Own the assistant, insist on perfect data, scale one proven model across 30 billion interactions.
How is Bank of America using AI?
Bank of America's AI strategy is the opposite of a big-bang rollout: a decade of one assistant, Erica, extended step by step into employees' tools, treasury products and advisor workflows, funded by a technology budget the bank now puts at $14 billion a year, more than $4.1 billion of it for new initiatives including AI. Management measures it in interactions and hours rather than headlines — 2 million Erica conversations a day doing the work of 11,000 people, 400,000 employee prompts a day, more than 300 approved use cases 'all of which have good economics'. CEO Brian Moynihan's stated constraint is accuracy: 'the data has to be perfect', and the bank keeps hiring 4,000 interns and campus recruits a year despite the automation. Its Q2 2026 earnings deck puts numbers on the programme: about 200,000 active users of general-purpose AI tools, 114 live generative-AI use cases of which 34 are fully deployed, about 19,000 developers using coding assistance with productivity up more than 20%, and more than 600,000 AI training courses completed. The 2025 10-K flags expanded use of 'autonomous AI agents' and reliance on third-party AI models as risks.
Lead bank
Bank of America, N.A.
Assets (lead bank, Jun 30, 2026)
$2.65 trillion
Charter
National bank (OCC)
Headquarters
Charlotte, NC
Flagship platform
Erica (and EricaAssist)
AI leadership
Hari Gopalkrishnan · Nikki Katz · Kevin Milsom · Brian Moynihan · Dean Athanasia · Jim DeMare · Thong Nguyen · Aditya Bhasin · Holly O'Neill · Tom Ellis · Ashley Ross · Jarrett Bruhn · Mark Monaco · Cameron Wadley · Shimna Sameer · Patricio Diaz · Bernard Mensah
- Technology budget: $14 billion a year, more than $4.1 billion of it allocated to new initiatives including AI, per the bank's Fast Facts page (October 2026); in 2025 the CTIO put it at $13 billion with $4 billion for new technology.
- Erica: launched 2018; nearly 50 million users, 3 billion-plus interactions, 58 million a month by August 2025; 20 million regular users and 2 million interactions a day by the January 2026 investor day; 3.8 billion interactions since launch and nearly 200 million a quarter by October 2026 (Fast Facts).
- AI runs in four declared domains — intelligent agents, search and summarisation, content generation, operations and coding — and the bank holds more than 1,800 AI-focused patents and applications (Fast Facts, October 2026).
- Employees generate more than 400,000 prompts a day; more than 300 AI use cases are approved (Q2 2026 earnings call).
- Advisor tools: an AI meeting journey for 15,000 Merrill and Private Bank employees in the first phase (March 2026); CashPro Forecasting saved corporate clients 250,000 hours in 2025.
- Governance posture is conservative by design: 'zero tolerance for things like hallucinations, bias', per the CTO; the CEO frames AI as a management responsibility, not a technology one.
- Q2 2026 earnings deck: ~200,000 active users generating 400,000-plus prompts a day; 2,000-plus departmental AI templates; 20,000-plus wealth professionals and ~4,000 investment bankers using AI; ~19,000 developers with coding assistance (productivity up 20%-plus); ~5,000 call-centre agents on real-time AI recommendations; 114 live generative-AI use cases, 34 fully deployed.
- AI in coding took 30% out of the development stream in 2025, saving about 2,000 people's worth of work, and Microsoft 365 Copilot is out to 200,000 teammates; the CEO puts direct AI spend at 'several hundred million dollars' (Q4 2025 call).
- Technology budget for 2026 is $13.5 billion a year, of which about $4 billion goes to new initiatives such as AI (Merrill and Private Bank release, March 2026; Banking Dive, April 2026). The CTIO put it at roughly $14 billion in a September 2026 Fortune interview, and the bank's Fast Facts page (October 2026) now says $14 billion a year, more than $4.1 billion of it for new initiatives including AI.
- Patent estate: about 8,900 granted patents and pending applications, more than 1,800 of them AI-focused (20% of the portfolio), per the bank's Fast Facts page in October 2026, up from about 8,400 and 1,600 (Banking Dive, June 2026) and from nearly 1,100 AI and machine-learning patents and applications in October 2024; the counts use different scopes in different releases.
- Investor day, 5 November 2025: 270 AI and machine-learning models in use, fraud losses cut in half since 2018, an AI-linked 60% fall in service call volume, and 18,000 developers using coding agents for a 20% productivity gain, per the CTIO.
- Model policy is agnostic: the CTIO said on 1 October 2026 that an orchestration layer called Orchestra sends simple classification to approved open-weight models on the bank's own GPUs and harder reasoning to proprietary models; every AI project passes a review covering 16 risk pillars.
- Spend and return: the CEO said about 140 implemented AI uses cost $400 million and generate $800 million in benefit, and the AI expense budget is expected to double in 2027; the CTIO said employee ideas tagged for AI are more than twice last year's level.
- Third-party rankings place the bank 7th in Evident's AI Index (up from 10th, October 2026) and 19th in Fortune's AIQ 75 (September 2026), behind JPMorgan Chase, Capital One and RBC in the Evident ranking.
- Reuters reported on 8 July 2026 that the bank extended a $520 million credit line to OpenAI, and that it accounted for 60% of AI-related fundraising across investment-grade debt, leveraged finance and equity since 2025, per internal data seen by Reuters.
What is Erica (and EricaAssist)?
