Bank of America adds generative AI to EricaAssist for service staff
The bank says EricaAssist now gives real-time recommendations to about 18,000 customer service employees during client interactions.
By Dominic Okoye · Staff Writer
· 3 min read
Bank of America has added generative AI features to EricaAssist, its internal assistant for customer service staff, the bank said Tuesday. The upgrade matters because it puts recommendation-making AI into the workflow of about 18,000 employees at a large regulated bank, rather than limiting the technology to consumer-facing chat.
The Charlotte, North Carolina-based lender said the enhanced tool can deliver personalized guidance within seconds while employees are speaking with customers. Bank of America did not disclose a vendor, model provider or implementation cost for the new capabilities.
EricaAssist is separate from Erica, Bank of America’s customer-facing virtual assistant. EricaAssist runs as a desktop widget for employees and is designed to surface information during live client interactions, according to Ashley Ross, the bank’s head of consumer client experience and business transformation.
The bank said the tool already reduces average call time by nearly one minute per interaction. That metric is useful, but incomplete: Bank of America did not provide total call volume, baseline handle time, customer satisfaction changes or error rates tied to EricaAssist.
What the upgraded tool does
Before the generative AI update, customer service employees used EricaAssist to pull up information while handling customer calls. The system could show why a customer was calling, display the customer’s tenure with the bank and answer employee questions during the interaction.
Tom Ellis, Bank of America’s chief information officer and head of consumer technology, said the new version can interpret what is happening in a conversation, summarize it and suggest a next step to the employee. The bank described the output as real-time recommendations aimed at resolving customer issues faster.
Ross said the goal is to help employees serve clients more efficiently. Ellis said Bank of America expects the feature to improve resolution time and customer experience, though the bank did not publish measured results for the generative AI features specifically.
Why banks are putting AI behind the counter
The rollout fits a common pattern in financial services: deploy AI first as an employee copilot, where staff remain in the loop and customer-facing risk is easier to contain. For a bank with $3.5 trillion in assets, shaving time from support interactions can be material, but the operational case depends on consistent accuracy, compliance controls and adoption by front-line staff.
Bank of America said it uses employee feedback to guide future EricaAssist development. The bank plans to expand the tool later this year to additional servicing scenarios and business lines.
The company’s technology budget provides context for the scale of the effort. Bank of America spends about $14 billion a year on technology, with roughly $4 billion allocated to new initiatives, including AI. The bank did not say how much of that budget is tied to EricaAssist or to generative AI more broadly.
For enterprise software vendors, the deployment is another signal that large banks are willing to use generative AI inside operational systems when the use case is narrow and measurable. For Bank of America, the next proof point is whether the upgraded EricaAssist can show gains beyond shorter calls, including better issue resolution and fewer escalations.
This story draws on original reporting from CIO Dive.
