Harmonized data enables powerful customer insights

The challenges of complex, siloed data

“It’s really important to execute credit risk right,” says Rodrigo Longo, data engineering manager at Banco Bradesco. “If we better understand the situation of our customers, we can adjust their credit limits or give them the right amount of credit.”

Financial status, existing products, active loans, known debts, payment histories, and customer lifecycles are a few factors that impact the credit worthiness of a client. “One of our biggest challenges is that credit card information is complex,” says Longo. “A customer can make lots of transactions, and usually they have more than one card in their wallet. We work with credit cards, debit cards, retail store cards, prepaid cards—and each one has its own unique system from which to extract data. This creates a really challenging environment.”

Bradesco’s merger and acquisition (M&A) activity makes this particularly difficult. “Since Bradesco has acquired a lot of financial companies, each company brings additional platforms,” says Marcos Alvarenga, senior commercial intelligence manager at Banco Bradesco. “This creates a lot of data silos that make it difficult to create a source of truth.”

Data integration and harmonization create a platform for Trusted AI

M&A activity has caused teams at Bradesco to second-guess and question data. Inconsistencies, multiple versions, complimentary but competing siloed data, and incomplete data means teams lack trust in knowing the data is complete, correct, and current.

“The ultimate goal is to have trusted data to support the business,” says Alvarenga. “Business users need to have confidence in the data to help them make decisions that are best for customers, improve customer relationships, and provide the best financial solutions.”

To overcome this, Banco Bradesco migrated from on premises to the cloud, providing greater flexibility, agility, and accessibility to data. Teradata VantageCloud on Microsoft Azure has become Bradesco’s platform for Trusted AI. Now, Bradesco unifies and harmonizes data to empower users across the business to use analytics and AI.

“In order to achieve Trusted AI, we must first have trusted data, which we have created in Teradata VantageCloud on Azure,” says Longo. “By ensuring we have a good flow of data that we can rely on and trust, we empower the business areas to better tend to customer needs. We also reduce the risk of incorrect outputs from AI. For example, when we model data in H2O.ai, we use the curated data in Teradata that has strong data governance."