Rogo

Rogo leads on scale, capital and the finished deliverable. Rogo has raised over $300m in two years, reports more than 50,000 users across 350+ institutions, and added five global banks to its cap table in September 2026. The product is built around what a banker actually delivers — models, comps, pitchbooks and memos in the firm's own templates — and Felix extends that to autonomous deal work. Since the start of September Rogo has completed the Arvo acquisition, opened Datasite data rooms under existing permissions and added an Octus credit-research partnership.

Enterprise posture is extensive as described; the public evidence is thinner. Rogo describes dedicated-cluster isolation, single-tenant and on-premises options, private model copies and permissions inherited from data rooms and Snowflake. The SOC 2 and ISO attestations were located only through third-party indexes of the trust centre, and we did not locate information-barrier controls, single sign-on or audit-log export documentation. Expert-network and aggregated broker research were not located.

We lower the rating to Established, at 7.24, with an Improving outlook. The change from 8.2 reflects recalibration to the full sub-criteria standard applied to every company covered since October rather than a change in Rogo's business: claims that we could not corroborate no longer lift scores, absent content and undocumented controls are scored as such, and BigFinanceBench is treated as a benchmark of models rather than of the Rogo product. The largest moves are in Enterprise Readiness (9.00 to 7.21), Data & Content (7.60 to 6.33) and Commercial Value (7.80 to 6.00). Momentum remains the highest in our coverage, and Rogo now sits in the Innovators quadrant; first-hand attestation evidence would move the score most.

Strengths

• Capital and vendor viability. About $310–345m raised from Khosla, Thrive, Sequoia and Kleiner Perkins, with a reported ~$2bn valuation in April 2026 and a September 2026 strategic round from global banks.
• Deliverable fidelity. Comps, DCF, LBO and merger models, pitchbooks, CIMs and memos are produced in Excel, PowerPoint and Word in the firm's own template and house style, with cell-level source links.
• Autonomous agents. Felix takes tasks by email and works continuously on deal screening, CIM generation, buyer outreach and data-room diligence.
• Deployment and isolation. Dedicated Kubernetes clusters and data stores, single-tenant, multi-cloud and on-premises options are described, with private model copies and permissions inherited from Datasite and Snowflake.
• Model-agnostic routing. Workflows route across OpenAI, Anthropic and Google models, and new frontier models are added within days.
• Named client base. Lazard, Rothschild & Co, Jefferies, Moelis, Nomura, Baird and Truist are named by the company or its investors, with more than 50,000 users reported.
• Adoption model. Forward-deployed teams of former bankers drive change management, supported by offices in New York, London, Singapore and Sydney.

Cautions

• Attestation evidence. SOC 2 Type I/II and ISO 27001/42001 are listed by third-party indexes of the trust centre; we could not read the trust centre directly.
• Model-provider terms. No-training and private-copy commitments are documented for the Amazon Bedrock path; zero-data-retention terms for the OpenAI and Google paths were not located.
• Information barriers. MNPI controls, information-barrier configuration and audit-log export were not documented publicly.
• Buy-side workflow fit. The product is built around investment-banking deliverables.
• Benchmark scope. BigFinanceBench is Rogo-authored and evaluates foundation models rather than the Rogo product, including its October 2026 extension to financial data; a third-party data-quality allegation was unresolved at our review.
• Leadership and integration. A new chief technology officer was appointed in July 2026, the co-founder CTO's current role was not located, senior AI staff moved to OpenAI, and four acquisitions — Subset, Offset, Rivanna and Arvo — are being integrated during international expansion.
• Pricing and metrics. Pricing is not published, and user and revenue figures are company-reported with shifting definitions; no audited ARR has been disclosed.
• Frontier-lab competition. OpenAI launched ChatGPT for Financial Services in September 2026 with investment-banking design partners, and Anthropic and Microsoft are moving into the same workflows.