← back to rankingBAC · Bank of America Corporation
Banks - Diversified · mkt cap $372.4B · calls: Q1 FY2026 vs Q4 FY2025
41.0 conviction · conf-adj 38
conf 3/10 partial
enthusiasm:24.0 · trend:8 · quantifies:5 · impact:0 · under_radar:0 · credibility:0 · business_impact:8 · disruption:0 · commitment:-4 · confirmation:0
Enthusiasm latest 8 / prev 7 (rising)
The AI thesis is mainly productivity and efficiency: management says AI helps “reduce manual work, lowered unit costs” and creates “positive pressure on the earnings.” Credibility is supported by specific headcount/productivity claims, including “30% out of the coding part” and “saves us about 2,000 people,” but revenue upside is less quantified and framed as future “greater market share and capabilities.”
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $191.6B · net income $30.5B · net margin 15.9% · diluted EPS 3.82
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: 0.0% · next-FY EPS uplift: 1.42% · vs analysts: behind · priced in: high · confidence: 3/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
90 AI installations live
other · soft | 90 installations | Count of deployments, no $ attached. Cannot translate to revenue or EPS without per-installation economics. | | |
200,000 teammates have AI access
other · soft | 200,000 teammates | Adoption reach, not a financial figure. Access != quantified saving. No arithmetic possible. Also serves as denominator for the coding-savings cost proxy. | | |
M365 Copilot across 200,000 teammates
productivity · soft | 200,000 teammates | 'Expect good leverage' is unquantified. No productivity % or $ disclosed, so no defensible saving can be computed. | | |
Erica worth 'thousands of teammates' avoided
cost · soft | thousands of teammates | Illustrative @ ~3,000 FTE x ~$150k = ~$450M gross, ~$355.5M after-tax = ~1.17% of $30,509M NI. But 'thousands' is vague and reflects today's run-rate (already in base), so incremental next-FY contribution ~0. Vague -> null/soft, no invented number counted. | | |
18,000 coders exposed to AI productivity
productivity · soft | 18,000 people | Denominator/context for the coding-savings claim; no standalone $ figure. Not separately quantifiable; 2,000/18,000 = 11.1% of coding workforce. | | |
30% taken out of coding workstream = ~2,000 people saved
productivity | 30% / 2,000 people | The single HARD figure. Two cost-per-teammate proxies: X = (gross profit $107.422B - operating income $37.695B)/200,000 = $348,635 -> 2,000*$348,635 = $697.27M gross, $550.84M after-tax = 1.806% of $30,509M NI; Y = $200k/FTE -> $400M gross, $316M after-tax = 1.04%. Both defensible; average ~1.42%. Stated past tense ('we've taken') so largely in 2025 base -> genuine incremental next-FY portion smaller. rev 0. | 0 | 1.42 |
AI expenditure 'several hundred million'
cost · soft | several hundred million $ | Spend, not a saving -> earnings HEADWIND, not uplift. Midpoint ~$400M pre-tax, ~$316M after-tax = ~-1.04% of NI, roughly offsetting the coding saving at today's run-rate. Vague exact figure -> null/soft, not counted in aggregate. | | |
15-20 AI projects underway
other · soft | 15-20 projects | Pipeline count, no $ per project. 'Laundry list of much bigger size' is forward and unquantified. | | |
99% of consumer interactions already digital
engagement · soft | 99% | Frames the automation opportunity but no take-rate, conversion, or revenue $ attached. Soft. | | |
90,000 sales force moving to agent force/AI
productivity · soft | 90,000 sales force | Forward 'we'll get more efficient' with no efficiency % or $ disclosed. No defensible number -> soft. | | |
Assumptions: Percentages reported as percent numbers (1.42 = 1.42%). Tax rate 21%. No incremental revenue margin used — every quantified claim is internal cost/productivity, not incremental revenue, so impact is bottom-line via after-tax saving / current NI ($30,509M). Fully-loaded cost/teammate proxied two ways: $348,635 (gross profit minus operating income / 200,000 teammates, grounded in the base) and ~$200k (assumed tech-talent loading); the hard coding saving averages these to ~1.42%. Phasing: the coding and Erica savings are described in present/past tense = largely already in the FY2025 base, so genuine next-FY incremental contribution is below the gross figures. 'Several hundred million' AI spend is a near-term ~-1% after-tax headwind funding these savings; the two roughly offset at today's run-rate. No revenue dollar claim was made.
Top line: Effectively zero quantifiable revenue uplift. No management AI claim attaches a dollar to revenue — all are internal cost/productivity (coding, Erica, Copilot) or unquantified efficiency (90k sales force) and reach (99% digital, 200k teammates). Forward 'drive greater market share' is asserted but unsized, so est_rev_uplift_pct = 0 with soft, unquantified upside only.
Bottom line: A pure adopter, bottom-line story. The one HARD figure — 30% of coding = ~2,000 FTE — is ~1.42% of $30.5B NI averaging two cost proxies ($697M gross/$550.8M after-tax at $348,635/teammate vs $400M/$316M at $200k). Critically, the 'several hundred million' AI spend is a ~-1% after-tax headwind that roughly offsets the coding saving at today's run-rate, and much of the saving is already in the FY2025 base; net genuinely-incremental next-FY adopter EPS uplift is therefore modest (~1-1.4%). Erica, Copilot, and broader AI access are directionally positive but unmodeled.
Consensus 2026 EPS ~$4.46 vs 2025 ~$3.82 = +16.8%; consensus revenue +9.4% YoY already bakes in strong positive operating leverage — exactly the bucket these AI savings fall into. The disclosed AI-specific EPS impact (~1.4% gross, partly offset by AI spend, much already in the base) is only ~10% of the ~17% EPS growth analysts assume and sits comfortably inside it, so it is priced in. (Note: supplied current revenue $191.567B vs consensus 2025 revenue ~$110.6B suggests the consensus set is on a different/segment basis, lowering revenue-comparison quality.)
MODEL CONSENSUS (impact)
partial
Both agree: pure adopter, 0% revenue uplift, priced-in high. Reconciled the lone hard coding saving to ~1.42% EPS (avg) and took the more conservative 'behind' verdict.
Conflicts reconciled
- est_eps_uplift_pct: X=1.806 vs Y=1.04 -> used 1.42 (average; X has grounded cost proxy, Y phasing-adjusted, both defensible)
- vs_analyst_expectations: X=behind vs Y=inline -> used behind (more conservative on a near-tie, confidence lowered)
- supplier_rev_uplift_pct: X=0 vs Y=null -> used 0 (no supplier-side claims exist)
- Erica eps_uplift_pct: X=null vs Y=1.17 -> used null ('thousands' is vague; don't invent)
- convention: X used fractions, Y used percent -> standardized to percent numbers
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | 0 | 0 |
| EPS uplift % | 1.0 | 0.018055108328689895 |
| Priced in | high | high |
| vs analysts | inline | behind |
| Confidence | 4 | 4 |
| Top line | Effectively zero quantifiable revenue uplift. No management AI claim attaches a dollar to revenue — all are internal cost/productivity (coding, Erica, Copilot) or unquantified efficiency (90k sales force) and reach (99% digital, 200k teammates). Forward 'drive greater market share' is asserted but unsized, so est_rev_uplift_pct = 0 with soft, unquantified upside only. | No hard adopter-side AI revenue claim was disclosed, so modeled next-FY revenue uplift is 0% of $191.567B. The 99% digital-interaction and 90,000 sales-force claims are potentially relevant to revenue, but lack conversion or volume math. |
| Bottom line | This is a pure adopter, bottom-line story. The one HARD figure — 30% of coding = ~2,000 FTE — is ~$400M gross / ~$316M after-tax = +1.04% of $30.5B NI. Erica's 'thousands' adds an illustrative ~1.17% but is vague and already in the run-rate. Critically, the 'several hundred million' AI spend is a ~-1.04% after-tax headwind that roughly offsets the coding saving at today's run-rate — management's thesis is that savings outgrow spend over time ('operating leverage will be there'). Net genuinely-incremental next-FY adopter EPS uplift is therefore modest, ~1%. | Hard quantified impact is the coding labor saving: 2,000 people * $348,635 opex proxy = $697.270M pretax, $550.843M after tax, or 1.806% of $30.509B net income. Erica, Copilot, and broader AI access are directionally positive but unmodeled. |
| Reasoning | Consensus 2026 EPS $4.46 vs 2025 actual $3.82 = +16.8%; consensus revenue +9.4% YoY — already bakes in strong positive operating leverage, which is exactly the bucket these AI savings fall into. The disclosed AI-specific EPS impact (~1%, and partly offset by AI spend, with much already in the base) is a small fraction of the ~17% EPS growth analysts assume and sits comfortably inside it. (Note: given consensus revenueAvg ~$111-121B vs the $191.6B reported base, the consensus set appears to be on a net-revenue basis; claim math uses the supplied $191.6B base as instructed.) Nothing in the quantified claims points clearly above embedded expectations. | Hard AI uplift is 0.0% revenue and 1.806% EPS. Consensus EPS rises from $3.81486 in 2025 to $4.46068 in 2026, or +16.929%; versus current EPS of $3.82, 2026 consensus implies +16.772%. The modeled hard AI EPS uplift is only about 10.7% of that consensus EPS growth. Consensus revenue growth from 2025 to 2026 is +9.413%, while hard AI revenue uplift is 0%; revenue comparison is low quality because supplied current revenue is $191.567B versus consensus 2025 revenue of $110.576B. |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
AI installations: 90 installations (Q1 FY2026 current, bottomline)
“And we've got 90 installations working, all 200,000 teammates have access to AI or can use it every day.”
AI access: 200,000 teammates (Q1 FY2026 current, bottomline)
“And we've got 90 installations working, all 200,000 teammates have access to AI or can use it every day.”
Copilot rollout: 200,000 teammates (Q4 FY2025 current, bottomline)
“under the 365 CoPilot rollout, which is now out across a total of 200,000 teammates and using it and learning from it, we expect to get good leverage of that.”
Erica labor substitution: thousands of teammates (Q4 FY2025 current, bottomline)
“today's activity in Erica in our consumer business alone is worth thousands of teammates that we don't have to have to do the great work we do for the customers.”
Coding workforce exposed to AI productivity: 18,000 people (Q4 FY2025 current, bottomline)
“we have 18,000 people on the company's payroll who code.”
Coding productivity: 30% (2025, bottomline)
“And we've using AI techniques. We've taken 30% out of the coding part of the stream of introducing a new product to service or change that saves us about 2,000 people.”
Coding labor savings: 2,000 people (2025, bottomline)
“And we've using AI techniques. We've taken 30% out of the coding part of the stream of introducing a new product to service or change that saves us about 2,000 people.”
AI expenditure: several hundred million dollars (Q4 FY2025 current, bottomline)
“I don't know off the top of my head the total expenditure, but it's several hundred million dollars.”
AI projects: 15, 20 projects (Q4 FY2025 current, bottomline)
“So there's, I don't know, 15, 20 projects going on, and there will be a laundry list of much bigger size as we go through the company now and are generating ideas now that people are using it and getting used to how to use it.”
Digital consumer interactions relevant to AI automation opportunity: 99% (Q1 FY2026 current, both)
“99% round numbers of all the interactions we have with our consumers are digital already.”
Sales force moving to agent force and AI: 90,000 sales force (Q1 FY2026 forward, both)
“But even on the customer facing with 90,000 sales force moving to agent force and AI and we'll get more efficient on that, too.”
PAST (realized)
- We saw productivity improvements through AI and digitalization more generally, and those enabled us to add client-facing associates as we eliminated work and roles in our operational support areas.
- today's activity in Erica in our consumer business alone is worth thousands of teammates that we don't have to have to do the great work we do for the customers.
- we have 18,000 people on the company's payroll who code. And we've using AI techniques. We've taken 30% out of the coding part of the stream of introducing a new product to service or change that saves us about 2,000 people.
CURRENT (now)
- The continued digitization of activities by our clients and inside our company, the application of artificial intelligence that detailed process reengineering all help reduce manual work, lowered unit costs, limited increase in our base cost structure.
- we maintain our sharp focus on operating leverage, including expanding our use of technology and AI to improve operational efficiency and sales effectiveness.
- AI gives us pieces to go, we haven't gone. And we've got 90 installations working, all 200,000 teammates have access to AI or can use it every day.
- under the 365 CoPilot rollout, which is now out across a total of 200,000 teammates and using it and learning from it, we expect to get good leverage of that.
FORWARD (guidance)
- there will always be positive pressure on the earnings due to the application technology and AI gives us a lot of efforts there.
- we feel very strongly that we will not only take advantage of AI will help us drive greater market share and capabilities in the future.
- the trends will be more technology, more intimacy with the customers more agentic versus prompt more built into the process rather than have it be delivered by teammates doing something, it's part of the process, more customers doing more with us and expense will be -- and the operating leverage will be there.
