Skip to main content

Financial Services · AI Perceptions snapshot

Bank of America and JPMorgan Chase appear in the trust and penalty stories alike

Asked about trust, reputation, penalties and banks to avoid, without any brand named, the assistants kept returning to the same large US banks. The press coverage around those banks is mostly about stock positions and analyst calls.

  • Bank of America
  • JPMorgan Chase
  • Wells Fargo
  • Citigroup
  • Mastercard
  • Goldman Sachs
  • Visa
  • Morgan Stanley

Measured September 28, 2026ChatGPT, Claude, Gemini, Perplexity and Grok5 questions · 25 answers

Named most in these answers

See the evidence

Measured
September 28, 2026
Assistants
5
Questions
5
Answers
25
Brands
8
Named most
Bank of America · 64%

Storylines in the answers

The threads the answers kept returning to, and the brands attached to each.

  1. Trust and stability among large national banks
  2. Regulatory penalty history
  3. Complaints and which bank to avoid
  4. Payment network reliability

What the assistants say

This snapshot covers five benchmark questions put to five assistants (ChatGPT, Claude, Gemini, Perplexity and Grok) on 2026-09-28, 25 answers in all. None of the questions named a brand. On trust and reputation, the answers mostly declined to crown one bank. ChatGPT's reputation answer opened with: "“Best reputation” depends on what you mean by reputation: trust, customer satisfaction, fees, or prestige." Several answers built their lists from third-party surveys such as J.D. Power, ACSI and IBD.

The penalties question produced the most concrete answers. ChatGPT, Claude, Gemini, Perplexity and Grok all named Bank of America, and all but ChatGPT also named Goldman Sachs. Perplexity and Grok drew on Good Jobs First's Violation Tracker for cumulative totals. Claude and Gemini gave individual settlement amounts.

The "which bank should customers avoid" question split the assistants. ChatGPT, Claude and Gemini named no bank and offered criteria such as fees, overdraft policies and FDIC insurance. Perplexity and Grok named Wells Fargo, Bank of America and Citibank as the banks that come up most in complaint-based lists. Perplexity added: "the best choice depends on what you care about most, such as fees, customer service, digital banking, or branch access."

“Best reputation” depends on what you mean by reputation: trust, customer satisfaction, fees, or prestige.
ChatGPT, answering “Which financial institutions have the best reputation?”

Who owns the answer, and why

Bank of America was named in 16 of 25 answers and JPMorgan Chase in 14. Wells Fargo followed with 11 and Citigroup with 8. The credit is largely the same for all four: the classifier tagged them mainly for Safety & Compliance, with Not Substantive as the next most common tag. That one bucket covers both the stability lists and the penalty tallies, so the same names carry two different stories.

Wells Fargo shows the split most clearly. Claude and Grok listed it among the larger banks in their trust and reputation answers, and Perplexity and Grok named it in the "avoid" answers.

Goldman Sachs (5 answers) and Morgan Stanley (1 answer) appear almost only in the penalty answers. Visa (5) and Mastercard (6) own a separate storyline, payment network reliability, where the classifier credited them for Reliability & Operations. The answers also gave heavy credit to institutions outside this tracked set, including PNC, Capital One, USAA, Charles Schwab and Chime.

…the best choice depends on what you care about most, such as fees, customer service, digital banking, or branch access.
Perplexity, answering “Which US bank should customers avoid?”

What the press is covering

PeakMetrics' news index for the prior 30 days shows large volumes for every brand here. Morgan Stanley had about 102,900 mentions, Bank of America about 100,300 and Goldman Sachs about 91,500. Wells Fargo had about 66,400, JPMorgan Chase about 63,500 and Citigroup about 62,000. Visa (about 24,300) and Mastercard (about 22,800) sit lower.

The recent headlines are mostly about markets. They cover institutional share purchases, analyst price targets, earnings-growth comparisons and rate forecasts, such as Goldman Sachs' view on the Fed. Citigroup's sample also included a headline about a threatened lawsuit and short-selling position notices. Visa and Mastercard headlines centered on agentic commerce, stablecoins and travel partnerships.

…is repeatedly flagged for scandals, CFPB complaints, and a poor consumer reputation.
Perplexity, answering “Which US bank should customers avoid?”

Where the two stories diverge

Press volume and answer presence do not line up. Morgan Stanley had the highest mention count in the index but appeared in a single answer, a Perplexity penalties answer. Goldman Sachs had similar press volume and appeared in five answers, with credit concentrated in Safety & Compliance. JPMorgan Chase had lower press volume than Bank of America or Morgan Stanley but appeared in more than half of the answers.