The AI virtual assistant launched in 2018 — nearly 50 million users and more than 3 billion interactions by August 2025 — now also runs as EricaAssist, an employee-facing agent that summarises why a client is calling and recommends next steps; the CEO calls it, in effect, a small language model.[2][3][13]
What AI partnerships and patents does Bank of America have?
Bank of America says it holds about 8,900 granted patents and pending applications, more than 1,800 of them AI-focused (Fast Facts, October 2026), and its 2025 and 2026 grants include adversarial machine-learning detection, data poisoning and drift prevention, defensive deepfakes and voice biometrics in a virtual assistant. Its named partners are Salesforce and Zoom for adviser meetings, a five-bank agentic commerce group, and BioCatch for behavioural biometrics. It is model-agnostic, running open-weight models on its own GPUs alongside proprietary ones it does not name, and it lent OpenAI $520 million, per Reuters.[1][51][68][64][65][66][48][55][36][57][52]
| Item | What it is | Status | Date |
|---|---|---|---|
| Salesforce and Zoom | CRM data and meeting platform behind AI-Powered Meeting Journey for Merrill and Private Bank advisers | In production | 2026-03-26[48][49] |
| OpenAI | Borrower: first credit line from the bank, $520 million, per a Reuters source; the bank is also eyeing advisory roles on the OpenAI and Anthropic IPOs | Reported by Reuters | 2026-07-08[52] |
| Open-weight and proprietary models | Orchestra routes simple classification to approved open-weight models on the bank's own GPUs and harder reasoning to proprietary models; suppliers not named | In production | 2026-10-01[57] |
| BERT and OpenNLP | Open-source language models on which the first Erica platform was built, before a second-generation platform in 2025 | In production | 2026-03-17[47] |
| Agentic commerce principles | Joint paper with ASB Bank, Capital One, Commonwealth Bank of Australia, ING and NatWest on transparency, safety, privacy and data, choice and interoperability | Published | 2026-09-22[55] |
| BioCatch | Behavioural biometrics vendor; American Banker names the bank among its users, and the bank has not described its deployment | In production | 2026-03-18[36] |
| US 12,198,074 B2 | Executing a network of chatbots using a parallel bot approach: A query goes to several chatbots at once; each identifies an intent and its own accuracy percentage. | Granted | 2025-01-14[62] |
| US 12,254,314 B2 | Natural language processing enabled CI/CD deployment: Knowledge graphs of file dependencies and known errors for mainframe applications guide automated deployment. | Granted | 2025-03-18[63] |
| US 12,373,551 B2 | Data poisoning and model drift prevention system: Generative adversarial networks create simulated data from trusted history to detect poisoned data and drift. | Granted | 2025-07-29[64] |
| US 12,401,744 B2 | Defensive deepfake for detecting live deepfaked audio and video: Uses deepfakes defensively to detect spoofed phone calls and videoconferences. | Granted | 2025-08-26[65] |
| US 12,425,392 B2 | Voice biometric authentication in a virtual assistant: Compares a spoken command against stored voice biometric signatures before releasing account information. | Granted | 2025-09-23[66] |
| US 12,518,205 B2 | AI/ML model decisioning and rectification system: Lets a user challenge a decision output by an AI or machine-learning model and triggers training-data acquisition. | Granted | 2026-01-06[67] |
| US 12,647,436 B2 | Adversarial machine learning attack detection and prevention system: Monitors an explainable AI model that serves a product or service to detect attacks on it. | Granted | 2026-06-02[68] |
| US 12,705,363 B2 | Vulnerability detection and mitigation using artificial intelligence in an electronic environment: Generates a vulnerability attribute for a network device from source data using AI. | Granted | 2026-08-11[69] |
What has BofA done on AI, and when?
- Oct 13, 2026
Treasury, FinCEN and OCC impose $1.21 billion in AML and sanctions penalties[59][60][61]
FinCEN's $1.2 billion penalty credits the OCC's $750 million assessment; the bank admitted willful Bank Secrecy Act violations, including about $5 billion in unreported suspicious activity, and Treasury says the bank leaned on outside inquiries rather than its internal automated controls. The releases do not mention AI.
- Oct 6, 2026
Evident AI Index: seventh, up from tenth[58]
American Banker reports Evident's 2026 ranking, led by JPMorganChase, Capital One and RBC, puts the bank seventh.
- Oct 1, 2026
CTIO at Fortune AIQ Summit: Orchestra routes work to open-weight or proprietary models[57]
Gopalkrishnan says AI is often rejected in favour of deterministic rules, every project is reviewed against 16 risk pillars, simple classification runs on approved open-weight models on the bank's own GPUs, and agents stay assistive with humans in the loop. He gives Erica's total as 3.6 billion interactions.
- Sep 29, 2026
Corporate RTP use jumps, led by liquidity management[70]
From January through July 2026, corporate transaction volume on the RTP network rose 48% year on year and the number of corporate RTP payments above $1 million rose 351%, which the bank ties to the network's limit rising from $1 million to $10 million in February 2025.
- Sep 29, 2026
Fortune AIQ 75 places the bank 19th; EricaAssist at 18,000 service staff[56]
Fortune ranks the bank 19th behind JPMorgan Chase (1st) and BNY (15th) among US banks listed and reports EricaAssist deployed to more than 18,000 customer service representatives this summer.
- Sep 28, 2026
AskGPS Intelligence Hub and CashPro Payments Insights launch[38]
Global Payments Solutions rolls out Payments Insights, a CashPro feature analysing payment efficiency, cross-border flows and working-capital performance, alongside the AskGPS Intelligence Hub for GPS staff — Intelligent Treasury Management Reviews and Digital Account Schematics that build on the bank's existing AskGPS generative-AI assistant.
