TL;DR
- Six of ten Korean banks crossed 40% mention rate in W20, Banking's debut week. No other industry in the index has that many tier-1 brands.
- Shinhan Bank leads Banking on mentions (59%) but is cited at just 2% — a 57-point gap, ranked sixth across all 103 brands tracked.
- The actual citation leader inside Banking is KB Kookmin Bank at 13% — ranked only fourth by mentions (49%). Mention rank and citation rank do not point at the same bank.
- Banking's own-domain citation share (26.8%) and tutorial-content citation share (14.6%) are both the highest of any of the 11 industries tracked — well above the all-industry averages of 19.5% and 7.5%.
Banking joins the index as the 11th industry tracked, and the debut-week numbers carry one clear signal on their own.
Six of ten Korean banks we now track — Shinhan Bank (59%), Woori Bank (52%), Hana Bank (52%), KB Kookmin Bank (49%), KakaoBank (42%), and Toss Bank (40%) — crossed 40% mention rate in W20. That is more tier-1 brands than any other industry in the index (Securities is next with five, Non-Life Insurance with four). Four commercial banks and two internet-only banks moved into tier-1 in a single week of measurement.
The real debut-week finding sits elsewhere. Shinhan Bank, the mention leader at 59%, is cited at just 2% — a 57-point gap, sitting in the W20 top six across all 103 brands tracked. The actual citation leader inside Banking is KB Kookmin Bank, ranked fourth by mentions at 49% but first by own-domain citation rate at 13%. In Banking, "the bank AI names most often" and "the bank AI cites most often" are not the same bank.
How AI is sourcing its answers
Every week, we run a fixed set of consumer-style prompts through the three AI surfaces Korean consumers actually use — Google AI Overviews (AIO), Naver AI Briefing (AIB), and ChatGPT. For each response we record which brand was mentioned, which URL the AI surfaced as a source, what type of domain that URL belongs to (the bank's own official site, an external blog, news, video, etc.), and — independently — what type of content the page itself is (curation, tutorial, news, PDP, homepage, etc.).
The 5,056 Banking citations in W20 break down by domain type as follows.
| Domain type | Citations | Share |
|---|
| External blog | 2,475 | 49.0% |
| Bank's own official site | 1,353 | 26.8% |
| News | 526 | 10.4% |
| Video (YouTube) | 253 | 5.0% |
| Brand-owned blog | 99 | 2.0% |
| Forum / Q&A | 56 | 1.1% |
| Social · wiki · eCommerce · other | 210 | 4.2% |
External blog at #1 (49.0%) matches every other industry we track — it is nearly identical to the all-industry average of 46.4%. What is different in Banking is #2. The bank's own official site takes 26.8% of citations, 7.3 points above the all-industry average of 19.5% and the highest official-site share of any of the 11 industries tracked. Credit Card Issuers are close behind at 26.4%, but the other nine industries are all under 25%.
Content type tells the same story from another angle. Banking has the lowest curation-content share of any industry at 21.2% — 13.5 points below the all-industry average of 34.7%. And Banking has the highest tutorial-content share at 14.6% — nearly double the all-industry average of 7.5%. When AI is assembling an answer to a banking question, it is leaning much less on "Top 5 banks for 2026" blog round-ups and much more on loan calculators, savings comparisons, and walkthroughs of how a specific product works.
What wins in each industry
To see where the debut-week finding lands, you have to look at the inside of the Banking category — which bank leads each measurement. Below are all ten banks tracked in W20 — four commercial banks, three internet-only banks, and three others — sorted by mention rate, with the W19 → W20 mention-rate change in the rightmost column.
| Bank (type) | Mention % | Own-domain % | Δ Mention |
|---|
| Shinhan Bank (commercial) | 59% | 2% | +2pp |
| Woori Bank (commercial) | 52% | 9% | +6pp |
| Hana Bank (commercial) | 52% | 5% | -3pp |
| KB Kookmin Bank (commercial) | 49% | 13% | -1pp |
| KakaoBank (internet-only) | 42% | 7% | -2pp |
| Toss Bank (internet-only) | 40% | 8% | -2pp |
| NH NongHyup Bank (commercial) | 32% | 0% | +1pp |
| K Bank (internet-only) | 26% | 0% | +1pp |
| IBK Industrial Bank (specialty) | 14% | 1% | +2pp |
| SC First Bank (foreign-owned) | 2% | 1% | 0pp |
The mention leader and the citation leader are different banks. Shinhan Bank is the Banking category mention leader at 59%, but its own-domain citation rate is only 2% — a 57-point gap that lands it in W20's top six gaps across the entire 103-brand index. The Banking citation leader is KB Kookmin Bank at 13%, ranked only fourth on mentions at 49%. Woori Bank (52% / 9%) and Hana Bank (52% / 5%) are tied on mentions but separated by four points on citation — the same pattern in miniature.
Internet-only banks lag on mentions but catch up on citation. KakaoBank (42% / 7%) and Toss Bank (40% / 8%) sit more than ten points below the four commercial banks on mention rate, but their own-domain citation rates are competitive with the commercial-bank average — and in Shinhan Bank's case, materially higher. Two of the top five W20 citation-rate risers across the entire 103-brand index are banks (KB Kookmin +4pp, Toss Bank +4pp). Mentions and citations move on different axes inside Banking, and the difference appears to track how much usable explainer content lives on the bank's own domain.
