Compare
Compare institutions, criterion by criterion
Compare up to three Canadian banks, credit unions or fintechs side by side on the Canadian Bank AI Index. You see each institution’s score out of 100, its five pillar scores and which of the 19 criteria it meets, with the public source for each. Two equal totals can hide very different strengths.
Figures as of November 1, 2026.Pick up to three. Every row is a scoring criterion; tap through to each profile for the sources.
Mitey, data editor · Evidence checked
| Criterion | TD84#1 · 13 live customer uses | RBC81#2 · 8 live customer uses | CIBC80#3 · 5 live customer uses |
|---|---|---|---|
| Customer deployment | 17/20 | 12/20 | 14/20 |
| Live customer use: Assistants & chat (3) | 3 | 3 | |
| Live customer use: Money insights (3) | 3 | 3 | 3 |
| Live customer use: Fraud & identity (3) | 3 | 3 | 3 |
| Live customer use: Credit & lending (3) | 3 | 3 | 3 |
| Live customer use: Wealth & investing (3) | 3 | ||
| Live customer use: Business banking (3) | 3 | ||
| Customers can use an AI assistant directly (2) | 2 | 2 | |
| Governance disclosure | 20/20 | 20/20 | 20/20 |
| Publishes responsible-AI principles (5) | 5 | 5 | 5 |
| Discloses board or committee oversight of AI (5) | 5 | 5 | 5 |
| Names AI as a risk in its annual reporting (5) | 5 | 5 | 5 |
| Names an executive accountable for AI (5) | 5 | 5 | 5 |
| Model-risk readiness | 12/20 | 20/20 | 20/20 |
| Says its own model-risk framework covers AI/ML (12) | 12 | 12 | 12 |
| Names OSFI E-23, or the AMF equivalent for Quebec-regulated institutions (8) | 8 | 8 | |
| Transparency to customers | 15/20 | 15/20 | 20/20 |
| Tells customers when AI or automated decisions are used (10) | 10 | 10 | 10 |
| Signed Canada’s Voluntary Code on generative AI (5) | 5 | ||
| At least 3 first-party AI disclosures in the last 24 months (5) | 5 | 5 | 5 |
| AI talent & capacity | 20/20 | 14/20 | 6/20 |
| Has a named AI lab or dedicated AI team (8) | 8 | 8 | |
| Discloses AI or data-science headcount (6) | 6 | ||
| Shows active AI hiring (6) | 6 | 6 | 6 |
Questions people ask
What does comparing scores tell me?
Where two institutions differ in what they have launched and disclosed about AI. It does not tell you which is better to bank with, or whose AI is more accurate.
Why compare criteria and not just totals?
Two institutions with the same total can be very different: one may ship many customer features while publishing nothing about oversight. The criteria rows show exactly where the points come from.
Two totals a few points apart can hide large differences in where the points come from.
| Customer deployment | 17 vs 12TD +5 |
|---|---|
| Governance disclosure | 20 vs 20Level |
| Model-risk readiness | 12 vs 20RBC +8 |
| Transparency to customers | 15 vs 15Level |
| AI talent & capacity | 20 vs 14TD +6 |
| Totals | 84 vs 81 |
| Gap in the total | 3 pointsLargest pillar gap: 8 points |

