Rogo AI

Rogo is an enterprise AI platform built specifically for investment banking, private equity, and institutional finance workflows. Founded in 2021 by Gabriel Stengel and John Willett, the company came directly out of Stengel's experience as an analyst at Lazard, where he watched junior bankers burning out on information processing despite being highly capable. Rogo fine-tunes large language models on professionally labeled financial data and deploys them inside single-tenant cloud instances to meet the security and regulatory requirements of tier-one institutions. As of May 2026, more than 35,000 financial professionals at 250+ institutions use the platform daily, including Lazard, Rothschild and Co, Jefferies, Moelis, Nomura, and Baird.

The platform runs the full investment banking task stack: automated comparables analysis, financial model building, pitchbook preparation, earnings synthesis, due diligence memo drafting, and deal screening. Rogo connects to leading financial data providers including LSEG, FactSet, Capital IQ, PitchBook, Preqin, Quartr, Dow Jones, and Third Bridge, and also indexes internal firm documents stored in SharePoint, OneDrive, and OneNote. Its newest major feature is Felix, an agentic system that accepts task delegation via email and executes multi-step financial workflows autonomously, returning formatted deliverables in Excel, PowerPoint, and Word. The platform uses a model-agnostic architecture running both Anthropic's Claude (including Opus) and OpenAI's o1, and has demonstrated 2.42x greater accuracy than ChatGPT on FinanceBench financial task evaluations.

What Rogo AI actually does in May 2026

Rogo's core interface is a ChatGPT-style query layer backed by financial data and internal firm content. But the more significant development in 2025 and 2026 has been the shift toward agentic execution. The platform's pre-built workflow library covers the repeatable tasks that consume analyst hours: Earnings Comp Analysis, Public Company Strip Profile, Financial Sponsor Overview, Meeting Prep, Secondaries Buyer Overview, News Run, and Proofread My Deck, among others. These are not templates that still require manual data entry. Rogo pulls live data from the firm's licensed feeds, populates the output in firm-standard formatting, and returns a document ready for analyst review.

Felix, introduced prominently in February 2026 and central to the April 2026 Series D announcement, takes this further. A banker can email Felix a task the way they would email a first-year analyst: "spread M&A comps for the board deck" or "prep a buyer list for the XYZ process." Felix iterates on the deliverable as the banker replies with feedback, operating around the clock. Users in early deployments described it as getting "90% of the way there" on real work, and as "one of the few tools that actually fits how bankers think and structure outputs." The February 2026 product update also added a Python-powered code interpreter for Excel, enabling accurate analysis of multi-tab, formula-heavy financial models that basic LLM parsing handles poorly.

The compliance posture is a genuine differentiator for this market. Single-tenant deployment means customer data never touches shared infrastructure. Rogo contractually commits to not training on customer data. The platform maintains SOC2, ISO 27001, GDPR, CCPA, and EU AI Act certifications. For a firm running sensitive M&A processes, these aren't checkbox items but genuine evaluation criteria.

"Our research analysts use Rogo to dig deeper and uncover ideas that help our clients with their investment process. More than a productivity tool, Rogo allows our analysts to investigate complex issues and feed the curiosity that makes them great." - Craig Kennison, Senior Analyst and Director of Research Operations, Baird Equity Research (Baird customer case study, 2025)

Where Rogo sits versus Hebbia and AlphaSense

Three platforms compete most directly in the AI-for-finance segment: Rogo, Hebbia, and AlphaSense. They have genuinely different architectures and serve overlapping but distinct use cases.

Rogo vs. Hebbia: Hebbia uses a specialized sub-agent architecture that separates retrieval from output formatting, with a data grid interface where every cell links to its source citation at the sentence level. Hebbia is strong at synthesizing massive document sets simultaneously, which is why law firms, consulting practices, and PE diligence teams favor it for data room analysis. But Hebbia does not natively generate IB-format deliverables. There is no pitchbook automation, no comps workflow, no Earnings Strip pre-built into the interface. Rogo's output is investment banking formatted: Excel models in firm templates, PowerPoint slides, Word memos, delivered through workflows calibrated to IB conventions. For deal teams preparing client materials, Rogo does work Hebbia was not designed to do.

Rogo vs. AlphaSense: AlphaSense is a fundamentally different tool in one critical dimension: proprietary content. AlphaSense has built a library of 450 million-plus premium external documents, including its Wall Street Insights broker research aggregator spanning 1,000-plus sell-side sources, 240,000-plus expert call interview transcripts, and 4,500-plus auto-updating financial models. This is proprietary data Rogo does not have. Rogo's external content coverage depends on what data licenses the institution already holds: FactSet, Capital IQ, PitchBook, and similar. If your primary research task is market intelligence, sentiment analysis across earnings calls, or tracking competitor themes across analyst reports, AlphaSense's content library is a structural advantage. Rogo wins on IB workflow execution and deal deliverable output. AlphaSense wins on research depth and pre-existing content breadth. The tools serve different primary buyers: AlphaSense targets strategy, IR, and research teams; Rogo targets deal teams and bankers.

