Context

The humanization of credit processes proved to be the safest way to measure the trust and credibility of customers in financial institutions, with a view to responding to Open Finance integration opportunities.

With this principle in mind, we worked on a digital transformation case study for the credit analysis area, where the main business problem was oriented towards the hypothesis of unifying the moment of communication of the sales executive, as an internal customer, by improving the self-service experience of the user of digital channels (end customer).

Problem

Instead of receiving several PDF files of bank statements, we seek to offer an internal and low-cost experience for the customer by qualifying the moment of contact with the final customer. The digital account integration solution aimed to minimize the need for manual income documents by offering account connection via open-banking in cycles of web scraping experiments.

In other words, we eliminate some issues involving business income offenders when we raise awareness of the verticals:

• minimizing "income proof" properly as a major pain point for the customer;

• the verification of unstable income information as operational (and manual) pain (from the credit analysis team and the commercial);

• wear with communication in the customer experience (commercial and CRM);

• mediation between "Trust vs. Risk" by the Corporate Business Intelligence areas.

Assumption: Sending multiple PDF files offers low end and internal customer experience

Hypothesis: Unify the moment of communication with the client to request income documents, offering connection to accounts via open-banking scraping

Solution

Optimized income tabulation, to be integrated via API, provided to the credit team in order to facilitate contact with the customer when obtaining the necessary documents for income tabulation.

Results

With this project, it was possible to:

1. Reduce origination CAC and increase the level of automation of the conveyor reflected in the E2E customer experience

2. Potentially offer a 49.9% reduction in the time from application to credit approval (with credit desk analysts)

3. 150+ Pre-built credit statements for building predictive credit models (with data science engineers);

4. 25% increase in risk score approval and termination rates (with portfolio team); 200 accounts connected in 2 weeks of experimentation (with the product, design and technology teams);

Insights

• New customer segments as promoters of Open Banking

• PF showed more confidence than PJ

• Digital ageism against all odds and unlike stereotypes

Learning

1. Data Driven sampling with the commercial, Business Process and Product Hacking team

2. Experiments to understand customer perception for certain financial behavior profiles

3. Quali and Quanti triangulated data patterns helped to identify and promote correlations to new commercial segments in the context of real estate lending.

Friction:

In the first two occurrences of the experiment, there was no public engagement and with the data obtained it was not possible to contribute and support evidence or conclusions. The high level of uncertainty was positively influenced by this bias, as we knew how much we didn't know about something. However, we promote cycles of experimentation with potential learning.

Action:

To convert this, we started exploratory experiments (Wizard of Oz) to reduce harmful uncertainties and identify the true potential to be worked on (building a roadmap for the future Open Finance integration). With the experiment, it was possible to readjust the target audience - starting with individual clients, we found more opportunities for corporate clients given the representativeness of the executed data - influencing a high flow of interactions during the test cycle because the possible impact generated by the high volume of PJ customers being captured in the experiment. As a result, it was possible to have an even greater reach of market share for the B2B2C audience, in addition to important new skills for the technology team to develop, considering batch executions.

Lessons learned:

The open finance validation use cases needed to be tested with the 60+ public, which represented the largest volume of customers in this business unit, but also a new customer segment as a promoter of Open Banking. Thus, in 3 months we evidenced key opportunities in our partners' ecosystem with data that represented the familiarity and interest in the studied technology, in addition to demonstrating greater confidence in the brand by the user/customer.