Bank of America
conversational financial assistant at scale
Erica, the conversational assistant in the Bank of America app launched in 2018, passed 2 billion interactions and 42 million clients in April 2024, then 3 billion interactions and nearly 50 million users in August 2025.
Key points
- Erica conversational assistant in the mobile app for balance, transactions, budget, and proactive alerts.
- In-house NLP and voice engine, integrated with core banking, on real-time account data.
- More than 3 billion interactions and 58 million per month; response in 44 seconds for 98% of clients.
- Evidence level A, confirmed active status.
Objective
Handle most of clients' daily needs in self-service (balance, transactions, budget, alerts, transfers) in the app, deflect simple requests away from costly channels, and push proactive insights to anchor app usage.
The deployment
Erica is the assistant built into the Bank of America mobile app, launched in 2018. The client queries it in natural language or by voice to check a balance, find a transaction, manage a budget, or receive an alert. The volumes are massive and documented by the bank: more than 2 billion interactions and 42 million clients helped announced in April 2024, then 3 billion interactions and nearly 50 million users in August 2025, with an average of more than 58 million interactions per month. Erica also pushes proactive personalized insights without the client having to ask.
Results Proof A
Volumes published by Bank of America itself in several consistent official press releases (T1), over several years (2 billion in 2024, 3 billion in 2025), in line with its digital engagement disclosures tied to quarterly reporting. Several consistent primary sources from the subject brand raise the level.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- outil Erica (moteur NLP maison) Natural language and voice understanding engine built in-house by Bank of America, integrated with core banking; the sources do not name any third-party LLM.
- outil Moteur d'insights proactifs (analyse predictive maison) Predictive analysis on transactions that detects events (unusual spending, upcoming due date, savings opportunity) and pushes personalized alerts.
- plateforme App mobile Bank of America Erica's single channel (text and voice), with deep integration to accounts and transactions in real time.
How it runs, concretely
For ops teams-
1Client request client
The client types or says their request in the app (balance, transaction, budget, transfer).
-
2Understanding and response AI
Erica interprets the intent, fetches the account data, and responds, on average in under 44 seconds.
-
3Proactive insights AI
Outside of any request, Erica detects events (unusual spending, upcoming due date, savings opportunity) and pushes a personalized alert.
-
4Supervision and improvement data team
The in-house team tracks volumes, satisfaction, and misunderstood cases to broaden the handled scope.
The intent expressed by the client (natural language or voice query) plus the transactional signals that trigger proactive insights. Without clean, up-to-date account data, the insights and alerts lose their relevance.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- Real-time account and transaction data
- History of client intents for training
- Event detection rules for the proactive insights
Org prerequisites
- Dedicated digital product team over time
- Banking security and compliance framework
- Governance of proactive messages (frequency, relevance)
Possible stack
- NLP/LLM engine
- Predictive analysis layer on transactions
- Deep integration with core banking and the mobile app
The plan, step by step
- Step 1Map the 20 most frequent client intents (balance, transaction, transfer, budget) from support and app logs.Deliverable: Intent repository prioritized by volume.
- Step 2Build the assistant in read-only mode on account data for these intents, in a test environment.Deliverable: Internal prototype tested on real data.
- Step 3Open a beta to a client segment in the app, with human escalation and transparency on the AI nature.Deliverable: Assistant in beta, resolution rate and satisfaction tracked.
- Step 4Broaden the scope to transactional actions and alerts, with banking security and compliance validation.Deliverable: Assistant generalized in the app.
- Step 5Add the proactive insights triggered by transactional signals, with governance of message frequency.Deliverable: Personalized alert engine in production.
First step: Map the 20 most frequent client intents in the app and connect the assistant to account data in read mode before adding the proactive layer.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 BofA's Erica Surpasses 2 Billion Interactions, Helping 42 Million Clients Since Launch Primary archive pending
- S2 A Decade of AI Innovation: BofA's Virtual Assistant Erica Surpasses 3 Billion Client Interactions Primary archive pending
- S3 Bank of America's Erica surpasses 2 billion interactions Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.