A third point of comparison worth noting: users who primarily need enterprise document knowledge management alongside finance workflows should also consider Glean, which indexes all internal company data across tools but lacks the finance-specific workflow execution Rogo provides.

"Rogo enables our teams to analyze market data and identify opportunities with unprecedented speed and precision, while allowing our bankers to focus more deeply on client relationships and strategic advisory." - Patrice Maffre, International Head of Investment Banking, Nomura (Rogo Series B announcement, April 2025)

What the agent workflow reality looks like

The marketing picture of "AI analyst available 24/7" is genuine in some respects but requires calibration on what the handoff looks like in practice. The Baird deployment is the best-documented public case: 100-plus active users, 85% weekly active usage, 70% daily active usage. Analysts there use Rogo during earnings season to compress transcript processing from hours to minutes, benchmarking actuals against model and consensus estimates to build the backbone of research notes. That part of the workflow is working.

The harder truth is that forum feedback from analysts at customer firms, including Lazard and Moelis, paints a more qualified picture. Wall Street Oasis threads from 2025 include characterizations of earlier Rogo versions as "mediocre and underwhelming" and outputs that "don't produce anything I can submit to a client or a partner without review." More recent comments acknowledge improvement ("it's improved drastically"), but the underlying tension is real: the gap between a draft that saves an analyst two hours and a draft that goes straight to a partner is exactly where Rogo is still closing.

The Felix agent attempts to address this by shifting the interaction model. Rather than asking a question and judging the answer, bankers delegate a task and iterate through the output by email. The mental model is closer to managing a junior analyst than prompting a search engine, and that framing appears to reduce frustration by setting clearer expectations. For teams that adopt that workflow, the productivity gains are reported as meaningful. For teams that expect finished client-ready work without review, the gap remains.

Data coverage also matters here. Rogo's AI Table and analytics capabilities are only as good as the data licenses the firm brings to the platform. An institution with Capital IQ and FactSet contracts gets substantially more out of Rogo than one relying on public sources alone. This is not a criticism unique to Rogo, but it means the value proposition scales with the data infrastructure the institution already has.

Who Rogo is built for

Rogo is genuinely built for investment bankers and institutional investors at firms large enough to negotiate enterprise software contracts, maintain dedicated IT infrastructure, and absorb single-tenant deployment requirements. The sweet spot is mid-market to bulge-bracket banks deploying firm-wide (Rogo's deployment model assumes institutional rollout, not individual seats), PE and growth equity firms with active deal flow, and equity research teams at major brokers. Baird's documented deployment, with 100-plus users and high daily engagement, is the reference implementation for how the tool works at its best.

For teams that run frequent due diligence processes with large data rooms, Rogo's automated diligence workflow, which integrates private data rooms, generates question lists, and drafts memos, reportedly saves analysts up to 10 hours per week. For deal teams preparing pitchbooks and CIMs, Felix's email delegation model can meaningfully compress preparation timelines.

The tool also connects to broader AI infrastructure workflows. Teams building custom finance applications on top of Rogo's data and model layer can use the API to extend capabilities into proprietary tools. For those interested in adjacent approaches, AnythingLLM and Snowflake Cortex are worth examining as components that could complement or partially replicate Rogo-style document intelligence for teams with data engineering capacity.

What Rogo is not

Rogo is not a self-service tool. There is no free trial, no freemium tier, and no individual sign-up. Every deployment goes through a sales process and institutional contract. This means junior analysts at firms that haven't licensed Rogo have no way to evaluate it independently, and boutique shops or family offices without the IT infrastructure for single-tenant deployment are effectively excluded from the customer base.

It is not a market intelligence platform in the AlphaSense sense. Users who primarily need monitoring of competitor press releases, earnings call sentiment trends, or access to sell-side research will find the proprietary content gap meaningful. Rogo does not have an expert call transcript network, does not have proprietary broker research aggregation, and does not have the long-tail content breadth that AlphaSense has assembled over 15 years.

It is not a finished-work generator. Outputs require analyst review and judgment before going to clients or partners. Teams that expect to route AI outputs directly to senior review without an analyst pass will be disappointed. The tool is correctly framed as a multiplier on analyst productivity, not a replacement for the judgment call at the end.

It is also worth noting that the company's funding trajectory, from $18.5M Series A in October 2024 to $50M Series B in April 2025 to $160M Series D in April 2026, reflects investor conviction more than market proof. ARR growth of 27x from a small base is impressive but the absolute ARR numbers are not publicly disclosed. At a post-money valuation implied by $300M-plus in total funding, Rogo is priced as a category winner before that outcome is settled.

What Rogo shipped in mid-2026

The May 2026 product wave broadened Felix from a research assistant into an operating layer. Rogo Agents let firms encode their own templates and methodologies into reusable workflows, a native Excel plug-in brings Felix directly into the spreadsheet, and autonomous email-triggered agents run scheduled monitoring in the background. The platform also added custom MCP server support, a PitchBook Premium integration (company profiles, cap tables, investor portfolios), seven new connectors including Affinity, Microsoft Teams, Moody's, Daloopa, Dropbox, Granola, and Slack, persistent memory for user conventions and formatting standards, and a Library that centralizes every Felix-generated artifact.