What does schwab.com do and who is it for? Explain it back to me.
- previous resource64%
- web search27%
- prior knowledge9%
The agent successfully assembled a comprehensive overview of Charles Schwab's business, pricing, and competitive positioning by combining direct site fetches (homepage, pricing, accounts, retirement pages) with web searches and prior knowledge. The site itself was moderately navigable—key pages like /pricing and /accounts existed and were reachable, but critical information like exact fee structures, account minimums, and detailed service comparisons were fragmented across multiple pages or required external research. The agent had to supplement site content with Wikipedia, SEC filings, and third-party comparisons to answer the full brief.
- ›The /pricing page (step 3) and /accounts page (step 5) both returned successfully but the HTML responses were truncated in the trajectory; the agent relied on web search results to extract concrete fee data (e.g., $0.65 per options contract, robo-advisor pricing tiers) rather than direct page parsing.
- ›Core service pages (/retirement, /accounts) existed and were found, but the site's navigation structure forced the agent to discover specific offerings (IRAs, checking accounts, wealth management, Intelligent Portfolios) through web searches (steps 9, 15) rather than a consolidated service overview—no single page answered 'what does Schwab do' comprehensively.
- ›Pricing and competitive differentiation (task requirements) were only partially visible on-site; the agent cited external sources (StockBrokers.com, NerdWallet, Bankrate) for the Fidelity/Interactive Brokers/Vanguard comparisons because Schwab's /compare-us page (step 11) did not provide concrete side-by-side data.
- ›The agent's final response explicitly flagged navigation friction: account minimums, monthly fees, advisor fee structures, exact ETF counts, and banking/brokerage relationship clarity were all cited as 'confusing or hard to find,' indicating the site lacked transparent, consolidated disclosure of these details.
- ›Prior knowledge and web search (27% + 9% = 36% of sources per the metadata) were necessary to fill gaps; the site's published pages alone did not provide enough structure to answer 'how is it priced' and 'how is it different' without external context.
Want to run your own?
Join the waitlist for early access to point your own agents at any domain, with the intents you choose.
or talk to us about agent readiness →