A top-three bank worldwide on the Evident AI Index four years running, with its own research institute (RBC Borealis), a proprietary transaction foundation model (ATOM), a C$700 million to C$1 billion AI value target for 2027 and, since February 2026, an AI Group reporting to the CEO.
How is RBC using AI?
Royal Bank of Canada (RBC), which reported C$2,498,817 million of total assets at 31 July 2026, is the only Canadian bank in the top ten of the Evident AI Index and ranked third of 50 global banks for AI maturity in 2025, its fourth consecutive top-three finish, behind JPMorgan Chase and Capital One. Its research institute RBC Borealis, launched in 2016, built the Lumina data and AI platform and ATOM, a proprietary foundation model trained on billions of client transactions and used across 15 products and processes in 2025. At its March 2025 Investor Day RBC set a target of C$700 million to C$1 billion of enterprise value from AI by 2027, measured net of investments, and on 3 December 2025 and again on 27 August 2026 management said it was on track. In February 2026 it created an AI Group led by Bruce Ross that reports to CEO Dave McKay. By May 2026 McKay told investors RBC had built more than 200 AI models and that large-language-model token usage had risen more than 500% since 2025. Almost 27,000 employees use the RBC Assist assistant, and the bank has partnered with Cohere (North for Banking), NVIDIA and, per American Banker, Anthropic. McKay said in June 2026 that most deployment is revenue enablement rather than back-office cost takeout. The bank's US arm, City National, is covered at /banks/city-national.
Lead bank
Royal Bank of Canada
Total assets (group, as reported, Jul 31, 2026)
C$2,498,817 million total assets[2]
Charter
Canadian Schedule I bank; federally regulated financial institution supervised by the Office of the Superintendent of Financial Institutions (OSFI)
Headquarters
Toronto, Canada
Flagship platform
Lumina, ATOM and RBC Assist
AI leadership
Dave McKay · Bruce Ross · Naim Kazmi · Foteini Agrafioti · Bobby Grubert · Lindsay Patrick · Neil McLaughlin · Chris Phillips · Gopala Narayanan · Kelly Bradley
- Ranked #3 of 50 banks in the 2025 Evident AI Index (published 8 October 2025), the fourth straight year in the top three, #1 in Canada and the only Canadian bank in the top ten; it also took the global #1 spot for AI talent development, according to RBC's release.
- Value target: C$700 million to C$1 billion of enterprise value generated from AI by 2027, defined as the annualised exit run-rate benefit in revenue, cost avoidance, expense savings, fraud reduction and lower cost of risk, net of investments, and incremental from fiscal 2024; the bank said on 3 December 2025 it was on track.
- Own research arm: RBC Borealis, launched in 2016, built Lumina (more than one billion business events, up to 10 billion transactions per minute) and ATOM, a foundation model trained on billions of client transactions and used across 15 RBC products and processes in 2025.
- Organisation: an AI Group created on 18 February 2026 reports to the CEO under Bruce Ross, previously Group Head of Technology & Operations for 12 years, and is meant to move AI projects from early stage to scaled client outcomes.
- Scale of use: almost 27,000 employees on RBC Assist and more than 8,000 Capital Markets staff on Aiden (February 2026); more than 200 AI models and a more than 500% rise in LLM token usage since 2025 (Q2 2026 call).
- Credit is the clearest production case: AI-assisted retail credit underwriting built on ATOM, which RBC says improved approval rates, shortened decision times and raised client acquisition, won an IDC CIO Award in August 2026.
- Partners named in RBC's own and press accounts include Cohere (North for Banking, January 2025), NVIDIA (agentic AI for Aiden), the Vector Institute (platinum sponsorship renewed to 2032) and MIT's FinTechAI@CSAIL.
- Its US bank is covered separately at /banks/city-national; the Canadian peer with a comparable AI value target is covered at /banks/td-bank-group.
What is Lumina, ATOM and RBC Assist?
