Inside Erica: The Making Of Bank Of America’s AI Advantage

Set in 2025, this case examines Bank of America’s (BofA) development and adoption of Erica, an artificial intelligence (AI) powered virtual assistant across its multiple business lines.

Developed in-house and launched in 2018, Erica by 2025 had evolved from addressing basic account queries to enabling transactions, issuing fraud alerts, and delivering proactive and personalised financial insights to over nearly 50 million active users. BofA’s first-mover lead, large scale proprietary data, sustained investment, in-built offramps to human support, and disciplined AI governance had led Erica to become a strategic asset. The bank expanded it beyond retail banking into its wealth-management businesses, supporting select functions such as proactive alerts, intelligent scheduling, document search, and workflow assistance. By Q3 2025, Erica supported three billion interactions and customer satisfaction was at an all-time high, service cost had significantly lowered, while employee productivity had increased manifolds.

Yet, as BofA prepared to establish Erica as an enterprise-wide “AI super-platform” with deeper inroads into its commercial banking and wealth management businesses, many concerns emerged. Wealth management, as a “high-trust, high-touch” business, represented a fundamentally different environment from retail banking with clients expecting tailored financial advice, discretion, and risk assessments. Moreover, interactions typically involved multiple stakeholders and multi-product portfolios along with higher regulatory exposure with more stringent fiduciary standards and privacy obligations. Did Erica’s existing architecture provide a strong enough foundation to expand into wealth management or was a fundamentally different model required? At the same time, the rise of generative AI (gen-AI) had created a new challenge. Could BofA use gen-AI to enhance Erica without compromising on its accuracy and reliability?

The case helps students analyse use of AI in enhancing existing capabilities and building new ones in highly regulated industries. They examine the role of customer-centricity, capability creation mindset, and feedback-based iterative evolution in developing emerging tech-based solutions, and the need for differentiated approaches across different businesses. Students also apply the transaction cost economics theory to assess the benefits of building AI capabilities in-house versus outsourcing.

Inspection copies and teaching notes are available for university faculty. To receive an inspection copy and teaching note, please email ccxshop@smu.edu.sg with your registered faculty email ID and a link to your contact information on the faculty directory at your university as verification. An inspection copy and teaching note will then be sent to your faculty email account.

Download information

SMU Faculty/Staff can download the case & teaching note on iNet with your SMU login ID & Password via the following links:

·      The Case (SMU-26-0004)

·      Teaching Note (SMU-26-0004TN)

For purchase of the case and supplementary materials via The Case Centre, please access the following links:

·      The Case (SMU-26-0004)

·      Teaching Note (SMU-26-0004)

For purchase of the case and supplementary materials via Harvard Business Publishing, please access the following links:

·      The Case (SMU-26-0004)

·      Teaching Note (SMU-26-0004)