The robots file welcomes every AI agent and the edge serves them cleanly. The card pages are ready to serve their facts through that same open door.
American Express Canada posts an Agent Readiness Score of 25 out of 100 in the AI Compass Canada July 2026 Index. The score splits into Technical Readiness at 16/50 and Content Architecture at 9/50. Crawler permission and edge behaviour are open and correct. Card product pages arrive consent-gated to navigation, and annual fee, interest rate, welcome bonus and earn rate reach an agent on the benefits pages today.
The access layer is right. The robots file carries four narrow rules under a single wildcard group, names no AI crawler, and declares a sitemap that resolves to a Canadian file with 1,476 URLs. Live tests confirmed GPTBot, ClaudeBot, PerplexityBot, CCBot, Google-Extended and Bytespider each receive 200 on the Canadian root and on card pages, with no bot challenge.
The card pages are where the opportunity sits. The Cobalt page returns 597,595 bytes that reduce to 1,987 characters of readable text, consisting of navigation and footer plus the literal string “Consent is required to load the content”. All-cards, Gold Rewards, Aeroplan Reserve and the compare page each return between 1,742 and 2,046 characters with the same result, so serving product facts unconditionally puts fee and rate into the first response.
Two Canadian paths are ready to open. Both the travel services root and the travel cards category answer “Access Denied” to ClaudeBot, GPTBot and a standard browser agent alike, so clearing that rule returns a whole category to agent answers.
The benefits estate is where Amex Canada does publish. The Plan It terms page serves “a Preferred Rate for Purchases of 21.99%” and an effective APR table reading “0.51% 11.30% 0.85% 18.83%”. The Platinum membership rewards page serves the full earn structure, “Earn 2 points for every $1 spent on eligible dining in Canada” through “1,000 POINTS = $10”. Those pages carry the answer, and the card pages are ready to carry it too.
One flagship offer was selected for each sub-category Amex Canada features on the en-ca root, with the instalment and rewards surfaces added where pricing was found. Each surface was read from the server response, without a browser session. Findings were independently verified against a second retrieval pass.
Amex Canada places no restriction on AI crawlers. The full ruleset under “User-agent: *” is four Disallow lines, none of which names an AI agent, and a sitemap is declared that resolves to 1,476 Canadian URLs.
The edge honours that policy and serves AI agents cleanly. Direct tests returned 200 for GPTBot, ClaudeBot, PerplexityBot, CCBot, Google-Extended and Bytespider on the Canadian root and on the Cobalt card page, with no bot challenge.
The instalment terms page publishes a real APR table. It serves “repayment of $1422.00 at a Preferred Rate for Purchases of 21.99%” and “Monthly Installment Fee Effective APR 0.51% 11.30% 0.85% 18.83% 0.90% 19.94%” as readable text.
The Platinum rewards page publishes the full earn structure. It serves “Earn 2 points for every $1 spent on eligible dining in Canada”, “Earn 1 point for every $1 in all other Card purchases”, “1,000 POINTS = $10” and “1,000 points = $7 at Amazon”.
The card pages are ready to serve their facts. The Cobalt page returns 597,595 bytes reducing to 1,987 characters of navigation, plus the literal string “Consent is required to load the content” where the card detail belongs, so serving product facts unconditionally puts fee and rate in the first response.
Price is ready to publish on every card page. The only dollar sign in 597 kilobytes of Cobalt HTML is the CSS selector “[class$=cardTitle]”, and All-cards, Gold Rewards, Aeroplan Reserve and compare each return under 2,050 characters today.
Two commercial paths are ready to open. The travel services root and the travel cards category answer “Access Denied” to ClaudeBot, GPTBot and a standard browser agent alike, so clearing that rule returns a whole category to agent answers.
The llms.txt is ready for Canadian coverage. It returns 200 as text/plain, headed “# American Express - LLM Discovery File”, and indexes one entry today, so Canadian card, rewards, travel and insurance pages are the next entries.
The encouraging read: the access layer is correct, which is the part that usually takes longest to change. The benefits estate already shows the publishing standard the card pages can adopt.
An agent asked what an Amex Canada card costs is served the card name, and the page is ready to serve the fee and the rate alongside it.
American Express Canada sits at the base of the Canadian field on a score that reflects rendering rather than product. The US property measured in the published index scores 67 on the same rubric. Scores shown are the nearest neighbours in the July 2026 Canadian field.
Every Agent Readiness Score splits into Technical Readiness and Content Architecture. The two halves point to different teams and different timelines.
Sequencing, effort estimates and the surface-by-surface build order sit in the full engagement plan. Item one carries the largest share of the available lift.
Move annual fee, purchase and cash advance rates, welcome bonus, earn rates and foreign transaction fee into the first HTML response on all card pages, so they read as text before any script runs.
The string “Consent is required to load the content” currently sits where the card details belong. Serve product facts unconditionally and keep consent scoped to tracking and personalisation.
The travel services root and the travel cards category are ready to open to any server-side reader, which returns a whole category to every agent answer.
Replace the single entry with Canadian card, rewards, travel and insurance pages, each with a one-line description naming the product and its fee.
The direction of travel favours American Express. The access layer is correct, which is the part that usually takes longest to change.
Amex Canada already runs an open door and a well written benefits estate. Serving the card facts through that door is the whole of the work.
AI Compass measures what an AI agent receives when it visits a financial institution website on a customer’s behalf. It reads the served response the way an agent does, without a browser session, without scripts, and without a human clicking to reveal a value.
Each product surface earns a score out of 100, built from Technical Readiness out of 50 and Content Architecture out of 50. The institution score is the average across the surfaces reviewed, up to seven.
Sub-categories come from the personal banking navigation on each brand’s Canadian root. Within each one, the flagship is the product the institution itself features first. Brands with fewer product lines are scored on the surfaces they actually sell, so a focused brand is measured on quality rather than on breadth.
Pages were read on 30 July 2026. Rates, bonus offers and terms move, and scores describe what each site served on that date. Where a site offers more than one language, the version served by default to a request with no language preference was used.
Scores come from observed page behaviour alone. No institution reviews or approves a score before publication, and every brand is measured against the same rubric. Outlier results are verified against a second independent retrieval pass before publication.
The score covers the public marketing site. Logged-in banking, the mobile app and the branch experience sit outside it. A site can score well here and still have work to do inside the login, and the reverse holds too.
American Express Canada carries an AI Compass Agent Readiness Score of 25 out of 100 in the July 2026 Canada Index. That is the sum of Technical Readiness (16/50) and Content Architecture (9/50), scored across 7 product surfaces and read from the server response on 30 July 2026.
Each product surface earns a score out of 100, built from Technical Readiness out of 50 and Content Architecture out of 50. Technical Readiness covers crawler permission, machine-readable discovery, server-side completeness, retrievability and disclosure availability. Content Architecture covers price in text, eligibility in text, comparison structure, question coverage and plain answers. The institution score is the average across the surfaces reviewed. Amex Canada sits at 16 on Technical Readiness and 9 on Content Architecture.
The strongest surface is Instalment plans at 42, with Membership Rewards at 39 behind it. The next points sit on server rendering the card facts and serving product content ahead of the consent gate, which moves every card surface at once.
Canada scores are refreshed each edition of the AI Compass Index, and a deep dive is refreshed whenever a brand ships a major site or content change. The July 2026 edition reads the served response on 30 July 2026. Subscribe at competitive-compass.com to get the next refresh.