RBC Lumina is the bank's internal enterprise data and AI platform, built by RBC Borealis. RBC says it collects and curates more than one billion business events and can analyse up to 10 billion transactions per minute, and that it runs on one of the largest GPU clusters among Canadian financial institutions. ATOM (Asynchronous Temporal Model), the bank's proprietary foundation model, was trained on large-scale financial datasets including billions of client transactions, was first used in 2023 and was applied across 15 RBC products and processes in 2025, including credit adjudication and personalisation. On top sit employee-facing tools: RBC Assist, an internal AI assistant that almost 27,000 employees were using in February 2026, and Aiden, the AI platform used by more than 8,000 Capital Markets employees and built with Borealis AI on NVIDIA AI Enterprise. In February 2026 the bank also described a co-developed enterprise generative AI product with Cohere, North for Banking.[13][10][15][1][18]
What has RBC done on AI, and when?
- Sep 9, 2026
McKay prefers AI to acquisitions[24]
At the Scotiabank Financials Summit the CEO says transforming the business with AI will drive 'by far the greatest shareholder return' and that making an acquisition would distract the bank right now.
- Aug 27, 2026
- Aug 13, 2026
IDC CIO Award for AI-assisted credit underwriting[20]
RBC's retail credit underwriting programme, built on ATOM, is recognised; early results include improved approval rates, faster decisions and increased client acquisition.
- Jul 30, 2026
Vector Institute sponsorship renewed to 2032[21]
RBC, a founding sponsor since 2017, renews its platinum sponsorship for five years; it cites 30 applied AI projects and 200 Vector-affiliated AI specialists recruited.
- Jul 16, 2026
OSFI bulletin on generative and agentic AI[26]
OSFI sets out sound practices for generative and agentic AI under its technology, operational-resilience and third-party guidelines (B-13, E-21, B-10), pointing to E-23 for model risk.
- Jun 16, 2026
McKay: AI is revenue enablement, not back-office cost takeout[19]
At a Bloomberg event the CEO says he reads an AI-generated report each morning and that most AI deployment is aimed at adding customers (target 25 million from 17 million), with workforce shrinkage expected through attrition.
- May 28, 2026
Q2 2026 call: 200-plus models, token use up 500%[7]
McKay reports more than 200 AI models, a more than 500% rise in LLM token usage since 2025, and AI contributions to over 24 million lines of code and over 120,000 code reviews.
- Feb 26, 2026
Q1 2026 call: AI benefits not yet visible in results[6]
McKay says the C$700 million to C$1 billion AI benefits have not yet shown up in results because the money has been invested, and that the bank remains on track.
- Feb 18, 2026
- Dec 3, 2025
Q4 2025: target 'net of investments', on track[9][1]
McKay says the AI value target is net of investments already made in data storage, GPU clusters, proprietary LLMs, risk governance and people; RBC Assist has launched to over 30,000 employees and AI has contributed over 5 million lines of code. The same day's annual report says AI is a key area of Board oversight.
- Oct 8, 2025
- Sep 11, 2025
OSFI publishes Guideline E-23 on model risk[25]
Canada's prudential supervisor finalises its model-risk guideline, effective 1 May 2027 for federally regulated institutions, with added clarity for AI and machine-learning models.
- Jul 29, 2025
ATOM foundation model disclosed[13]
RBC announces ATOM, built by RBC Borealis on billions of client transactions and part of Lumina; first used in 2023 and already behind several products, including credit adjudication.
- Jul 18, 2025
Joins MIT FinTechAI@CSAIL[22]
A three-year membership gives RBC early access to research in machine learning, secure computation, cybersecurity and data science, plus access to talent.
- Mar 27, 2025
- Jan 9, 2025
Cohere partnership for North for Banking[14]
RBC and Cohere announce they will co-develop and securely deploy North for Banking, an enterprise generative AI platform integrating RBC's and Cohere's foundation models and RBC's internal platforms.