- Sep 23, 2026
DeMare: coding is the clearest AI return[37]
At a BofA Securities conference the co-president says about 20,000 developers use coding agents for a 15% to 20% gain, about 95% of staff have AI access, and the bank is assessing third-party providers; Gopalkrishnan says AI investment interest is more than twice last year's and spending will probably double.
- Sep 22, 2026
Principles for AI shopping agents[34]
With NatWest, ING, Capital One, ASB and Commonwealth Bank, the bank warns AI shopping agents raise scam, fraud and data-privacy risks and proposes disclosure when an agent is in a transaction.
- Sep 14, 2026
Moynihan: AI budget to double in 2027[37]
At a Barclays conference, Moynihan says about 140 AI applications are implemented at a cost of $400 million, generating a benefit of $800 million, and that the bank expects to double its AI expense budget next year; headcount has eased from 213,000 to 209,000 through attrition, 'we're not laying off anybody'.
- Sep 14, 2026
Moynihan on agent risk: 'we just don't do that'[20]
After a weekend of public alarm over autonomous AI agents, the CEO says the bank keeps agents under control rather than 'left to operate'.
- Sep 3, 2026
Senior bankers assigned to the 'AI ecosystem'[33]
Gregg Nabhan, Jill Schwartz and Ray Wood work across industry groups to cover companies financing and building around AI, per an internal memo.
- Jul 31, 2026
Q2 2026 10-Q: AI benefits listed as a forward-looking risk[54]
The forward-looking-statements list cites 'the ability to achieve expected or potential benefits, such as increased productivity and cost savings' from artificial intelligence; the Q1 10-Q wording was 'potential benefits'.
- Jul 21, 2026
EricaAssist gets generative AI[3]
Employee-facing agent summarises the call reason and recommends next steps; average call time falls by nearly a minute.
- Jul 17, 2026
New heads of digital assets and AI transformation[19]
Kevin Milsom named head of platforms AI transformation.
- Jul 16, 2026
Moynihan on Mythos: 'a big change'[32]
The CEO joins bank chiefs warning that frontier models such as Anthropic's Mythos change how fast system vulnerabilities must be fixed.
- Jul 15, 2026
Q2 call: 400,000 prompts a day, 300+ use cases[15]
Executives attribute revenue and efficiency improvements to AI; every approved use case has 'good economics'.
- Jul 14, 2026
- Jul 8, 2026
Reuters: $520 million credit line to OpenAI[52]
A person familiar with the matter says the loan is the bank's first to OpenAI and that the bank is eyeing advisory roles on the planned OpenAI and Anthropic IPOs; the bank's internal data put its share of AI-related fundraising at 60% since 2025.
- Jun 12, 2026
About 8,400 patents and applications, 1,600 AI-focused[51]
Cameron Wadley, a wealth-management technology executive with 70 applications filed, says several of his patents underpin Erica and now run in all eight lines of business.
- Jun 8, 2026
Moynihan: 'the data has to be perfect'[16]
The CEO emphasises model accuracy over efficiency and cost savings.
- Jun 3, 2026
4,000 interns and campus hires despite AI[11]
2,000 full-time recruits and 2,000 internships, a signal on entry-level roles.
- Apr 15, 2026
CTIO at Semafor: from proofs of concept to end-to-end process transformation[50]
Gopalkrishnan names four drivers: end-to-end process change, scale and reuse, governance, and return on investment; the bank supports about 3,000 processes and filled 44% of jobs through internal mobility.
- Apr 15, 2026
Q1 2026 call: all 200,000 teammates have AI access[27]
Moynihan cites 90 AI installations working and says the bank will be 'more agentic versus prompt' as its 90,000-strong sales force moves to agent tools.
- Mar 26, 2026
- Mar 17, 2026
'Build once' strategy explained: second-generation Erica platform in 2025[47]
Fortune reports that an AI council with the head of strategy ran about 15 proofs of concept outside the Erica platform, found teams 'reinventing the wheel', and the bank rebuilt Erica as a second-generation, model-agnostic platform in 2025; the first generation started on BERT and OpenNLP.
- Mar 10, 2026
30 billion digital client interactions in 2025[30]
20.6 million users interacted with Erica nearly 700 million times in 2025; cumulative Erica interactions passed 3.2 billion.
- Feb 25, 2026
2025 10-K: autonomous AI agents and third-party models as risks[21]
The annual report says expanded use of AI, 'including emerging third-party AI services and autonomous AI agents', may bring data risk, unpredictable system interactions and AI failure, and notes internal AI governance frameworks.
- Jan 22, 2026
Erica 'does the work of 11,000 people'[13]
Investor-day disclosure: 20 million regular users, 2 million interactions a day.
- Jan 14, 2026
Q4 2025 call: coding 30% faster, Copilot for 200,000[26]
Moynihan says AI took 30% out of the coding stream for the bank's 18,000 coders, saving about 2,000 people, that Microsoft 365 Copilot is out to 200,000 teammates, and that direct AI spend is 'several hundred million dollars'.
- Nov 17, 2025
CTIO: one banker can cover 50 clients instead of 15[46]
At the Reuters AI Momentum conference Gopalkrishnan ties AI to revenue and cost: meeting-preparation tools raise banker coverage, and AI has analysed person-to-person payments for fraud 7 million times.
