company

Bank of America

Financial Services / Banking
72Good
Aio Composite Score
Confidence
ApplicationHeavy
high
ImpactMixed
medium
OpennessTransparent
medium
Bank of America is one of the most advanced AI adopters in the financial services sector, deploying AI at scale across customer-facing and internal operations. Its flagship virtual assistant Erica has surpassed 3 billion client interactions and is used by more than 20 million active customers, while Erica for Employees is used by over 90% of its 213,000-person workforce. The bank publicly emphasizes human oversight, responsible AI governance, and a 16-parameter vetting framework, maintaining a generally consistent narrative across sources. While headcount reductions and AI-driven efficiency gains are both occurring, the causal link between AI and job displacement is emerging but not yet fully evidenced from named roles or restructuring announcements explicitly tied to AI automation.
Application
How is AI being used?
Heavy
high
Bank of America has deployed AI at enterprise scale across all eight lines of business, with 270+ AI models in production, 90%+ employee adoption of…
Impact
How have people been affected?
Mixed
medium
Positive outcomes are evidenced — 50% reduction in IT service calls, productivity gains for financial advisers, and efficiency savings in payments…
Openness
Company honesty about AI use.
Transparent
medium
Bank of America's public narrative — emphasizing responsible AI governance, human oversight, a 16-parameter vetting process, and the AI oversight…
AI Tools & Platforms Detected
Erica (customer virtual assistant)Erica for Employees (internal virtual assistant)AskGPS (AskGlobal Payments Solutions – generative AI assistant for transaction banking)CashPro Chat (corporate/commercial client AI interface)Merrill virtual assistantBank of America Private Bank virtual assistantAI fraud detection modelsAI-generated market information for financial advisersAI coding toolsLarge Language Model (LLM)-based proof-of-concept tools
AI Hiringmoderate
30AI roles/ 33165 total
Artificial Intelligence Engineering ManagerArtificial Intelligence Senior Security EngineerSr. Information Security Officer – Generative Artificial IntelligenceAI/ML Engineer (AIML Engineer)Data Scientist – Data and AI team, Global Payment Solutions
Key Evidence (12)
More than 90% of Bank of America's 213,000 employees use Erica for Employees AI assistant, which has driven a 50% reduction in IT service desk calls since its 2020 launch.
2024
web searchSource ↗
Bank of America launched AskGPS in September 2025, a generative AI assistant built in-house on 3,200+ internal documents, designed to save tens of thousands of employee hours annually for its GPS team serving 40,000+ business clients.
2025-09-30
web searchSource ↗
Erica has surpassed 3 billion client interactions since 2018 and averages more than 58 million interactions per month, with 20 million active users.
2025-11-05
web searchSource ↗
Bank of America holds more than 1,200 AI and machine learning patents and pending applications, representing 17% of its total patent portfolio of nearly 7,400 granted patents and applications — the most U.S. granted patents of any financial services company.
2025-04-08
web searchSource ↗
Bank of America invests $13 billion annually on technology, with $4 billion directed to new technology initiatives in 2025, a 44% increase in strategic tech spending over the past decade.
2025-11-05
web searchSource ↗
Bank of America's AI approach explicitly includes human oversight, transparency, and accountability; the bank applies 16 parameters to vet AI capabilities before deployment and maintains an AI oversight council.
2025-04-08
web searchSource ↗
What insider reports could clarify
  • How are AI-generated outputs reviewed before they reach clients or are acted upon?
  • Have any specific roles or teams been formally replaced or eliminated due to AI automation?
  • Do financial advisers receiving AI-generated market briefs have discretion to override or flag AI outputs?
  • Are employees informed when AI, rather than a human, has handled their HR or IT request?
  • How does the 16-parameter AI vetting framework handle failures or post-deployment issues?
Submit a ReportClaim This Profile