Bank of America AI Search Visibility: Recommendation Coverage Declined Across All Six AI Platforms

Research status: Exploratory longitudinal research. Bank of America's AI recommendation momentum has not been validated as a predictor of deposit growth, loan growth, account openings, card activity, wealth flows, revenue, earnings, analyst revisions, valuation, or stock returns.

Observation window: July through September 2026

Ticker: BAC

Public parent: Bank of America Corporation

Tracked entities: Bank of America and Bank of America Corp.

Exposure type: Direct/core brand with entity variants merged

AI platform families: ChatGPT, Gemini, Google AI Mode, Google AI Overviews, Microsoft Copilot, and Perplexity

Current methodology version: V0

Answer Capsule

Bank of America recorded one of the clearest negative AI recommendation-coverage changes in the initial LLM Authority Index public-company panel.

Across 442 matched prompt-platform cells, Bank of America recommendation coverage declined from 35.3% in July to 26.7% in September 2026, a change of -8.6 percentage points.

The exploratory 95% interval ranged from approximately -12.8 to -4.4 percentage points, remaining fully below zero. All six measured AI platform families declined:

• ChatGPT: -16.13 pp
• Gemini: -13.73 pp
• Google AI Mode: -1.03 pp
• Google AI Overviews: -9.04 pp
• Microsoft Copilot: -1.64 pp
• Perplexity: -25.00 pp

Bank of America therefore meets the V0 Negative AI divergence candidate rules and receives a High AI-measurement confidence classification.

That classification refers only to the measured AI-side movement. It does not mean Bank of America deposits, loans, revenue, earnings, analyst estimates, valuation, or stock performance will decline.

Simple presence also moved lower, from 87.6% to 81.4%, a decline of 6.1 percentage points. However, average recommendation rank improved from approximately 3.95 to 3.49 among responses where Bank of America was recommended.

The resulting pattern is therefore specific:

1. Bank of America appeared in fewer eligible matched responses;
2. it was recommended in a smaller share of those responses;
3. all six platform families moved negatively on recommendation coverage; and
4. when Bank of America was recommended, its average position improved.

That last point matters. Even one of the strongest cross-platform negative recommendation signals in the current panel does not imply every AI visibility metric moved in the same direction.

This is why AI recommendations, mentions, citations, presence, and rank should remain separate measurement layers.

The forward-looking research question remains the one defined by the AI Commercial Momentum Hypothesis: do persistent changes in unbranded AI recommendation behavior contain incremental information about later commercial outcomes after accounting for information already available when the signal was measured?

The company-level signal construction is documented in How We Measure AI Commercial Momentum, while the complete initial public-company panel is preserved in Initial Findings From 25 Public Companies.

Bank of America AI Recommendation Momentum at a Glance

Additional V0 signal properties:

Bank of America also appears in the broader Bank Stocks and AI Search analysis. In that six-company research slice, Bank of America is one of four negative AI divergence candidates and one of two companies, together with Goldman Sachs / Marcus, whose recommendation coverage declined across all six measured AI platform families.

Questions This Section Answers

• How large was Bank of America's recommendation-coverage decline?
• Was the decline broad across AI platforms?
• Why did average rank improve while recommendation coverage fell?

How Large Was Bank of America's Recommendation-Coverage Decline?

Bank of America recommendation coverage declined 8.6 percentage points, from 35.3% in July to 26.7% in September 2026.

The exploratory interval remained below zero, from approximately -12.8 to -4.4 points. Under the V0 framework, that means the observed recommendation change is directionally negative within the current AI-side measurement system.

This result is supported by a relatively large matched panel: 442 matched prompt-platform cells across 334 normalized prompt clusters and all six platform families.

The signal therefore receives High AI-measurement confidence.

High confidence here has a narrow meaning. It indicates that the measured recommendation decline is comparatively well supported by the current matched-panel design and remains stable under the present data-cleaning sensitivities.

It does not mean there is high confidence Bank of America will experience weaker deposit growth, loan growth, revenue growth, earnings growth, or stock performance.

Was the Decline Broad Across AI Platforms?

Yes. Bank of America recommendation coverage declined on all six measured platform families.

The magnitude was not uniform. Perplexity and ChatGPT showed the largest declines, while Google AI Mode and Microsoft Copilot moved only slightly lower.

That distinction matters because cross-platform breadth and aggregate magnitude are separate characteristics.

The cross-platform portability analysis identified Bank of America as one of only three public-company rows in the initial panel whose aggregate recommendation direction was shared by all six platforms.