Where does RBC use AI today?
| System | What it does | Status | Use case |
|---|---|---|---|
| ATOM-based retail credit adjudication[13][20] | ATOM supports credit adjudication using transaction histories and non-traditional data, including extending credit to clients with less credit history such as newcomers; in August 2026 RBC cited improved approval rates, faster decisions and increased client acquisition. | In production | Credit scoring & underwriting |
| AI agents in commercial and wholesale credit[7][3] | Clients' financials are ingested and spread using AI (Q2 2026 call); the December 2025 strategic update lists improved document-ingestion accuracy, shorter credit-memo preparation and auto-decision targets for small business, delivered through AI agents launching throughout 2026. | Rolling out | Credit scoring & underwriting |
| RBC Assist employee assistant[9][15] | Internal AI assistant launched to over 30,000 employees (December 2025); almost 27,000 using it by February 2026. | In production | Generative & agentic AI |
| North for Banking with Cohere[14][1] | Enterprise generative AI platform co-developed with Cohere, integrating RBC's and Cohere's foundation models and RBC's internal platforms. | Announced | Generative & agentic AI |
| Advice Centre generative AI and conversational banking[3][7][1] | GenAI supports Personal Banking advisors with knowledge access and task execution; about 25% of clients interacting with the Advice Centre did so through conversational banking in 2025, and an AI-powered search of policy and procedure articles handles about two million searches a month. | In production | Generative & agentic AI |
| Aiden: electronic trading and AidenResearch[18][16][17] | Aiden began as a deep reinforcement learning electronic trading platform and now includes AidenResearch, agents that generate earnings content, summarise calls and give real-time access to filings, built with Borealis AI on NVIDIA AI Enterprise and used by more than 8,000 Capital Markets employees; AidenBanker, for pitchbooks and meeting preparation, is in development. RBC analysts cover about 1,700 companies, up from about 1,500 in 2023, with a goal of 2,500. | In production | Trading & capital markets |
| Genesis graph analytics and User Behaviour Analytics[1] | Genesis, an in-house graph analytics tool launched in 2025, proactively identifies attempted financial crimes; User Behaviour Analytics uses eight AI models and eight risk indicators to flag threats such as phishing. | In production | Fraud detection |
| AI-assisted software development[7][9] | AI has contributed to over 24 million lines of code and over 120,000 code reviews (Q2 2026 call), up from over 5 million lines and 55,000 reviews cited in December 2025. | In production | AI-generated code & coding agents |
| Responsible AI principles and Board oversight[23][1] | Four principles (privacy and security, accountability, fairness, transparency and responsible disclosure); the 2025 annual report says AI is a key area of Board oversight and that a risk governance framework guides GenAI development. | In production | AI governance (general) |
Use-case links open the tracker's rules-by-use-case matrix: which documents govern each system type.
What has RBC disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Total assets | C$2,498,817 million[2] | Jul 31, 2026 |
| AI enterprise-value target by 2027 (net of investments) | C$700 million to C$1 billion[9][3] | Dec 3, 2025 |
| Evident AI Index 2025 rank | 3rd of 50 global banks; #1 in Canada[10][11] | Oct 8, 2025 |
| AI models built | More than 200[7] | May 28, 2026 |
| LLM token usage since 2025 | Up more than 500%[7] | May 28, 2026 |
| RBC Assist users | Almost 27,000 employees using it; launched to over 30,000[15][9] | Feb 18, 2026 |
| Aiden users in Capital Markets | More than 8,000[15][18] | Feb 18, 2026 |
| Lines of code AI has contributed to | More than 24 million, with over 120,000 code reviews[7] | May 28, 2026 |
| ATOM training data | Over 65 billion client financial transactions[5] | Mar 27, 2025 |
| Borealis talent | More than 100 PhD scientists and over 850 AI developers and data engineers[5] | Mar 27, 2025 |
| Patent filings since 2019 | 1,200+, of which 635 related to AI[1] | Oct 31, 2025 |
