Competitive Compass Competitive Compass Signal for Financial Leaders
Mystery Agent  ›  Credit Cards  ›  Chase
Credit Cards · August 2026

Chase publishes card pricing in text; applications open a script shell

An AI agent ran five ordinary credit cards tasks on the public pages of Chase and scored what each page delivered.

Run 6 August 2026 · five tasks · public pages, logged out · Issuer
75/100
Mystery Agent score
Credit Cards · rank 4 of 25
Purchase APR
92Served as text
Card comparison
88Served as text
Pricing and terms
92Served as text
Welcome offer
90Served as text
Application page
15Arrives after scripts
Summary
Chase also appears in the Deposits study, where it scores 47 and ranks 19 of 25. See that scorecard.

Task by task

Each task is scored out of 100 on completion, answer quality and how directly the answer was reached. Each note describes what the page delivered, with a link to the page the answer was found on.

Task 01Purchase APRServed as text92 out of 100

The Sapphire Preferred card page states its purchase APR in visible text as 19.24% to 27.49%, next to the $95 yearly fee line, confirmed word for word on the linked pricing and terms disclosure.

Task 02Card comparisonServed as text88 out of 100

Comparing two cards, an automated reader read the Sapphire Preferred's $95 fee and travel credit against the Freedom Unlimited's $0 fee and unlimited 1.5% cash back, both stated in plain visible text.

Task 03Pricing and termsServed as text92 out of 100

A link on the Sapphire Preferred page led to a complete Schumer-box disclosure delivered as ordinary HTML: purchase APR 19.24% to 27.49%, cash advance APR 28.49%, a $95 yearly fee, and every listed penalty and transaction fee.

Task 04Welcome offerServed as text90 out of 100

The Sapphire Preferred page states its welcome offer in visible text: 75,000 points after spending $5,000 in purchases within the first 3 months from account opening, sitting in the same block as the APR.

Task 05Application pageArrives after scripts15 out of 100

A plain request to the guest application page returned 191,021 bytes of mostly scripts and assets; stripped of markup, only 1 character of visible text remained, with zero input fields present in the response.

Earn 75,000 points after you spend $5,000 in purchases in the first 3 months from account opening.

The application page

What the account opening page states to a reader that runs no scripts.

191,021Bytes delivered
1Characters of visible text
0Form fields in the first response
After scripts runWhat an applicant needs

Strengths and opportunities

What is working

  • Chase publishes its full Schumer-box pricing and terms disclosure as plain readable HTML rather than a locked PDF, so an agent can read every rate and fee (purchase APR, balance transfer APR, cash advance APR, penalty APR, all fees) in one fetch with no extraction work.
  • The core shopping facts an agent needs, purchase APR, annual fee, rewards rates, and the welcome offer bonus, spend threshold, and time window, all sit in the visible delivered text of the card detail page itself, not behind a script call or a hidden data layer.
  • Navigation from the homepage to a specific card's full pricing detail took only three to four plain GET hops with no 403s, no bot challenges, and fast response times (roughly one to two and a half seconds per page).

Opportunities

  • The application entry point currently serves a near-empty JavaScript shell: zero form fields and one character of visible text in the raw response. Publishing a lightweight static landing step ahead of the single-page app, even just a summary of what information applicants will need, would let an agent (and many browsers) read the eligibility and requirements picture before the interactive form loads.
  • an llms.txt file is available to publish; the URL returns Chase's standard site-wide 404 page. A short agent-facing file pointing to the card comparison hub and the pricing-and-terms host would give agents a documented, intentional entry point instead of requiring homepage link discovery.

Access

Every page the agent requested was delivered on the first attempt. Publishing an llms.txt file would give agent traffic a documented entry point.

Back to the credit cards leaderboard  ·  Read the report  ·  How the study works