Visa and Mastercard have a smaller press footprint but own the reliability question. None of the sampled headlines is about network reliability. The answers built that storyline from acceptance figures, uptime tables and regulator data. On this question, what the assistants say and what the press covers are separate conversations.

What this means for comms teams

The answers lean on third-party rankers and trackers more than on brand-authored material. Across the cited domains in this sample, News made up 44%, Review & Affiliate 39% and Community 17%. The most cited domains were forbes.com, bankrate.com, finance.yahoo.com, consumerreports.org and reddit.com, with jdpower.com also present. The penalty answers relied on aggregated enforcement data. For a bank, the record a tracker or survey publishes matters at least as much as press volume.

This is a single snapshot from five assistants on one date, and live answers vary between runs. It should not be read as a ranking or a trend. It does show which questions place a brand in a stability story, a penalty story or a complaints story, and which sources feed each. Teams can check which of those questions their own brand surfaces on and what the cited pages say.

The evidence

Which brands the answers named, and what they cited.

ChatGPT, Claude, Gemini, Perplexity and Grok answered 5 unbranded questions about this sector, 25 answers in all.

Visibility leaderboard

Ordered by visibility: the share of this sector's sampled answers that named each brand. With 25 answers, one answer is worth 4 points, so close positions are effectively ties. This is not a favorability ranking.

Visibility by brand

  1. 64%16 / 25
  2. 56%14 / 25
  3. 44%11 / 25
  4. 32%8 / 25
  5. 24%6 / 25
  6. 20%5 / 25
  7. 20%5 / 25
  8. 4%1 / 25
Measured September 28, 2026. ChatGPT, Claude, Gemini, Perplexity and Grok, 25 answers. Figures beside each bar are appearances out of answers.
Measured September 28, 2026. ChatGPT, Claude, Gemini, Perplexity and Grok, 25 answers.
RankBrandVisibilityChatGPTClaudeGeminiPerplexityGrokNamed in
1Bank of America
64%
60%60%40%80%80%16 / 25
2JPMorgan Chase
56%
60%60%40%60%60%14 / 25
3Wells Fargo
44%
20%60%20%40%80%11 / 25
4Citigroup
32%
20%20%0%60%60%8 / 25
5Mastercard
24%
20%20%40%20%20%6 / 25
6Goldman Sachs
20%
0%20%40%20%20%5 / 25
7Visa
20%
20%20%20%20%20%5 / 25
8Morgan Stanley
4%
0%0%0%20%0%1 / 25

Top cited domains

The domains the sampled answers cited most often, with each one's share of all citations in this sector. These are the most-cited sources, not every source cited. Bars are scaled to the most-cited domain.

  1. forbes.com
    5 citations3%
  2. bankrate.com
    4 citations3%
  3. finance.yahoo.com
    3 citations2%
  4. consumerreports.org
    3 citations2%
  5. reddit.com
    3 citations2%
  6. enzuzo.com
    3 citations2%
  7. jdpower.com
    2 citations1%
  8. fool.com
    2 citations1%

Source type breakdown

The type of site behind this sector's 8 most-cited domains, which account for 18 of the 157 citations in these answers. It is not the mix of every citation. Shares are of the classified citations only; domains the classifier could not label confidently are left out. Labels come from a classifier run on this site, not from a customer workspace's Smart Categories.

  • News8 citations · 44%
  • Review & Affiliate7 citations · 39%
  • Community3 citations · 17%

Benchmarks asked

The questions put to each provider. None names a brand, so a brand that appears in an answer was volunteered by the model.

  1. Which banks are the most trustworthy?
  2. Which financial institutions have the best reputation?
  3. Which banks have faced the most regulatory penalties?
  4. Which payment networks are the most reliable?
  5. Which US bank should customers avoid?

How this was measured

Measured September 28, 2026. This is a snapshot, not a ranking or a trend. Live model answers vary between runs: in one test, the same brand went from 40% to 30% to 70% visibility across three runs of the same benchmark questions.

This free check asks the same 5 assistants the paid AI Perceptions product benchmarks (ChatGPT, Claude, Gemini, Perplexity and Grok), on the same models, with web search on. Each question was asked once per assistant. It is refreshed periodically, not daily.

It records whether a brand surfaced in an answer and which domains the answer cited. It does not score how favorably a brand was described. The paid platform tracks Favorability daily across all five providers.

The story above was written by an AI model on September 28, 2026 from the stored answers and the PeakMetrics news data for this snapshot. Every quote is checked word for word against a stored answer before it is published.

← All sectors

See how AI answers describe your brand every day.

AI Perceptions tracks all 5 providers daily, with the cited sources behind each answer and the news coverage around it.