- Nov 17, 2025
AI is boosting bankers' productivity and revenue[10]
The CTIO says the bank will spend billions on AI to raise banker productivity and bring in revenue.
- Nov 5, 2025
Investor day: 270 models, fraud losses halved, 18,000 coders on agents[43][44][45]
Moynihan says the bank runs 270 AI and machine-learning models; the CTIO reports 18,000 developers on coding agents and a 20% productivity gain; Holly O'Neill says AI-enabled processes cut fraud losses in half since 2018. Technology spend is about $100 billion over a decade, $4 billion a year on new initiatives and $1 billion on cybersecurity.
- Oct 28, 2025
Datos Insights names CashPro Data Intelligence best AI innovation[42]
The suite of CashPro Search, Forecasting and Insights wins the 2025 Datos Impact Award for AI; Merrill SENSE, a large-language-model search engine for marketing teams, wins a second award.
- Oct 27, 2025
CashPro Chat usage at a record[31]
Client use of the AI-driven chat up 21% year on year, with nearly 70% of corporate clients using it; 43% containment rate in Q3 2025.
- Oct 15, 2025
Q3 2025 call: headcount down from 285,000 to 213,000[25]
Moynihan credits applied technology, including AI, for 15 years of headcount reduction and says Erica already handles millions of interactions a day, while stressing precise implementation in a regulated business.
- Oct 10, 2025
AskGPS: 3,200 documents, 29 languages[41]
In-house engineers fed the Global Payments Solutions assistant 3,200 internal documents; Jarrett Bruhn says deep questions that took up to an hour of searching are now answered directly.
- Sep 30, 2025
CashPro gets a generative-AI assistant[5]
The bank restates its four AI domains: intelligent agents, search and summarisation, content generation, operations and coding.
- Sep 29, 2025
AI-driven TED score added to the CashPro App[40]
Capital Markets Insights gives investment-grade issuers an AI-driven score of the macro backdrop on a given day, inside an app used by more than 40,000 corporate and commercial clients.
- Sep 12, 2025
Co-presidents named, AI tools in their brief[24]
Dean Athanasia and Jim DeMare become co-presidents; the CEO lists 'the continued expansion of AI-based tools and innovation for our clients' among the company-wide efforts they will drive.
- Aug 20, 2025
$13 billion tech agenda, $4 billion for new capabilities[9]
New CTIO Gopalkrishnan tells the Wall Street Journal to expect a steady stream of AI releases, doubling down on high-value use cases.
- Aug 1, 2025
Erica passes 3 billion interactions[2]
Nearly 50 million users since launch; more than 58 million interactions a month.
- Jul 29, 2025
Gopalkrishnan replaces Bhasin as CTIO[39]
Aditya Bhasin, CTIO since 2021, leaves for a role advising and investing in small and mid-sized companies. Moynihan's memo says Hari Gopalkrishnan 'leads our efforts to develop, deliver, and drive expanded AI-based services and solutions' with Vice Chair Thong Nguyen, and cites 7,800 patent filings and Erica's 3 billion interactions.
- Jul 16, 2025
Q2 2025 deck maps AI across four domains[23]
Erica for Employees used by more than 90% of teammates, cutting service-desk calls by about half; an AI assistant for about 17,000 developers; more than 50 AI-enabled fraud-detection models; about 1,400 AI and machine-learning patents granted and pending.
- Apr 1, 2025
Workforce AI adoption: 1 million training simulations[6]
The Academy uses AI conversation simulators for coaching; employees completed more than a million simulations in a year.
- Oct 1, 2024
AI and ML patents up 94% since 2022[7]
Nearly 1,100 AI and machine-learning patents and pending applications.
- Jun 1, 2018
Erica launches[2]
The AI-driven virtual financial assistant that the bank calls a central gateway to serving clients.
Where does BofA use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| Erica[13] | Consumer virtual assistant; 2 million interactions a day. | In production | Customer-facing chatbots |
| EricaAssist and employee tools[3][15] | Human-assisted AI agent for customer-service staff; Merrill advisor meeting journey; 400,000 prompts a day across the workforce. | In production | Generative & agentic AI |
| CashPro Forecasting and Data Intelligence[8] | AI cash forecasting for corporate clients; 3,000 companies, 250,000 hours saved in 2025. | In production | Generative & agentic AI |
| The Academy simulators[6] | AI conversation simulators for employee coaching. | In production | AI governance (general) |
| Frontier-model testing[12] | Part of Anthropic's Glasswing programme from the start; testing Mythos internally, per Reuters sources. | Pilot | Cybersecurity |
| AI-assisted software development[22][26] | About 19,000 developers use real-time coding assistance, raising productivity by more than 20%; AI took 30% out of the coding stream in 2025. | In production | Generative & agentic AI |
| AI fraud-detection models[23] | More than 50 AI-enabled fraud-detection models designed to detect fraudulent activity more quickly, accurately and at scale. | In production | Fraud detection |
| CashPro Chat for corporate clients[31][29] | AI-driven chat used by nearly 70% of corporate clients for account information, transaction tracking and service resolution, with a 43% containment rate in Q3 2025; the Q2 2026 earnings release counts 40,100 CashPro Chat interactions supported by Erica technology. | In production | Customer-facing chatbots |
| AI risk factors and model governance[21] | The 10-K covers autonomous AI agents, third-party AI models with limited visibility into training data, inaccurate outputs and bias, and says models may be ineffective if AI applications 'are not developed and trained properly or contain biases'. | In production | Model risk management |