The other two were Goldman Sachs / Marcus and Happen / LendingClub.

This makes Bank of America a useful future test case for whether broad multi-platform movement contains more information than platform-specific movement. That proposition has not yet been validated.

Why Did Average Rank Improve While Recommendation Coverage Fell?

Average recommendation rank and recommendation coverage measure different things.

Recommendation coverage asks how often Bank of America was actually recommended across eligible matched responses.

Average recommendation rank asks where Bank of America appeared among recommendations in the subset of responses where it was recommended.

From July to September:

• recommendation coverage fell from 35.3% to 26.7%;
• presence fell from 87.6% to 81.4%; but
• average recommendation rank improved from 3.95 to 3.49.

A company can therefore be recommended less frequently while appearing in a better position in the smaller subset of responses where it remains recommended.

This is not contradictory. It is evidence that frequency and position capture different aspects of AI behavior.

For investors and researchers, this matters because a single composite visibility score could hide a pattern like this. Bank of America became less broadly recommended across the matched prompt universe, but its average position improved within the recommendation set that remained.

Why This Matters for Investor Research

Bank of America provides a strong prospective case for the AI Commercial Momentum research program because three AI-side characteristics line up in a relatively unusual way:

• the aggregate recommendation decline is large enough to clear the V0 magnitude threshold;
• the exploratory interval remains below zero; and
• all six platform families move in the same direction.

Those features make the signal easier to define and freeze prospectively than many mixed cases in the panel.

The correct next question is not whether Bank of America is financially weak. The correct question is whether this dated AI-side observation later proves to have any measurable relationship with downstream commercial or financial data.

Potential future comparison variables include:

• branded search demand;
• consumer checking and savings interest;
• credit-card and lending consideration;
• digital traffic and app engagement;
• deposit balances and account growth;
• consumer and small-business loan trends;
• card spending and payment activity;
• Merrill and wealth-management flows where relevant;
• analyst revenue and earnings estimate revisions;
• reported segment revenue and profitability; and
• stock returns only after commercial and financial relationships have been tested.

The purpose is to test incremental information value, not to reverse-engineer a narrative after outcomes are known.

Bank of America Compared With Other Financial-Services Signals

The initial public-company panel provides useful contrast cases.

Axos Financial

Axos Financial moved in the opposite aggregate direction, with recommendation coverage increasing 19.7 percentage points. Its platform pattern was broad but not universal.

This creates a useful positive-vs-negative comparison for later bank-sector validation.

Citi

Citi declined only 1.8 percentage points on aggregate recommendation coverage and remained Mixed / neutral because the interval crossed zero and platform direction was evenly split.

Ally Financial

Ally Financial declined 5.2 percentage points, with five of six platforms worsening and a High AI-measurement confidence classification.

American Express

American Express declined 8.4 percentage points, with five of six platforms worsening. Like Bank of America, American Express also showed an improving average recommendation rank despite lower recommendation coverage.

Goldman Sachs / Marcus

Goldman Sachs / Marcus is another all-six-platform negative case in the initial panel.

These contrasts are why the sector analysis should not be reduced to one average or one ranking table. Different companies show different combinations of aggregate direction, uncertainty, platform breadth, presence movement, rank movement, and entity mapping.

Information Timing and Prospective Validation

Prospective testing requires a strict information cutoff.

Bank of America's official investor-relations materials show that Q2 2026 results were released July 14, 2026, inside the July-to-September observation period. Those results were already public and must therefore be treated as baseline information in any future predictive model.

Bank of America's investor calendar schedules its Q3 2026 earnings conference call for October 14, 2026, after the September AI signal freeze.

That creates a clean prospective checkpoint, but one quarter by itself cannot validate the hypothesis.

Any later comparison should ask whether the frozen AI signal adds information beyond what was already observable through Q2 results, market prices, analyst estimates, macroeconomic conditions, interest-rate expectations, credit conditions, search demand, and other conventional variables.

Current official source pages:

• Bank of America Investor Relations: https://investor.bankofamerica.com/
• Bank of America Events: https://investor.bankofamerica.com/events-and-presentations/events

What This Does Not Mean

The Bank of America result does not establish that:

• deposits will decline;
• consumer or commercial loan growth will weaken;
• card activity will slow;
• Merrill or wealth-management flows will weaken;
• revenue or earnings will decline;
• analyst estimates will fall;
• valuation is too high or too low;
• the stock will underperform; or
• AI systems caused any future commercial outcome.