The biggest single-week mover in the entire W20 index also came from Banking. Woori Bank lifted its mention rate from 46% to 52%, a +6-point gain — the largest single-brand mention-rate increase across all 103 brands tracked. Producing both the index-wide top mover and the index-wide #6 gap in the same debut week is itself a signal of how densely competitive Banking is on day one.
What this means for AI visibility budgets
Three implications follow directly from Banking's W20 debut data.
In Banking, treating "mention leader" and "citation leader" as the same problem is the wrong starting move. Shinhan Bank is #1 by category mention rate and roughly #8 by own-domain citation rate. KB Kookmin Bank is #4 by mention and #1 by citation. "Increase how often AI names us" and "Increase how often AI cites our website" are not two phases of the same workstream — at least not on Banking's debut-week data. They should be planned and budgeted as separate measurement targets until evidence accumulates that they share inputs.
The bank's own site is unusually strong as an AI citation source in Banking. The category's own-domain citation share is 26.8%, the highest of any of the 11 industries we track — 7.3 points above the all-industry average of 19.5%. Unlike telecom (W19, 11.9%) or home appliances (16.1%), where AI almost never lands on the brand's own site, in Banking the bank.com itself is the second-largest citation source after external blogs. The question is not whether SEO matters in Banking — on this data, it clearly does — but which content format on the bank's own domain converts a mention into a citation.
The format that works in Banking is explainer, not curation. Banking has the lowest curation-content citation share of any industry at 21.2% and the highest tutorial-content citation share at 14.6% — nearly double the all-industry average of 7.5%. KB Kookmin Bank's position as the Banking citation leader at 13% is not unrelated. Loan calculators, mortgage explainers, savings-rate comparisons, and "how to apply" walkthroughs are more likely to land in AI's citation path than product-detail pages — and that effect is bigger in Banking than in any other industry we measure.
What we'll publish each week
AI Visibility Weekly publishes every Monday. We track 103 Korean enterprise brands across 11 industries (660 prompts per week, 47,715 classified citations in W20) and pull out the single most meaningful shift in how AI is talking about them. Week 01 gave the all-industry source distribution, Week 02 surfaced the curation concentration in credit cards, Week 03 framed the mention-citation gap in Korean telecom, and this Week 04 marks Banking's debut as the 11th tracked industry.
The newsletter is the digest, the blog is the analysis. Both are free, and they run alongside the public Enterprise AI Visibility Index where readers can find their own brand at any time. The next category expansion will get the same debut-data treatment when it lands.
The question to take into the week from Issue 04 is direct. If AI names your bank more often than any other bank in its category but almost never cites your website, how should the AI visibility budget split between external-blog curation, your own explainer content, and own-site SEO?
FAQ
Why does Banking start being tracked in W20?
Banking is the 11th industry to enter the index, debuting in W20. We track ten Korean banks: the big four commercial banks (Shinhan, Woori, Hana, KB Kookmin), three internet-only banks (KakaoBank, Toss Bank, K Bank), plus NH NongHyup, IBK Industrial, and SC First. Sixty prompts are run weekly across all three AI surfaces. In its debut week, the category produced 5,056 classified citations.
Why is the mention leader and the citation leader a different bank in Banking?
Shinhan Bank leads Banking mentions at 59% but is cited at just 2%. KB Kookmin Bank leads Banking citations at 13% but ranks fourth by mentions (49%). How often AI names a bank and how often AI cites that bank's own website move on independent axes. The format of the content placed on the bank's own domain is the most likely cause of the divergence.
How does Banking's citation pattern differ from other industries?
Banking has the highest own-domain citation share of any of the 11 industries we track, at 26.8% — 7.3 points above the all-industry average of 19.5%. Banking also has the highest tutorial-content citation share at 14.6%, nearly double the all-industry average of 7.5%. In the opposite direction, Banking has the lowest curation-content citation share at 21.2%, against an all-industry average of 34.7%.
Why does Banking have more tier-1 brands than any other industry?
Six of ten Korean banks crossed 40% mention rate in W20 — more tier-1 brands than any other industry. Securities follows with five, Non-Life Insurance with four. Banks are the entry point for everyday financial decisions for Korean consumers, so AI tends to name all four major commercial banks plus the two largest internet-only banks when answering a banking-adjacent question. The debut week compresses an unusually high level of competition into a single category.
Which AI surfaces does this data cover?
Three: Google AI Overviews (AIO), Naver AI Briefing (AIB), and ChatGPT. We run a fixed set of consumer-style prompts through all three every week and classify the URLs each surface returns by both domain type and content type.
How often is this report updated?
Every Monday. The data covers 103 Korean enterprise brands across 11 industries (660 prompts and 47,715 classified citations in W20). The latest snapshot covers W17 to W20.
Find your brand → bubbleshare.io/resource/enterprise-index
Want to track your own prompts, internally? AI Visibility Weekly runs on the same engine we offer to enterprises as VIVI: private, configurable, scoped to your competitors and your category. Talk to us.
Subscribe to AI Visibility Weekly: a 3-minute Monday digest of how AI search is moving across Korean enterprise brands.