| Annual technology spend | More than C$5 billion[16] | Mar 19, 2026 |
| Vector Institute collaboration since 2017 | 30 applied AI projects; 200 Vector-affiliated AI specialists recruited[21] | Jul 30, 2026 |
“Importantly, our target is net of investments, including building on investments already made in data storage, GPU clusters, proprietary LLMs, risk governance and in people”[9]
Dave McKay · President and CEO · Dec 3, 2025
“Since 2025, LLM token usage has increased by over 500%, reflecting the speed at which AI is being integrated into daily workflows and critical business processes.”[7]
Dave McKay · President and CEO · May 28, 2026
“Most of the deployment of our AI models now are not cost takeouts in the back office; they are revenue enablement models, adding more customers, making our frontline more efficient and more effective”[19]
Dave McKay · President and CEO · Jun 16, 2026
“ATOM represents the future of banking at RBC”[13]
Foteini Agrafioti · SVP, Data and AI and Chief Science Officer · Jul 29, 2025
“(RBC) has few rivals when it comes to the sheer number of documented AI applications”[11]
Alexandra Mousavizadeh · Co-founder and CEO, Evident Insights · Oct 8, 2025
“We're at the beginning of the beginning”[16]
Bobby Grubert · Head of AI and Digital Innovation, RBC Capital Markets · Mar 19, 2026
Who runs AI at RBC?
Bruce Ross
Group Head, AI (leads the AI Group; previously Group Head, Technology & Operations)[15][21]
Naim Kazmi
Group Head, Technology & Operations (from February 2026)[15]
Lindsay Patrick
Chief Strategy and Innovation Officer, RBC Capital Markets[16]
Neil McLaughlin
Group Head, Wealth Management and Insurance (AI in advisor productivity and asset management)[8]
Chris Phillips
Senior Vice President, Global Functions Technology[20]
Gopala Narayanan
Senior Vice President, Retail Risk and Chief Operating Officer, Group Risk Management[20]
Kelly Bradley
Chief Human Resources Officer (AI fluency and workforce training)[15]
Which regulators govern RBC's AI?
| Authority | Why it matters here | Documents |
|---|---|---|
| OSFI | RBC's home prudential supervisor. Guideline E-23 on model risk management takes effect for federally regulated institutions on 1 May 2027 and adds clarity for AI and machine-learning models; OSFI's July 2026 bulletin sets sound practices for generative and agentic AI under B-13, E-21 and B-10. | |
| OCC | City National Bank, RBC's US bank, is a national bank; group AI platforms used there fall under OCC model-risk and third-party expectations (covered at /banks/city-national). | OCC Bulletin 2026-13 OCC Semiannual Risk Perspective, Spring 2026 |
| Federal Reserve | RBC's US operations sit under the Federal Reserve's holding-company supervision, so group-provided AI services reaching US entities are third-party and model-risk matters. | SR 23-4 SR 26-2 |
| UK (BoE / PRA / FCA) | RBC describes itself as the fifth-largest UK wealth advisory firm (December 2025 strategic update); AI used in its UK businesses meets the UK model-risk and AI regime. | PRA SS1/23 FS2/23 |
What does the public record suggest about RBC's AI strategy?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice.
01
A research lab first, a product organisation second
Most banks buy their AI platform. RBC built Borealis in 2016, then Lumina and ATOM from its own transaction data, and only in 2026 added an enterprise AI Group to turn research into products. Evident's reading is that the lab is why RBC is a 'powerhouse for channelling applied research into practical application'.
02
The AI target is net, and not yet visible
The C$700 million to C$1 billion figure is stated net of investments and measured as an exit run-rate at the end of fiscal 2027. In February 2026 the CEO said the benefits had not yet appeared in results; by August the language had moved to 'accelerating'. Quarterly calls are where progress can be checked.
03
Growth framing rather than cost framing
McKay's June 2026 remarks tie AI to adding customers (17 million to 25 million) rather than back-office savings, while also expecting a smaller workforce through attrition. That puts the debate on whether AI-assisted underwriting and advice can be scaled without proportionate hiring.