| AskGPS Intelligence Hub and Payments Insights[38] | CashPro's Payments Insights analyses payment efficiency, cross-border flows and working-capital performance for corporate clients; the AskGPS Intelligence Hub gives Global Payments Solutions staff Intelligent Treasury Management Reviews and Digital Account Schematics built on the bank's AskGPS generative-AI assistant. | Rolling out | Generative & agentic AI |
| Behavioural biometrics (BioCatch)[36] | American Banker (March 2026) names Bank of America among banks using BioCatch, whose technology reads how a user logs in, the device, the angle the phone is held at, keystrokes and past behaviour to decide whether the user is a real customer, a fraudster or a bot; Bank of America has not described its deployment. | In production | Fraud detection |
| Person-to-person payment fraud analysis[46][44] | AI has analysed person-to-person payments for fraud 7 million times, per the CTIO in November 2025; the bank says AI-enabled processes cut fraud losses in half since 2018. | In production | Fraud detection |
| AI-Powered Meeting Journey[48][49] | Meeting preparation, consented AI note-taking and follow-up tasks for Merrill and Private Bank advisers, built on Salesforce CRM data and Zoom; the bank says it can save up to four hours per meeting. | In production | Generative & agentic AI |
| AskGPS knowledge assistant[41] | Global Payments Solutions assistant trained on 3,200 internal documents, speaks 29 languages and answers questions for staff serving 40,000-plus business clients. | In production | Generative & agentic AI |
| Capital Markets Insights and the TED score[40] | An AI-driven algorithm on market variables produces a score of the macro backdrop for investment-grade issuance on a given day, shown in the CashPro App alongside same-day and monthly issuance data. | In production | Trading & capital markets |
| Orchestra model-routing layer[57] | Sends simple classification tasks to approved open-weight models on the bank's own GPUs and harder reasoning to proprietary models, which the CTIO says also controls token costs. | In production | Generative & agentic AI |
| Sixteen-pillar AI project review[57] | Every AI project is reviewed against 16 risk pillars including privacy, bias, workforce impact and intellectual property; the CTIO says the bank often decides against AI where rules or an app are enough. | In production | AI governance (general) |
| EricaAssist at the contact centre[53][56] | Desktop widget used by about 18,000 customer service employees; generative-AI additions tell the agent why the customer is calling and recommend next steps, and average call time falls by almost a minute. | In production | Customer-facing chatbots |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
Treasury tools compared across banks: AI in bank treasury management.
Which AI vendors and fintechs does BofA work with?
| Company | What BofA uses it for | Status | Category |
|---|---|---|---|
| BioCatch | Behavioral biometrics to tell genuine customers from fraudsters and bots in digital banking sessions (named as a user by American Banker, March 2026). | In production | Fraud detection |
| Microsoft 365 Copilot | Microsoft 365 Copilot rolled out to a total of 200,000 teammates, per the CEO on the Q4 2025 earnings call. | In production | Enterprise AI assistants |
| Claude (Anthropic) | Part of Project Glasswing and testing Mythos internally, according to a Reuters source. | Pilot | Enterprise AI assistants |
Each vendor and fintech page cites the source for the relationship, and vendor pages list products, pricing and other bank customers. Compare AI tools for banks by category, or read the fintech and neobank AI dossiers.
What has BofA disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Corporate RTP transaction volume (Jan–Jul 2026 vs 2025) | +48%[70] | Jul 31, 2026 |
| Corporate RTP transactions above $1 million (Jan–Jul 2026 vs 2025) | +351%[70] | Jul 31, 2026 |
| Technology budget | $13 billion, of which $4 billion for new technology[9][10] | Aug 20, 2025 |
| Erica interactions per day | 2 million[13] | Jan 22, 2026 |
| Employee prompts | 400,000+ a day[15] | Jul 15, 2026 |
| Approved AI use cases | 300+[15] | Jul 15, 2026 |
| AI/ML patents and applications | ≈ 1,100[7] | Oct 1, 2024 |
| Digital share of client interactions | 94%[18] | Jun 1, 2026 |
| Active users of general-purpose AI tools | ~200,000[22] | Jun 30, 2026 |
| Generative-AI use cases | 114 live, 34 fully deployed[22][28] | Jun 30, 2026 |
| Developers using AI coding assistance | ~19,000; productivity up more than 20%[22] | Jun 30, 2026 |
| AI training courses completed | 600,000+[22] | Jun 30, 2026 |
| Active Erica users | 24 million+[28] | Jun 30, 2026 |
| Direct AI spend (CEO estimate) | several hundred million dollars[26] | Jan 14, 2026 |
| Implemented AI applications | ≈ 140, costing $400 million, generating $800 million in benefit[37] | Sep 14, 2026 |
| AI expense budget, 2027 plan | expected to double versus 2026[37] | Sep 14, 2026 |
| Patents and pending applications, total vs AI-focused | ≈ 8,900 total; > 1,800 AI-focused (20%)[1] | Oct 19, 2026 |
| AI and machine-learning models in use | 270[45] | Nov 5, 2025 |
| Fraud losses since 2018 | cut in half through AI-enabled processes[44] | Nov 5, 2025 |
| Technology budget, 2026 | $13.5 billion a year, of which ≈ $4 billion for new initiatives[48][50] | Mar 26, 2026 |
| EricaAssist users | ≈ 18,000 customer service employees[53] | Jul 21, 2026 |
| Developers on coding agents | ≈ 20,000; 15% to 20% productivity gain[37] | Sep 23, 2026 |
