign!

Another thing that AI can help with is big-picture campaign management. Typically most of the big banks have strict contact policies that limit how often customers can be contacted for marketing promotions. If they contact you with an offer, sales reps typically need to wait 4–6 months to offer you something else. So if you offer a client a product that they are not in the market for at all, then suddenly you’ve just disqualified them for all of your upcoming campaigns. Maybe even a campaign that would be a perfect fit for them. This can be a costly mistake.

Modern AI/ML solutions also help banks map out which products they should run campaigns for, which customers to include, and how to space out these campaigns to optimize their conversions without violating their contact policies. For example, if there’s a large group of clients who are in the market for two products, AI can help make sure to schedule those two campaigns far enough apart so that you can include these customers in both campaigns.

It’s a common misconception that customers hate ads or sales calls. The reality is that they just hate ads and sales calls for products that they don’t want! However, if you can give them an offer that is personalized, relevant, and timely, you’ll find that more times than not they’ll even thank you for calling!

External

This content is provided by an external author without editing by Finextra. It expresses the views and opinions of the author.

Join the Community

21,754
Expert opinions
43,842
Total members
451
New members (last 30 days)
188
New opinions (last 30 days)
28,612
Total comments

Trending

Dirk Emminger

Dirk Emminger Managing Director at knowing finance

Competition and Cooperation: In an AI-Dominated World (A2)

Sireesh Patnaik

Sireesh Patnaik Chief Product and Technology Officer (CPTO) at Pennant Technologies

Empowering the Lending Industry: How Low-Code, No-Code, Pro-Code Platforms are Driving Innovation

Now Hiring