Centauri AI

What do they actually do

Centauri AI is a hosted software product that turns finance documents (credit agreements, servicer reports, spreadsheets, presentations) into structured, reusable data and analyst workflows. Teams upload files, the platform extracts key terms into rows and columns, and analysts review and correct outputs before using them for monitoring, reporting, and portfolio analysis. Their public site shows an app, docs, and login, indicating an active SaaS offering (homepage, docs).

Today it’s used by private credit and structured‑finance teams to automate extraction from credit agreements and portfolio documents. In a published benchmark, they report over 92% extraction accuracy on real agreements and a reduction in processing time from about 30 minutes to roughly 20 minutes per file, with large reductions in manual typing (credit‑agreements case study). They’ve also implemented fine‑grained access controls suitable for enterprise governance, signaling fit for regulated institutions (Permit.io write‑up).

Who are their target customer(s)

• Private credit / structured‑finance analysts: They manually copy clauses and numeric terms from long PDFs and Excel reports into spreadsheets, which is slow, error‑prone, hard to audit, and difficult to scale across many deals.
• Portfolio managers at alternative‑asset funds: They need timely, comparable metrics across positions, but inconsistent document formats make covenant monitoring and portfolio health tracking fragmented and delayed.
• Investment operations and data/ETL teams: They spend time validating and reworking manual extractions from vendor reports and inconsistent files instead of building reliable pipelines and analyses.
• Risk, compliance, and audit teams at institutional investors and banks: They require traceability and strict permissions, needing reproducible extraction histories, explainable decisions, and fine‑grained access controls to meet internal and regulatory standards.
• Deal / diligence teams: High‑volume contract review during transactions creates bottlenecks; manual processes risk missing nonstandard clauses or amendments and can delay deal timelines.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run high‑touch, time‑boxed pilots via founder networks and YC introductions, targeting private‑credit and structured‑finance teams and anchoring outreach with the credit‑agreement case study to show concrete value (case study, YC listing).
• First 50: Convert successful pilots to paid contracts, hire an SDR/AE pair to replicate outreach into similar funds and deal teams, publish 1–2 detailed case studies, and streamline procurement with standardized security/compliance materials and references (enterprise access controls).
• First 100: Productize repeatable automations and templates for faster onboarding, add self‑serve and metered pricing for smaller teams, and build integrations with servicers and vendor platforms; support with a lightweight CS function and recurring industry content to drive referrals.

What is the rough total addressable market

Top-down context:

Centauri serves institutional investment teams that need to convert credit and alternative‑asset documents into structured, auditable datasets. The market is a slice of document automation and data platforms for financial services, driven by the number of credit/alt‑asset teams and the analyst seats they can replace or augment.

Bottom-up calculation:

Assuming 2,000 target institutions globally adopt the product, with an average of 20 analyst/ops seats per customer at $3,000 per seat per year, the core TAM is roughly $120M annually (2,000 × 20 × $3,000). This excludes adjacent legal and banking teams that could expand the market.

Assumptions:

• ~2,000 relevant institutions globally across private‑credit funds, alt‑asset managers, and banks with credit/structured‑finance workflows.
• Average 20 recurring users per organization (analysts, ops, risk/compliance).
• Average $3,000 per seat per year; platform fees and add‑on modules not included.

Who are some of their notable competitors

• Eigen Technologies: Document‑AI vendor used by banks and asset managers to extract data from complex loan, regulatory, and legal documents; competes on accuracy and governance for contract‑level extractions (overview, acquisition note).
• Kira Systems (Litera): Longstanding contract‑review tool for law firms and corporate legal teams, strong in M&A/diligence workflows; less focused on portfolio monitoring or finance‑native analytics (product, acquisition coverage).
• eBrevia (DFIN): AI contract‑analytics tool feeding data rooms and deal workflows; overlaps on extraction for diligence/reporting but sells more through legal‑tech/data‑room channels (DFIN acquisition).
• Ocrolus: High‑volume financial document automation for lenders (bank statements, paystubs, tax forms) integrated into underwriting; overlaps on ingestion/table extraction but focuses on consumer/mortgage lending rather than institutional credit contracts (homepage, docs).
• Luminance: Legal AI and contract lifecycle platform for law firms and enterprises; competes on clause extraction and governance but is oriented to legal ops/CLM more than analyst‑ready finance datasets (overview, financial services use cases).