The term Negative AI divergence candidate is an internal V0 research classification for unusual AI-side recommendation movement under predefined thresholds. It is not an investment recommendation and is not equivalent to the later concept of AI Visibility Market Divergence against investor expectations.

The distinction is developed in AI Visibility Market Divergence and AI Recommendation Share vs. Market Share.

Methodology

The V0 Bank of America signal uses the same core framework described in How We Measure AI Commercial Momentum.

Matched-panel denominator

The primary change estimate compares recommendation outcomes across 442 matched prompt-platform cells present in both the July baseline and September observation.

This controls for changing prompt composition more effectively than comparing unmatched monthly totals.

Prompt-level uncertainty

The exploratory interval is based on change values averaged within normalized prompt clusters. The standard error is calculated across prompt-level means and converted to an approximate normal 95% interval.

For Bank of America, the resulting interval is approximately -12.8 to -4.4 percentage points.

This is an exploratory measurement interval, not a causal confidence interval and not a forecast interval for any financial variable.

Platform breadth

The framework separately records whether recommendation coverage improved, worsened, or remained stable on each of the six platform families.

Bank of America declined on 6 of 6.

Entity normalization

The V0 row merges Bank of America and Bank of America Corp. into one mapped public-parent exposure.

This normalization avoids treating obvious naming variants as economically separate companies.

It does not imply that every Bank of America subsidiary, geography, product, business line, or customer segment is fully represented by the prompt universe.

Sensitivity checks

The no-dedupe sensitivity difference is approximately -0.17 percentage points.

The capture-average sensitivity difference is approximately +0.25 percentage points.

Both are small relative to the -8.6-point primary estimate and remain well within the current High-confidence sensitivity thresholds.

Confidence rules

Bank of America receives a High V0 AI-measurement confidence classification because it has:

• more than 200 matched cells;
• all six platform families;
• a directional exploratory interval; and
• cleaning sensitivities within the current High-confidence thresholds.

Again, this confidence rating applies to the AI-side measurement, not to future company performance.

Limitations

The Bank of America observation has several important limitations.

1. The observation window is short

The current signal compares July and September 2026. A two-endpoint change may not persist.

2. AI systems can change independently

Models, retrieval systems, answer policies, indexes, and product behavior can change during the observation period.

3. Prompt coverage is not the same as customer demand

The prompt universe is commercially oriented but is not a direct sample of Bank of America customers or account openings.

4. Recommendation exposure is not market share

AI recommendation coverage should not be compared directly with deposit share, loan share, card share, or revenue share without denominator alignment.

5. Bank of America is diversified

The company spans consumer banking, global wealth and investment management, global banking, markets, cards, lending, deposits, and other activities. A single consumer-facing AI recommendation measure cannot represent every economic exposure.

6. Cross-platform agreement does not prove causality

Six-platform directional agreement makes the AI-side observation more portable. It does not prove the signal causes or predicts financial outcomes.

7. Already-public information may explain part of the movement

Q2 results and other market information released during the observation period may influence both AI outputs and investor expectations. Future validation must control for those data.

What We Will Test Next

The Bank of America observation is now a frozen prospective signal.

Future validation should test whether the July-to-September AI recommendation decline precedes any measurable change in:

1. branded search and digital demand;
2. consumer banking interest and account acquisition;
3. deposit growth and mix;
4. card and lending activity;
5. segment revenue growth;
6. analyst estimate revisions;
7. revenue or earnings surprise; and only then
8. later excess stock returns.

The preferred design remains a walk-forward framework with conventional financial and commercial variables in the baseline model first, followed by AI variables to test incremental explanatory or predictive value.

A null result is a valid outcome. If Bank of America's broad AI recommendation decline is not followed by commercially relevant changes, that evidence weakens the hypothesis that cross-platform AI recommendation momentum contains useful leading information.

Related LLM Authority Index Research

• Can AI Search Signal Future Revenue Growth? The AI Commercial Momentum Hypothesis
• Initial Findings From 25 Public Companies
• How We Measure AI Commercial Momentum
• AI Investor Signal Tracker
• Bank Stocks and AI Search
• Axos Financial AI Search Visibility
• Citigroup AI Search Visibility
• Ally Financial AI Search Visibility
• American Express AI Search Visibility
• Chime AI Search Visibility
• Goldman Sachs and Marcus AI Search Visibility
• AI Recommendations vs. Mentions vs. Citations
• Cross-Platform AI Recommendation Portability

The live public-company series is maintained in the AI Investor Signal Tracker.