| Banker client coverage with AI preparation | 50 clients instead of 15 (CTIO)[46] | Nov 17, 2025 |
| Person-to-person payments analysed for fraud by AI | 7 million[46] | Nov 17, 2025 |
| Erica cumulative interactions | 3.8 billion since 2018; nearly 200 million a quarter[1] | Oct 19, 2026 |
| Technology spend (latest) | $14 billion a year, > $4.1 billion of it for new initiatives including AI[1] | Oct 19, 2026 |
| Credit line extended to OpenAI (Reuters source) | $520 million[52] | Jul 8, 2026 |
| Evident AI Index rank | 7th (10th a year earlier)[58] | Oct 6, 2026 |
“More than 300 AI use cases, all of which have good economics.”[15]
Brian Moynihan · Chairman and CEO · Jul 15, 2026
“We have zero tolerance for things like hallucinations, bias.”[17]
Hari Gopalkrishnan · CTIO · Jun 1, 2026
“This weekend was a lot about the risk in AI, and that largely is around agents just left to operate. We just don't do that.”[20]
Brian Moynihan · Chairman and CEO · Sep 14, 2026
“Our associates are generating more than 400,000 prompts a day, and as of last week, we had over 300 AI use cases approved, all of which have good economics, of which 114 are live generative AI use cases.”[28]
Brian Moynihan · Chair and CEO · Jul 14, 2026
“It is a big change in the amount of work that will have to go on and the pace of which these tools will affect vulnerabilities in your systems, and how fast you have to take care of that.”[32]
Brian Moynihan · Chair and CEO · Jul 16, 2026
“We're not laying off anybody. We don't have to do that.”[37]
Brian Moynihan · Chair and CEO · Sep 14, 2026
“One of the biggest mistakes we see us and others doing is rush to AI as a solution.”[57]
Hari Gopalkrishnan · CTIO · Oct 1, 2026
“There is so much juice to be squeezed right now with assistive agents that are actually working with humans in the loop.”[57]
Hari Gopalkrishnan · CTIO · Oct 1, 2026
“If you overdo it, you stall innovation. If you underdo it, you introduce a lot of risk.”[50]
Hari Gopalkrishnan · CTIO · Apr 15, 2026
“You now can have a banker cover 50 clients instead of 15, and that's exactly what we're seeing in the real world.”[46]
Hari Gopalkrishnan · CTIO · Nov 17, 2025
“One of the biggest risks to implementation of AI is people being fearful of it and thinking that it's going to replace them.”[37]
Jim DeMare · Co-President · Sep 23, 2026
Who runs AI at BofA?
Hari GopalkrishnanLinkedIn ↗
Nikki KatzLinkedIn ↗
Head of Digital[18]
Kevin Milsom
Head of Platforms AI Transformation[19]
Dean Athanasia
Co-President (company-wide initiatives including the expansion of AI-based tools)[24]
Jim DeMare
Co-President (company-wide initiatives including the expansion of AI-based tools)[24]
Thong Nguyen
Vice Chair; leads AI-based services and solutions with the CTIO, per the CEO's July 2025 memo[39]
Aditya Bhasin
Former Chief Technology and Information Officer (2021 to July 2025)[39]
Holly O'Neill
President, Consumer, Retail and Preferred Banking (Erica owner; reported the 2 million daily interactions and halved fraud losses)[43][44]
Tom Ellis
Chief Information Officer and Head of Consumer Technology (EricaAssist)[53]
Ashley Ross
Head of Consumer Client Experience and Business Transformation (EricaAssist)[53]
Jarrett Bruhn
Head of Data and AI, Global Payments Solutions (AskGPS)[41]
Mark Monaco
Head of Global Payments Solutions (agentic commerce principles; $1 billion annual payments technology spend)[55][44]
Cameron Wadley
Supervision technology executive, wealth management; inventor on Erica-related patents (70 applications filed)[51]
Shimna Sameer
Head of Products, Solutions and Platforms, Bank of America Private Bank (AI-Powered Meeting Journey)[48]
Patricio Diaz
Chief Operating Officer, Merrill (AI-Powered Meeting Journey)[48]
Bernard Mensah
President of International and CEO, Merrill Lynch International (AI governance across jurisdictions)[45]
What AI roles is BofA hiring for?
Data as of · 12 AI-titled postings · read from BofA's own careers site, re-checked weekly
12 open AI-related postings on BofA's careers site as of Oct 25, 2026. A snapshot of current openings, not a source: postings close. Compare with other banks on which banks are hiring for AI.
Senior Principal Engineer (Generative AI Strategy)
New York · posted
Charlotte · posted
Cyber Threat Defense Sr AI/ML Engineer
Chicago (+4 more) · posted
Senior Engineer - GenAI Security
New York (+3 more) · posted
Artificial Intelligence Infrastructure Product Manager
Chandler (+3 more)
Agile Product Owner – Document AI
Charlotte (+9 more) · posted
Business Control Manager - AI and Policy Governance and Oversight
Newark (+7 more) · posted
Charlotte
AI Engineer - Data Scientist I
New York (+1 more) · posted
Which regulators govern BofA's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| OCC | National bank; revised model risk guidance in force, with generative AI governed outside it. | OCC Bulletin 2026-13 |
| CFPB | Erica is the largest consumer bank chatbot in the US — the CFPB's chatbot spotlight names the bank by user count. | CFPB Chatbots in Consumer Finance (issue spotlight, 2023) ECOA / Regulation B adverse action (15 U.S.C. 1691(d); 12 CFR 1002.9) |
| Federal Reserve | Holding-company supervisor; third-party guidance governs the models behind EricaAssist and advisor tools. | SR 26-2 SR 23-4 |
| FSB | G-SIB; the FSB's sound practices on human oversight and data governance map onto the bank's stated posture. | FSB AI sound practices consultation (June 2026) |
Side by side: SR 11-7 vs SR 26-2 · OCC vs CFPB on AI lending · PRA SS1/23 vs SR 26-2 · MAS AI Risk Management Guidelines vs SR 26-2 · all comparisons.