04
Several model suppliers, one in-house core
RBC works with Cohere, NVIDIA, Databricks and, per American Banker, Anthropic, but keeps ATOM and Lumina in-house and says client data never leaves its control. How model-risk rules such as OSFI E-23 treat that mix is an open supervisory question.
What is RBC's AI value target?
RBC aims to generate C$700 million to C$1 billion of enterprise value from AI by 2027, measured as annualised exit run-rate benefit in revenue, cost avoidance, expense savings, fraud reduction and lower cost of risk, net of investments and incremental from fiscal 2024. It set the target at its 27 March 2025 Investor Day and said on 3 December 2025 and 27 August 2026 that it was on track.
How does RBC rank for AI?
RBC ranked third of 50 banks in the 2025 Evident AI Index (October 2025), behind JPMorgan Chase and Capital One, its fourth consecutive top-three finish. It is the only Canadian bank in the global top ten.
What are Lumina, ATOM and Aiden?
Lumina is RBC's enterprise data and AI platform built by RBC Borealis; ATOM is its proprietary foundation model trained on billions of client transactions; Aiden is the AI platform used in RBC Capital Markets, from electronic trading to research agents, and used by more than 8,000 employees.
Does RBC use AI in its US business?
RBC's US bank is City National, covered separately at /banks/city-national. RBC's group AI platforms are provided from Toronto; the bank has not published a separate AI programme for City National.
Is RBC cutting jobs because of AI?
CEO Dave McKay said on 16 June 2026 that most AI deployment is revenue enablement, not cost takeout, and that the bank may need fewer people over time mainly through attrition and demographic change. RBC describes itself as having more than 101,000 employees (July 2026).
- Royal Bank of Canada Annual Report 2025 (EX-99.1, filed with the SEC)
- RBC Third Quarter 2026 Report to Shareholders
- Strategic Update, December 3, 2025
- 2025 RBC Investor Day Presentation
- Royal Bank of Canada Investor Day on March 27, 2025 (earnings call transcript)
- Royal Bank Of Canada (RY) Q1 FY2026 earnings call transcript
- ROYAL BANK OF CANADA (RY.TO) Q2 FY2026 earnings call transcript
- ROYAL BANK OF CANADA (RY.TO) Q3 FY2026 earnings call transcript
- RBC Expects to Gain Over $700 Million of Enterprise Value From AI
- RBC ranks #1 in Canada and #3 globally for AI maturity in financial services for fourth straight year
- RBC retains top spot, TD stumbles as AI race in banking intensifies
- RBC is making a multibillion-dollar bet on AI. Will it pay off for Canada's largest company?
- RBC's proprietary AI foundation model for financial services aims to enhance personalized client experience
- RBC and Cohere partner to develop the next generation of highly secure generative AI solutions for financial services
- RBC creates new AI Group to help bring top AI opportunities to market
- RBC says its focus on AI is paying dividends
- #44 At RBC, Bobby Grubert is moving AI from a tool to a collaborator
- RBC Capital Markets Advances Research and Banking Capabilities With Agentic AI (undated page, retrieved 2026-10-05)
- RBC's McKay says he's using AI every day to help reshape how he runs the bank
- RBC wins IDC CIO Awards Canada 2026 for AI-driven credit underwriting transformation
- Royal Bank of Canada extends Vector Institute sponsorship, advancing responsible AI innovation and talent development through 2032
- RBC joins Massachusetts Institute of Technology's (MIT) CSAIL fintech research initiative on the role of AI in the future of finance
- Artificial Intelligence at RBC (undated page, retrieved 2026-10-05)
- Banks say AI is allowing them to recover unpaid loans, cut call centre volumes and save workdays
- Guideline E-23 – Model Risk Management (2027) - Letter
- Generative and Agentic Artificial Intelligence: Implications for Technology, Cyber Security, and Operational Resilience
When RBC 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 ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning