Schwab’s Two AI Front Doors
Download case PDF (6 pages)Request teacher’s pack
Teaching case
Information cutoff: October 3, 2026, end of day Dubai time
Audience: MBA students and working professionals
A portfolio of announcements
On September 30, 2026, Charles Schwab announced Charley, a generative AI assistant for eligible U.S. retail clients. The planned October rollout would bring Schwab information, tools, and service into a voice- or text-based conversation on Schwab.com and Schwab Mobile. Clients could ask questions, receive information and suggested next steps, complete select account tasks, and reach a Schwab representative.[1]
The announcement followed a different AI commitment two weeks earlier. Schwab Advisor Services said it was working with Anthropic to bring Claude for Financial Advisors to the 16,000-plus independent registered investment advisors (RIAs) it serves. The proposed advisor experience connects with Schwab Advisor Center and tools used in CRM, custody, portfolio reporting, financial planning, estate planning, and meeting workflows.[2]
For classroom purposes, imagine a cross-functional Schwab leadership group reviewing the two announcements on October 3. The group is fictional; no public source says Schwab had convened it or that a resource-allocation decision was pending. Students are asked to advise on how Schwab should sequence and govern its retail and advisor AI propositions as they move from announcement to implementation.
One company, two users
Schwab’s business spans wealth management, securities brokerage, banking, asset management, custody, and financial advisory services. Its retail clients include self-directed investors and clients who use managed solutions. Separately, Schwab Advisor Services provides custody, trading, and support services to independent advisors; those RIAs are not owned by Schwab.[3]
At June 30, 2026, Schwab reported $13.08 trillion in total client assets and 48.0 million client accounts. During Q2, clients opened 1.4 million brokerage accounts, core net new assets were $119.8 billion, and daily average trades reached a record 11.9 million. Net revenue was $7.072 billion, up 21% from Q2 2025; GAAP net income was $2.8 billion.[4]
The company reported revenue from several sources: $3.357 billion of net interest revenue, $1.825 billion of asset-management and administration fees, $1.215 billion in trading revenue, $333 million in bank deposit account fees, and $342 million in other revenue. These Q2 figures describe Schwab’s company-wide business before Charley’s announcement and do not isolate any AI contribution.[4]
Exhibit 1. Schwab reported scale and Q2 financial context
| Measure | Q2 2026 / June 30, 2026 | What it measures |
|---|---|---|
| Total client assets | $13.08 trillion | Company-reported total at June 30 |
| Client accounts | 48.0 million | Total at June 30 |
| Core net new assets | $119.8 billion | Q2 inflows under Schwab’s definition |
| Average daily trades | 11.9 million | Record average for Q2 |
| Net revenue | $7.072 billion | Q2; +21% year over year |
| Net interest revenue | $3.357 billion | Q2 |
| Asset-management and administration fees | $1.825 billion | Q2 |
| Trading revenue | $1.215 billion | Q2 |
Source: Schwab’s July 21, 2026 Q2 earnings release, “2Q26 Client and Business Highlights,” “2Q26 Financial Commentary,” and consolidated income statement.[4] Dollar amounts are reported in millions in the release. Revenue subtotals are not AI-specific.
The scale raises the stakes on both sides. An assistant used for routine service could touch many client interactions. An advisor workflow tool could affect how independent firms prepare for client meetings and handle records. In each channel, a mistake could consume time, undermine trust, or require correction. Schwab had not publicly disclosed per-product costs, adoption, error rates, productivity effects, or incremental revenues for either new announcement by the case cutoff.
Charley: retail information and selected actions
Schwab described Charley as a phased rollout beginning in October for eligible U.S. retail clients. The release listed help with transfers and payments, trading and account-action questions, account opening and maintenance, troubleshooting, quotes and market information, research tools, balances, transactions, positions, statements, and tax documents.[1]
Charley’s announced selected actions included adding or updating watchlists, enrolling eligible securities in dividend reinvestment, adding beneficiaries, updating contact or eligible trusted-contact details, and setting or managing alerts. The announcement described neither autonomous securities trading nor personalized investment recommendations by Charley. Clients could connect with a representative for additional help.[1]
Schwab said Charley uses large language models grounded in Schwab data and expertise. Its disclosure also said results may be inaccurate or incomplete, users should verify information and confirm account changes, third-party AI providers process inputs as needed, and interactions may be recorded and monitored for quality assurance and service improvement.[1]
Earlier in May, Schwab had rolled out Portfolio Insights, a separate retail feature that summarizes portfolio changes, selected market news, and Schwab investment research. Schwab described that tool as informational and educational, not investment advice or a recommendation; its disclosure warns that AI output can be inaccurate, hallucinated, stale, or incomplete.[5]
Claude for Financial Advisors: support for independent firms
The September 14 announcement positioned Anthropic’s Claude for Financial Advisors around the work of independent advisors. Schwab said the offering was designed to connect to advisors’ existing tools and help with research, meeting preparation, analytics, financial-plan updates, drafting client follow-up for advisor review, and CRM updates. Schwab stated that RIAs would access Schwab Advisor Center through an authenticated connection and that administrative audit logs would be available.[2]
This professional workflow differs from Charley’s retail service channel. The RIA remains an independent business, and an advisor can review AI-generated material before it reaches a client. The announcement did not report adoption, task accuracy, time savings, paid conversion, effect on custody relationships, or the allocation of responsibilities between Schwab, Anthropic, and individual advisory firms. Schwab also did not disclose whether the retail and advisor offerings share technical infrastructure.
The market around the decision
AI-supported service and advisor tools were not new categories. Bank of America said its Erica virtual assistant had surpassed 3 billion client interactions by August 2025, with more than 58 million interactions per month at that time. It also described employee-facing assistance and other AI capabilities.[6] Erica’s longer operating history shows that a financial institution can build an AI interface into a service model, but it is not a direct measure of the adoption or economics of Schwab’s new generative AI products.
Morgan Stanley Wealth Management announced in March 2023 that it was developing an internal OpenAI-powered assistant to search and synthesize its own research for financial advisors, with source-document links. The company framed the assistant as advisor-facing and intended to support the human advisor relationship.[7] That announcement establishes an earlier professional-channel precedent; it does not establish comparative product performance against Schwab’s announced Anthropic integration.
Schwab also surveyed its own clients and prospects. A September survey of more than 500 respondents with at least $10,000 in retail assets found that 65% believed AI could help with routine tasks or play a meaningful role in investing when paired with human expertise. Schwab reported that respondents were most comfortable with AI for questions, account information, and routine updates, and preferred a person for more meaningful account actions.[1] The survey measures stated views in a Schwab-selected, asset-screened sample—not Charley use or realized outcomes.
In a separate Schwab Advisor Services study, 63% of 533 surveyed Schwab-custodied RIAs reported using AI in some capacity. Of AI users, 82% said they used generative AI; Schwab reported that most firms remained in early stages, often using AI for administrative work such as note-taking and email drafting.[8] Schwab said the survey was voluntary and uncompensated, and the sample was limited to advisors who custody at Schwab. The survey does not establish demand for, or adoption of, the Anthropic integration.
Exhibit 2. Two announced propositions and their disclosed boundaries
| Retail: Charley | Independent advisors: Claude for Financial Advisors | |
|---|---|---|
| Announced | Sept. 30, 2026 | Sept. 14, 2026 |
| Intended user | Eligible U.S. retail clients | Independent RIAs Schwab serves |
| Disclosed tasks | Questions, Schwab/account information, suggested next steps, selected account updates, representative escalation | Research, meeting prep, analytics, financial-plan updates, client follow-up drafts for advisor review, CRM updates |
| Connection | Schwab.com and Schwab Mobile, voice or text | Authenticated access to Schwab Advisor Center and advisor tools |
| Stated human/control element | Client can reach a Schwab representative; user is told to verify information and account changes | Advisor review of client follow-up; administrative audit logs |
| Status on Oct. 3 | Phased October rollout announced; broad uptake not reported | Partnership announced; uptake and realized results not reported |
| Important unknowns | Accuracy, adoption, containment, escalations, cost, retention or revenue effects | Accuracy, adoption, time saved net of review, custody effects, economics and implementation responsibility |
Sources: Schwab Charley announcement; Schwab/Anthropic announcement.[1][2] The table separates each offering’s announced scope; it does not imply a common model, architecture, or maturity.
An open strategic choice
Schwab’s task is not simply to decide whether to use AI. It has announced two applications with different users, data contexts, possible harms, and paths to human review. The retail assistant might improve access to routine service and make Schwab’s tools easier to navigate. The advisor integration might help independent firms use existing workflows more efficiently and strengthen the usefulness of Schwab’s custody platform. A shared capability layer could create reuse, but it could also make different permission and review needs harder to manage.
The evidence available on October 3 does not establish which proposition will produce more value. Company-wide Q2 growth predates Charley and cannot be attributed to it. The surveys report stated attitudes and existing experimentation, not outcomes from either launch. Schwab has not publicly supplied the costs, task-level performance, client effects, or advisor productivity data needed for an ROI comparison.
For this classroom exercise, assume leadership has finite management attention and must choose a sequence for the next implementation phase. This is a teaching assumption, not a reported Schwab constraint. The group can consider three postures:
- Retail-first: concentrate on improving Charley’s task completion and service reliability, keeping its authority narrow while data accumulate.
- Advisor-first: concentrate on implementing the Claude integration in RIA workflows, where a professional can review outputs and Schwab can learn from authenticated, auditable use.
- Advance both with distinct gates: invest in reusable foundations where appropriate, while maintaining separate models, data permissions, human-review rules, and release criteria for retail and advisor use.
Each posture has costs. Retail-first could focus on a large client base but exposes Schwab directly to customer-facing errors. Advisor-first could help custody clients and place a professional in the review loop, but depends on adoption and workflow fit across independent firms. Advancing both could preserve strategic options and reuse controls, but spread leadership attention before either product has demonstrated results. None can be evaluated from announcement claims alone.
Decision assignment
You are advising the fictional leadership group on October 3, 2026. Recommend a sequence for Schwab’s retail and independent-advisor AI propositions. State:
- which customer and use cases should come first, and which should wait;
- how the permissions and human-review boundaries should differ between Charley and the Anthropic integration;
- what evidence Schwab should collect before widening either proposition;
- which measures would demonstrate client or advisor value, operational efficiency, and acceptable risk; and
- what findings should pause, narrow, or expand a rollout.
Discussion questions
- How do Charley and the advisor integration fit Schwab’s separate retail and RIA value propositions?
- What capabilities or controls might be reusable across the two offerings, and which should remain channel-specific?
- How should Schwab distinguish routine information support, account maintenance, portfolio insight, and client-specific advice or action?
- Which objective measures and user perceptions should Schwab track for each channel? What comparison would make those measures informative?
- What evidence would justify scaling one offering before the other? What evidence would justify advancing both?
Sources
- Charles Schwab, “Meet Charley, Schwab’s New AI-Powered Assistant Designed to Help Clients Get Answers and Take Action,” Sept. 30, 2026, official announcement, product scope, rollout and disclosures.
- Charles Schwab, “Charles Schwab and Anthropic to Bring Claude to Independent Registered Investment Advisors,” Sept. 14, 2026, official announcement, offering scope, authenticated access, tasks and audit logs.
- The Charles Schwab Corporation, Q2 2026 Form 10-Q, period ended June 30, 2026, SEC filing, cover and company description. The filing was not used for unquoted AI product claims.
- Charles Schwab, “Schwab Reports Record Quarterly Revenue and Earnings,” July 21, 2026, official earnings release, Q2 highlights, financial commentary and consolidated income statement.
- Charles Schwab, “Charles Schwab Launches AI-Powered Capability That Helps Investors Understand Portfolio Performance and Market Activity,” May 5, 2026, official announcement, Portfolio Insights scope and disclosure.
- Bank of America, “A Decade of AI Innovation: BofA’s Virtual Assistant Erica Surpasses 3 Billion Client Interactions,” Aug. 20, 2025, official release. Company-reported usage, not independently verified and not comparable to Schwab’s announced products.
- Morgan Stanley Wealth Management, “Key Milestone in Innovation Journey with OpenAI,” Mar. 14, 2023, official release, advisor-facing assistant announcement and intended source-linked research workflow.
- Charles Schwab Advisor Services, “Schwab Study Reveals RIA AI Adoption More Than Doubles—But Most Firms Still in Early Stages,” Jan. 22, 2026, official release, results and study disclosure.
This case uses public information available through October 3, 2026. The leadership group, finite-attention constraint, choice set, and decision assignment are constructed for classroom discussion; they do not describe a disclosed Schwab meeting, budget, or management intention.