Competitive Compass Competitive Compass Signal for Financial Leaders
Competitive Compass
AI Compass
By Competitive Compass
Deep Dive · CITI June 2026
Top 10 Banks

Citi.

Agent Readiness Score
36 / 100
Technical Readiness
16 / 50
Content Architecture
20 / 50

Citi posts an Agent Readiness Score of 36 out of 100 in the June 2026 deep dive refresh, up from 30 in the May Index. Crawl access is genuinely open, the deposit FAQs read beautifully, and the Personal Loans page proves the playbook works by shipping a real 9.99% APR in the raw HTML. One pattern holds the score down: the structure ships server-side while the numbers wait for JavaScript. Fixing that single habit moves Citi out of the early tier.

Surface Scorecard

How each product page scores.

Citi is scored across all 7 standard product surfaces. Lending blends the Personal Loans page with the Mortgage hub, and Investing maps to Citi Self Invest and the Wealth overview, consistent with how full-service banks are normalized across the index.

Surface
Tech /50
Content /50
Total /100
Homepage
14
8
22
Credit Cards (Strata, Double Cash & AAdvantage)
22
30
52
Checking (Access & Regular Checking)
20
34
54
Savings (Citi Accelerate & CDs)
18
28
46
Investing (Citi Self Invest & Wealth)
12
10
22
Lending (Personal Loans & Mortgage)
19
21
40
Help/Support
8
6
14
Strengths and Gaps

What lifts and drags the score.

Top Strength

Citi opens the front door wide. robots.txt welcomes every crawler with a sitewide Allow and lists 11 product sitemaps covering cards, banking, mortgage, small business, travel, and personal loans, so ClaudeBot, GPTBot, and PerplexityBot all read the site freely today. The Checking and Savings pages carry real FAQ blocks with verbatim-quotable answers: an agent asking about monthly fees gets the $5 Access and $15 Regular numbers, the $250 direct deposit waiver, and the no-overdraft-fee policy straight from the HTML. The Personal Loans page is the model for the whole site, publishing its 9.99% APR floor, the 2-14-2026 as-of date, worked payment math, and a clean FAQ in fully server-rendered HTML.

Top Gap

The numbers wait for JavaScript. Card pages render every pricing field blank in the raw HTML: the Double Cash page reads “Earn $ cash back bonus,” “Intro APR %,” and a Pricing Details paragraph with every dollar figure, rate, and month count missing. Savings pages publish APY footnotes with as-of dates from March and April while the yield itself arrives only after render. The Mortgage hub serves “You need to enable JavaScript to run this app,” and Help, Contact Us, and Citi Self Invest all resolve to empty app shells titled “Citibank Online.” The homepage greets agents with a sign-on module and a maintenance banner. llms.txt returns empty, and FAQPage and FinancialProduct JSON-LD are absent from the surfaces scored.

The encouraging part: Citi already owns both halves of the fix. The content team writes plain-language answers that agents can quote today, and the Personal Loans team ships real numbers in real HTML. Extending the Personal Loans rendering pattern to cards, savings, and mortgage is an engineering decision, and each surface that adopts it converts existing content into agent-readable answers overnight.

An agent reading the Double Cash page sees the architecture of a great answer with every number missing. Earn $ cash back. % intro APR for months. The frame is built. The data waits for JavaScript.
Peer Context

Where Citi sits in the index.

Position within the AI Compass Index of 36 brands. Closest peer brands shown in order. The June refresh lifts Citi from 30 to 36, closing most of the gap to the cluster at 38.

25
JPMorgan Chase
47
25
PNC Bank
47
25
Synovus
47
28
Wells Fargo
41
28
Cadence Bank
41
28
Affirm
41
31
Robinhood
39
32
KeyBank
38
32
Huntington
38
32
Fifth Third
38
35
Citi
36
36
NC SECU
18
The Two Dimensions

Where the work lives.

Every Agent Readiness Score splits into Technical Readiness and Content Architecture. The two halves point to different teams and different timelines.

Technical Readiness
16 / 50
Crawl access is a solved problem: robots.txt allows all agents and 11 sitemaps map the estate. The lift sits in rendering. Publish card pricing, APYs, and mortgage content as real numbers in server-rendered HTML, stand up a static Help hub and llms.txt, and add FAQPage and FinancialProduct JSON-LD to pages that already hold the answers. The Personal Loans page is the internal proof that the stack supports all of it today.
Content Architecture
20 / 50
The writing is ahead of the rendering. Deposit FAQs answer real consumer questions in quotable plain language, card pages explain benefits with genuine depth, and the Personal Loans page teaches its product with worked examples. Extending that FAQ discipline to Mortgage, Investing, and a public Help center, and keeping every visible as-of date current, gives agents a complete answer library to draw from.