- Next year, we should get more out of it as we figure out and apply it across.
TRACK RECORD — PROMISE vs DELIVERY
—/100 (no quantified promises) no-quantified-promises 6 calls reviewed
Bank of America showcases substantial real AI deployment (Erica at ~58M interactions/month, ~17,000 developers on AI coding tools at 10-15% savings, 250+ ML models, AI-assisted headcount reduction), but consistently frames these as current achievements rather than quantified forward-looking AI targets with both a number AND a future date, so there is no judgeable AI promise to score. (Note: the Q3 FY2025 transcript in this set is actually Banc of California, an unrelated company, and was disregarded.)
PRICED-IN (REFINED)
HIGH (already in)Est. revisions rising · Fwd P/E 16.4 · EV/Sales 2.9x
AI claim maps to Consumer Banking Segment, Global Wealth and Investment Management Segment, Global Banking Segment
Analyst mix has migrated upward over recent months, with fewer holds and a higher buy/strong-buy share, while forward EPS and revenue estimates imply solid multi-year growth already embedded in consensus. Price targets are roughly flat rather than being aggressively raised, but the valuation is rich for a diversified bank at 16.4x forward earnings and 2.9x EV/sales. AI benefits would most plausibly show up through efficiency, personalization, underwriting, advisory, and client-service gains in Consumer Banking, Global Wealth and Investment Management, and Global Banking. Rising estimates make the thesis more priced-in, and combined with a premium valuation this points to high priced-in risk.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
5Q4 FY20245Q1 FY20258Q2 FY20251Q3 FY20257Q4 FY20257Q1 FY2026
AI enthusiasm across 6 calls — trend ↗ rising
BAC moved from Erica/digital metrics to explicit AI-driven productivity and cost leverage, with one non-BAC call showing no relevant story.
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
7/10 qualitative impact material near-term · mixed evidence
Where AI matters: operating efficiency, software development, customer service, sales effectiveness
BAC has real deployed AI at enterprise scale: 90 installations, 200,000 teammates with access, Erica labor substitution, and a hard coding-productivity claim worth roughly 2,000 people. The upside is mainly cost and operating leverage rather than clearly incremental revenue, so it is material for earnings efficiency but not transformational to the banking model.
Caveats: AI savings may already be embedded in the cost base and consensus operating leverage; AI spend of several hundred million dollars offsets part of the gross productivity benefit; Revenue and market-share upside is asserted but not quantified; Model, cyber, privacy, and regulatory risks are high in banking
AI DISRUPTION / CANNIBALIZATION RISK two-sided · 3/10
AI can pressure parts of banking through AI-native fintech interfaces, automated advice, lower servicing costs, and commoditized basic financial guidance. But BAC's core model is still protected by deposits, regulation, balance sheet scale, distribution, trust, and risk controls, so AI is more competitive pressure than core cannibalization.
OPTIONS / MARKET STRUCTURE
option liquidity: good
proxy inputs — dollar-ADV $2.0B · beta 1.221 · px $52.48
source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.
CONFIRMATION — INSIDERS · 13F · LANGUAGE
Undercutting — insiders selling, institutions flat, management language 3/10 hedged.
INSIDERS selling 7 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) flat as of 2026-03-31: 187 new / 345 closed positions; 1544 increased / 1522 reduced; institutional ownership -1.68pp; -158 net 13F holders
MGMT LANGUAGE 3/10 hedged AI was barely discussed; one firm productivity claim, but no AI-specific figures, timelines, or detailed ownership.
commit “the application of artificial intelligence that detailed process reengineering all help reduce manual work, lowered unit costs”
hedge “we're able to convert scale, productivity and macro tailwinds and operating leverage over time”
VERBATIM AI QUOTES
“The continued digitization of activities by our clients and inside our company, the application of artificial intelligence that detailed process reengineering all help reduce manual work, lowered unit costs, limited increase in our base cost structure.”
— Brian Moynihan, Q1 FY2026
“we maintain our sharp focus on operating leverage, including expanding our use of technology and AI to improve operational efficiency and sales effectiveness.”
— Alastair Borthwick, Q1 FY2026
“AI gives us pieces to go, we haven't gone. And we've got 90 installations working, all 200,000 teammates have access to AI or can use it every day.”
— Brian Moynihan, Q1 FY2026
“we're still in the early stages of what all this will do, but we're seeing real benefits out of it today.”
— Brian Moynihan, Q1 FY2026
“we are a beneficiary of the impacts of all technology, including AI.”
— Brian Moynihan, Q1 FY2026
“there will always be positive pressure on the earnings due to the application technology and AI gives us a lot of efforts there.”
— Brian Moynihan, Q1 FY2026
“we feel very strongly that we will not only take advantage of AI will help us drive greater market share and capabilities in the future.”
— Brian Moynihan, Q1 FY2026
“I think AI really helps us internally just to make it straightforward.”
— Brian Moynihan, Q1 FY2026
“the trends will be more technology, more intimacy with the customers more agentic versus prompt more built into the process rather than have it be delivered by teammates doing something, it's part of the process, more customers doing more with us and expense will be -- and the operating leverage will be there.”
— Brian Moynihan, Q1 FY2026
“We saw productivity improvements through AI and digitalization more generally, and those enabled us to add client-facing associates as we eliminated work and roles in our operational support areas.”
— Alastair Borthwick, Q4 FY2025
“Beyond that, productivity improvements from AI and digitalization continued to help offset higher wages, benefits and technology investments.”
— Alastair Borthwick, Q4 FY2025
“today's activity in Erica in our consumer business alone is worth thousands of teammates that we don't have to have to do the great work we do for the customers.”
— Brian Moynihan, Q4 FY2025
“under the 365 CoPilot rollout, which is now out across a total of 200,000 teammates and using it and learning from it, we expect to get good leverage of that.”
— Brian Moynihan, Q4 FY2025
“we have 18,000 people on the company's payroll who code. And we've using AI techniques. We've taken 30% out of the coding part of the stream of introducing a new product to service or change that saves us about 2,000 people.”
— Brian Moynihan, Q4 FY2025
“I don't know off the top of my head the total expenditure, but it's several hundred million dollars.”
— Brian Moynihan, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Glenn Schorr): where are you in the AI journey in terms of how that might bring a little bit larger headcount reduction or not replacing all the attrition going forward?
A: AI gives us pieces to go, we haven't gone. And we've got 90 installations working, all 200,000 teammates have access to AI or can use it every day. Erica is more understood out there, but it's been brought across lots of platforms that the models. And so -- we're still in the early stages of what all this will do, but we're seeing real benefits out of it today.
Q (Q1 FY2026, Michael Mayo): why is Bank of America and AI beneficiary? And if you could just frame it, somewhere I know that the question was asked already, but revenues per employee, how much would you expect that to increase or something around that?
A: we are a beneficiary of the impacts of all technology, including AI. We've applied it and we'll continue to apply our team's job, and we've got catalyst efforts going on, on a corporate wide basis to bring all the ideas to bear. Our team's job is to benefit from the technology.
Q (Q1 FY2026, Michael Mayo): How to use AI to improve the trust of customers, whether it's with a cyber risk. I'm not sure if you were of the CEOs went down to DC or just trust with data and identity and the relationships?
A: we keep our data out of the models, we keep -- so that our customer data, et cetera, and take advantage of the models coming into us, but not feeding them. And that's what we owe our customers.
Q (Q1 FY2026, Michael Mayo): And then last follow-up, just in 5 years from now, when we look back, say, okay, AI, tech, where should we see the benefits? Is it just the stickiness of the relationship? Or is it efficiency or where should that show up?
A: I think AI really helps us internally just to make it straightforward. 99% round numbers of all the interactions we have with our consumers are digital already. So there's no person involved. So as you start to think about that, you go the inverse, the cost of that 1% is a pretty high number, and we're working on that with all this technology, and we're working on the efficiency even of the 99% in house delivered.
Q (Q4 FY2025, Michael Mayo): If you could just give more of an update on technology. What do you expect your spend to be this year versus last year, your spend on AI?
A: under the 365 CoPilot rollout, which is now out across a total of 200,000 teammates and using it and learning from it, we expect to get good leverage of that.
Q (Q4 FY2025, Michael Mayo): And specifically on AI investments, like how much do you spend on that? Or the number of people, if you could dimension that and kind of what kind of outcomes you're looking for, especially as we say here at the start of the year?
A: we have 18,000 people on the company's payroll who code. And we've using AI techniques. We've taken 30% out of the coding part of the stream of introducing a new product to service or change that saves us about 2,000 people.