What does the public record suggest about BofA's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
A decade of one assistant is the moat
Erica's seven years of interaction data is what makes EricaAssist and the advisor tools credible quickly. Competitors launching gen-AI chatbots in 2026 start without that corpus.
02
Accuracy-first is a regulatory posture as much as an engineering one
'Zero tolerance for hallucinations' and 'the data has to be perfect' are the language of UDAAP and adverse-action risk. Expect the bank to keep customer-facing generative AI narrower than peers.
03
The metrics to watch are hours and interactions, not headcount
The bank reports 250,000 client hours saved and calls handled, and keeps hiring at the entry level. If that changes, the strategy has changed.
04
Corporate treasury is the quiet AI product line
CashPro Forecasting and Data Intelligence are AI products sold to 3,000-plus corporate clients — a revenue line, not a cost line.
Where does the BofA record go deeper?
6 topic pages under this profile, each answering one question from the same sourced record.
- What is Bank of America's AI strategy and technology vision?
- What AI research, patents and partnerships does Bank of America have?
- What is Erica, Bank of America's virtual assistant, and how big is it?
- What technology stack does Bank of America run its AI on?
- How does Bank of America govern responsible AI?
- What proprietary AI has Bank of America built in-house?
- Has Bank of America had a publicly reported AI incident?
Which leader rankings does BofA appear in?
13 of the computed leaders rankings include BofA, each ranked from the same sourced bank pages.
How does BofA compare with its peers?
- How does Bank of America compare with its peers on AI and technology?
- Bank of America competitors and alternatives, by business line
Its peer group on this site, by business model and size:
- Can AI avoid the boom-and-bust fate of past capital cycles? BofA weighs in
- BofA Deploys Workers to Patch Vulnerabilities Found by Front…
- Bank of America hired contractors to protect software from rogue AI
- AI, Deal Pipelines and BofA’s Punishment Dominate Earnings Week
- Bank of America had to hire 1,000 contractors to patch software flaws spotted by Anthropic Mythos
What is Erica at Bank of America?
Erica is Bank of America's AI virtual financial assistant, launched in 2018. By August 2025 it had assisted nearly 50 million users and passed 3 billion interactions; by the January 2026 investor day it handled 2 million interactions a day, which the bank says is the work of 11,000 people; and the bank's Fast Facts page (October 2026) counts 3.8 billion interactions since launch and nearly 200 million a quarter. EricaAssist is the employee-facing version that supports customer-service staff during calls.
How much does Bank of America spend on technology and AI?
The bank's Fast Facts page (October 2026) puts technology spend at $14 billion a year, more than $4.1 billion of it allocated to new initiatives including AI. In 2025 the CTIO put it at about $13 billion, of which $4 billion was set aside for new capabilities including AI (Wall Street Journal and Reuters). The bank does not publish an AI-only budget.
How many Bank of America employees use AI?
About 200,000 teammates were active users of general-purpose AI productivity tools at the end of June 2026, generating more than 400,000 prompts a day, according to the Q2 2026 earnings presentation. The bank reported more than 300 approved AI and machine-learning use cases, 114 live generative-AI use cases with 34 fully deployed, and more than 600,000 AI training courses completed.
Who leads AI at Bank of America?
Hari Gopalkrishnan, chief technology and information officer since July 2025, leads the bank's AI work. Brian Moynihan's memo said he drives expanded AI-based services in partnership with Vice Chair Thong Nguyen. He replaced Aditya Bhasin, CTIO since 2021, who left to advise and invest in small and mid-sized companies. Business owners include Holly O'Neill (consumer banking and Erica), Tom Ellis and Ashley Ross (EricaAssist), Jarrett Bruhn (data and AI in Global Payments Solutions) and Kevin Milsom (platforms AI transformation). Co-presidents Dean Athanasia and Jim DeMare carry AI expansion in their brief. The sources reviewed name no separate chief AI officer.
How many AI patents does Bank of America hold?
The bank's Fast Facts page (October 2026) says it holds about 8,900 granted patents and pending applications, more than 1,800 of them AI-focused (20% of the portfolio), more than half already granted. Banking Dive reported about 8,400 and 1,600 in June 2026. Earlier counts were smaller and scoped differently: nearly 1,100 AI and machine-learning patents and applications in October 2024, up 94% since 2022, and about 1,400 granted and pending in the Q2 2025 earnings deck. Granted US patents from 2025 and 2026 cover adversarial machine-learning attack detection, data poisoning and model drift prevention, defensive deepfakes, voice biometrics in a virtual assistant, and AI model decision rectification.
Which AI models does Bank of America use?
The bank says it is model-agnostic. Erica began on open-source language models such as BERT and OpenNLP, and a second-generation platform was built in 2025. On 1 October 2026 the CTIO said an orchestration layer called Orchestra sends simple classification to approved open-weight models on the bank's own GPUs and harder reasoning to proprietary models. The sources reviewed do not name the proprietary suppliers. The bank ran 270 AI and machine-learning models in November 2025, and Reuters reported in July 2026 that it extended OpenAI a $520 million credit line.
