Read a bank against
a fair peer set.
Peer Lens runs an AI agent to identify a defensible peer group from FDIC regulatory data, compares it across 44 metrics, and an advisor-backed model writes CFO-grade takeaways - interpretation by AI, every figure grounded in the Call Report.
Agentic peer identification
An AI agent assembles candidates from FDIC data - asset band, business model, Fed region, proximity - then an advisor model pressure-tests whether the set is defensible. Sweep and trust banks are screened out; swap any member by hand.
44 regulatory metrics
Scale, profitability, funding, capital, and asset quality - FDIC Call Report-derived and bank-level. Computed in code, not guessed, for clean comparability.
Direction-aware
Efficiency, cost of funds, NIE/assets, and charge-offs read better-when-lower. Percentile is position, not virtue.
AI analysis, grounded
An AI analyst - an executor model with a stronger advisor - writes mechanism-level findings, watch items, and caveats. Every figure is cited from the computed stats; the model never invents one.
Regulatory-first, screened for outliers,
and defensible to a CFO.
Start from FDIC BankFind. An agent screens candidates to a 0.5x-2.0x asset band, gates on business model and Fed region, ranks by proximity, and an advisor model checks the set's defensibility - sweep and trust banks fall into a tray you can pull back in. The metrics are computed in code and grounded in the Call Report; the AI interprets them, it never invents a number.
Pick. Compare. Read.
Pick a target
Type a bank name or FDIC CERT. Ambiguous names disambiguate by certificate.
Agent builds the set
An AI agent proposes a defensible ~11-bank group and an advisor model pressure-tests it; remove, swap, or pull screened-out banks back in.
Read the AI analysis
Percentile strips, a performance-frontier scatter, and AI-written CFO-grade findings - every figure traced to FDIC.
Benchmark a bank in under a minute.
Peer Lens replaces the hand-built comp sheet with agentic peer identification and AI analysis, grounded in regulatory data.