Kepler’s models extract the right line item from a 10-K 94% of the time. Frontier models alone score 38-46%. A recent Anthropic customer profile details the architecture.
Kepler, the verifiable AI platform for financial research, expanded on the architecture documented in a recent Anthropic customer profile, “How Kepler built verifiable AI for financial services with Claude.” The profile details a system built on a single thesis: language models alone cannot meet the rigor financial workflows require, and traditional financial software cannot match the flexibility analysts need from AI. Kepler’s architecture combines them. That combination is what makes the answers defensible in an IC memo or under audit.
“One wrong assumption early in a financial analysis breaks everything downstream”Share
Without an architecture around them, language models fabricate numbers, drop constraints mid-analysis, and produce answers no analyst can defend in an investment committee. Traditional financial software is rigid, slow to build, and breaks the moment a question departs from a pre-defined template. Other industries have moved on AI. Finance has held back, because no AI system has been able to clear its bar for verification.
Kepler’s architecture splits the work. Language models interpret the analyst’s question, decompose it into a plan, and generate the final narrative. Code retrieves the data, runs the calculations, and renders the citation. A financial ontology mediates between the two. It is the dictionary that maps the language analysts use (“EBITDA,” “free cash flow,” “segment revenue”) to the exact line items in the underlying filings. Every formula is explicit. Every calculation is reproducible. The model never invents a number.
Kepler’s specialized models read a 10-K and identify the right line item 94 percent of the time. Frontier models alone score 38 to 46 percent on the same task. The platform indexes 26 million SEC filings, 50 million additional public documents, and 1 million private documents across 14,000 companies and 27 global markets. Every figure on every page clicks through to its source.
“In finance, the model can’t be the whole system,” said Dr. John McRaven, CTO of Kepler, in the Anthropic profile. “We treat it as one stage in a pipeline whose job is to hand the model exactly what it needs to succeed at exactly that stage. Prompt engineering optimizes a call. Content engineering optimizes the system around it.”
“On our workloads, Claude was the model that consistently held the plan together. Other models would start strong and then quietly drop a constraint by step five,” said Vinoo Ganesh, CEO of Kepler. “That behavior matters more than any benchmark score. One wrong assumption early in a financial analysis breaks everything downstream.”
For the buy-side analyst, that architecture changes three things in the daily workflow. Every figure in a Kepler answer clicks through to its filing, page, and line item, so a number in an IC memo can be defended at the level of the underlying 10-K. Calculations are explicit and reproducible, so running the same query twice returns the same number. Outputs are auditable end-to-end, so compliance reviews and examiner requests do not require redoing the work. Buy-side analysts at private equity firms, hedge funds, and investment banks use Kepler today. The platform was built in under three months, informed by interviews with 147 financial firms. The analysts Kepler serves want answers they can put their name on. The company is continuing to scale across the buy-side research stack, with expansion into private credit underway.