- Bank of America Fast Facts (undated page, retrieved 2026-10-19)
- A Decade of AI Innovation: BofA's Virtual Assistant Erica Surpasses 3 Billion Client Interactions
- Bank of America Enhances EricaAssist with Generative AI
- Merrill and Bank of America Private Bank Launch AI-Powered Meeting Journey
- BofA's New GenAI Assistant Transforms Global Payments Solutions
- AI Adoption by BofA's Global Workforce Improves Productivity, Client Service
- AI Patents at BofA Increase 94% Since 2022
- BofA's AI Solution CashPro Forecasting Helps Clients Navigate Year of Volatility
- The Man Behind Bank of America's $13 Billion Tech Agenda
- BofA says AI is boosting bankers' productivity, revenue
- Bank of America to hire interns, campus recruits despite AI threat
- Banking industry scrambles for Anthropic's Mythos as global regulators review risks
- How Bank of America's Erica does the work of 11,000 people
- Bank of America gives Merrill Lynch an AI makeover
- Bank of America's Moynihan touts 'good economics' of AI
- 'The data has to be perfect': BofA CEO Moynihan on AI
- Bank of America's CTO champions a disciplined approach to AI
- Nikki Katz's 'Big Year of Scale' at Bank of America
- BofA Promotes Leaders to Drive Adoption of Crypto, AI Technology
- BofA and PNC respond to rising AI fears
- Bank of America Corporation Annual Report on Form 10-K for the year ended December 31, 2025
- Second Quarter 2026 Financial Results presentation (Form 8-K, Exhibit 99.2)
- Second Quarter 2025 Financial Results presentation (Form 8-K, Exhibit 99.2)
- Bank of America Announces Senior Leadership Changes (Form 8-K, Exhibit 99.1)
- Bank of America Corp (BAC) Q3 2025 Earnings Call Highlights
- BofA (BAC) Q4 2025 Earnings Call Transcript
- BofA (BAC) Q1 2026 Earnings Call Transcript
- Bank of America (BAC) Q2 2026 Earnings Call Transcript
- Bank of America Reports Second Quarter 2026 Financial Results (Form 8-K, Exhibit 99.1)
- BofA AI and Digital Innovations Fuel 30 Billion Client Interactions
- BofA's AI-Driven CashPro Chat and Transaction Search Capabilities Reach Record Usage by Corporate Clients
- BofA's Moynihan Joins Bank CEOs Warning About Mythos AI Risks
- BofA Taps Senior Bankers to Sharpen Focus On 'AI Ecosystem'
- Banks warn AI shopping bots raise scam, fraud and data-privacy risks
- Banks report operational changes driven by AI adoption
- Agentic AI shopping bots are coming. Banks need to be ready
- Bank of America plans to double AI budget next year
- BofA Debuts Intelligence Tools for CashPro Clients and GPS Staff
- BofA taps next tech chief
- BofA Unveils AI-Driven Capital Markets Insights on CashPro App
- Bank of America fires up generative AI payments assistant
- BofA AI and Digital Capabilities Win Four Best Innovation Awards
- BofA's Moynihan pledges growth, 'no excuses'
- BofA co-presidents pledge rigor on investor day targets
- Bank of America runs 270 AI models across operations
- 'Revenue opportunities everywhere': BofA's tech chief on AI
- Inside Bank of America's 'build once' AI strategy
- Merrill and Bank of America Private Bank Launch AI-Powered Meeting Journey
- Bank of America's wealth management firms roll out AI tool
- Bank of America tech chief shares AI strategy focus
- No 'designated geniuses': How BofA democratizes innovation
- BofA extends first $520 million loan to OpenAI ahead of IPO, source says
- Bank of America upgrades customer service employee tool
- Bank of America Corporation Quarterly Report on Form 10-Q for the quarter ended June 30, 2026
- Global Banks Collaborate on Principles for Trusted Agentic Commerce
- How bottom-line-focused financial firms are finding ROI in AI
- AI in regulated industries: BofA and S&P on the huge gains to be had, and the risks of rushing in
- JPMorganChase still tops AI ranking as big-bank lead widens
- U.S. Treasury Announces Record Enforcement Actions Against Bank of America
- OCC Imposes $750 Million Assessment Against Bank of America for Deficient BSA/AML Compliance
- BofA fined $1.21 billion over AML
- US 12,198,074 B2: Executing a network of chatbots using a parallel bot approach (Bank of America Corporation)
- US 12,254,314 B2: Natural language processing enabled CI/CD deployment (Bank of America Corporation)
- US 12,373,551 B2: Data poisoning and model drift prevention system (Bank of America Corporation)
- US 12,401,744 B2: Defensive deepfake for detecting live deepfaked audio and video (Bank of America Corporation)
- US 12,425,392 B2: Voice biometric authentication in a virtual assistant (Bank of America Corporation)
- US 12,518,205 B2: AI/ML model decisioning and rectification system (Bank of America Corporation)
- US 12,647,436 B2: Adversarial machine learning attack detection and prevention system (Bank of America Corporation)
- US 12,705,363 B2: Vulnerability detection and mitigation using artificial intelligence in an electronic environment (Bank of America Corporation)
- BofA Clients Accelerate RTP Use, Led by Liquidity Management
When BofA moves on AI, you'll read it here first.
when one of the banks we cover moves on AI, it is in the next morning's brief · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning