RavenPack / Bigdata.com: Initiation of Coverage

GAII Intelligence defines the AI Research & Intelligence market as software that applies generative and agentic AI to the research process of professional investors: locating candidate securities, assembling and interrogating evidence, producing written analysis with traceable sources, and monitoring positions once held. Buyers are hedge funds, long-only managers, multi-manager platforms, family offices and allocators; the economic buyer is typically the chief technology officer, the head of research, or a portfolio manager with a technology budget.

Capabilities we treat as mandatory

• Evidence retrieval at scale: the ability to search and reason across regulatory filings, transcripts and news, and across documents the client supplies.
• Source traceability: every material claim in an output must be reachable back to the passage that supports it.
• Institutional data handling: a stated position on where client data goes, how long it is held, and whether it is used for model training.

Capabilities we treat as differentiating

• Persistent state: the system remembers a thesis, a coverage list or a position between sessions and acts on it without being re-prompted.
• Workflow surfaces: delivery inside the applications analysts already use — the spreadsheet, the research management system, email and messaging.
• Owned or exclusive content: material that a competitor cannot obtain by signing the same third-party contract.
• Published evaluation: a measured, disclosed accuracy methodology rather than an assertion of quality.

Bigdata.com competes with Fiscal.ai and Daloopa as a data layer for AI platforms, and with AlphaSense, Bloomberg and LSEG on content. Its model-neutral positioning means it is often complementary to research platforms such as Hebbia and Rogo rather than a replacement.

The following are the capabilities and characteristics that, in our assessment, most distinguish Bigdata.com within the AI Research & Intelligence market. Each is drawn from publicly documented product capability, published interviews with management, or independent hands-on review.

• Entity knowledge graph: More than 12 million entities resolved to tickers and identifiers, with point-in-time history built over two decades.
• Licensed premium content: Financial Times (licensed for generative AI), S&P Global and FactSet transcripts, Quartr, Washington Post and Caixin, among 300+ sources.
• Assistant distribution: An official Claude connector, ChatGPT and Microsoft Copilot access, and a remote MCP server used from Cursor and Gemini.
• Shipping cadence: About 50 MCP releases in 17 months, including Deep Research, private content and watchlist tools.
• Transparent pricing: Published tiers from free to $160 a month, per-run cost estimates and credit for licences a client already holds.
• Tenure and backers: A 23-year operating history, tier-one bank and asset-manager clients, and a strategic investment from FT Ventures.

The following are factors that prospective clients should weigh, and areas where our public review did not find sufficient evidence to reach a settled view. Several are common across vendors at this stage of the market and are not specific to Bigdata.com. We encourage prospective clients to raise each of them directly with the company; we have done the same, and will update this note as answers arrive.

• Security evidence: The homepage cites both SOC 2 Type II and a "Type III" that does not exist; the trust centre could not be read. Buyers should request the report.
• Training terms: The homepage pledges no training on customer data, while the terms allow searches and responses to be used to improve the service. Buyers should seek a contractual exclusion.
• Retention: Zero data retention is required of the customer's model, not offered by the vendor on prompts.
• Access controls: SSO is Enterprise-only; SCIM, admin audit logs and MNPI controls were not located.
• Content scope: One expert network and independent research only; no bulge-bracket broker research.
• Related-party evidence: The only quantified case study comes from Bullhound Capital, affiliated with an investor; the CFO is a former Bullhound executive.
• Inconsistent figures: Archive depth (20 or 25 years), source counts and token savings vary across company pages.

RavenPack was founded in 2003 by Armando Gonzalez and contracts through RavenPack International S.L.U. in Marbella, Spain, with a New York office and about 200 staff. Its News Analytics datasets have served quantitative funds and banks for two decades. Bigdata.com, launched publicly in October 2024, is its generative-AI platform.

RavenPack raised $5m from Draper Esprit in 2017, $10m from GP Bullhound in 2019 and $20m led by GP Bullhound's DoubleDown fund with the EIB Group in July 2024 to build Bigdata.com; FT Ventures invested an undisclosed amount alongside a content licence in December 2025. About $35m is disclosed; valuation and revenue are not.

Company-reported figures are shown as such and are not independently verified. Verification is carried into the vendor questionnaire.

Bigdata.com is offered as an app and as a content layer for other assistants and agents. Figure 2 sets out our assessment of how it maps onto the investment workflow.

Bigdata.com app

Chat, Deep Research, watchlists, briefs, workflows, document Q&A and tearsheets on web and mobile.

Remote MCP server

Search, tearsheets, screener, events calendar, watchlist and private-content tools, with a Deep Research MCP in beta.

Assistant connectors

Official Claude connector (April 2026), ChatGPT and Microsoft Copilot access, and open-source skills.

API and datasets

Search and research-agent APIs, a Python SDK and RavenPack's legacy News Analytics datasets.

Scope of the product today

We did not locate an Excel or Office add-in, native deck building or a VPC deployment option.

Bigdata.com runs retrieval on Vespa, combining approximate-nearest-neighbour vector search with machine-learned ranking over a knowledge graph of more than 12 million entities. Its own models moved to Amazon Bedrock; the client's assistant does the final reasoning.

Evaluation evidence is limited to one vendor-run test of 32 reports, in which grounding scored 82% with Bigdata against 61% with web search.

Content spans more than 300 premium sources in 35+ languages: SEC filings and exchange feeds, S&P Global, FactSet, Quartr and SCRIPTS Asia transcripts, the Financial Times since 2012, MT Newswires, Benzinga and Caixin, FactSet estimates, Preqin and alternative data.

Expert content comes from one network and research from independent providers. Clients can upload documents and connect their own sources.

Bigdata.com fits monitoring and research inside an existing assistant: watchlists and briefs that report what changed, and grounded answers in Claude, ChatGPT or Copilot.

It does not build models or decks; those depend on the host assistant.

RavenPack reports ISO 27001 certification in 2025 and SOC 2 Type II, offers SSO and data-residency controls on its Enterprise tier and runs on AWS. Figure 3 sets out what our public review located.

We would encourage buyers to request the SOC 2 report, a contractual no-training clause, SCIM and audit-log capability, and named model providers.

The offsetting factor is tenure: RavenPack has supplied banks and quant funds for two decades, implying repeated vendor diligence.

Pricing is published: Starter free with a $10 credit, Pro at $40 a month, Max at $160 a month and Enterprise by quote, with content charged per token at roughly $0.57–2.69 per workflow.

Existing licences such as FactSet or FT Professional are honoured, so content is not paid for twice.

Bigdata.com sits in the Emerging quadrant of the GAII Intelligence Matrix, held back by enterprise documentation and content gaps, close to the Established band on overall score.

GAII initiation universe, 20 October 2026. Scores as currently published on the GAII Intelligence page; preliminary and based on public information.

11.1 Where Bigdata.com is strongest relative to peers

• Against content incumbents, on per-token access through the assistants analysts already use.
• Against AI-native data layers, on entity resolution and licensed news depth.
• Against research platforms, on model neutrality and price.

11.2 Areas of narrower functional focus

Bigdata.com has built depth in content infrastructure rather than end-user deliverables or enterprise paperwork.

• Enterprise: SOC 2 claim inconsistent; training terms permissive.
• Content: thin expert and broker research.
• Deliverables: no models or decks.
• Evidence: vendor-run evaluation only.

Scores are on a 0–10 scale. Each dimension score is the mean of its sub-criteria, and the GAII Overall Score is the weighted mean of the six dimensions. The detailed sub-criteria scores that build each dimension are set out in Appendix A.

These are the areas where our public review did not find sufficient evidence to form a settled view. We have put them to the company and will update this note when they are answered. Prospective clients may find them a useful agenda for their own evaluation call.

• The SOC 2 report type, period and auditor.
• A contractual no-training clause covering searches and responses.
• Zero-data-retention terms for prompts on the vendor side.
• SCIM and admin audit-log availability.
• Information-barrier and MNPI controls.
• Model providers used for internal processing.
• Bigdata.com customer count and revenue.
• Licence scope for each premium source.
• Plans for broker and expert content.
• VPC or private deployment options.
• Basis for archive-depth and source-count figures.
• Independent evaluation of grounding accuracy.

This note is an assessment of a vendor against a defined framework, not a recommendation to buy or to decline. The following are the considerations we would put at the centre of an evaluation of Bigdata.com specifically.

If you are a fundamental equity manager

• Connect it to your current assistant and test grounding on live names.
• Use watchlists and briefs for monitoring.
• Check which of your licences it honours.

If you are responsible for technology or vendor risk

• Open with Figure 3; request the SOC 2 report.
• Negotiate the no-training clause.
• Confirm SSO and audit logging before rollout.

If you are assessing the category rather than the company

• Content layers are being sold into assistants rather than terminals.
• Per-token content pricing is spreading.
• Licensed news for generative AI is becoming scarce and valuable.

Each dimension score in Section 12 is the mean of the sub-criteria below.