Instabase

Instabase is an AI-native applied-AI platform for enterprise financial services, founded in 2015 at MIT by Anant Bhardwaj and headquartered in San Francisco with offices in New York, London and Bangalore. Where Rossum was architected as a purpose-built transformer-first IDP engine for transactional documents, and Hyperscience sits as the "modern" CNN-plus-template rung with enterprise analyst depth, Instabase positions further up the stack — as an app-building platform for unstructured-data workflows, with a content-understanding core (since 2023 rebranded and extended as "AI Hub") that layers LLMs on top of its own document-AI models.

What Instabase actually is. Three things bundled into one platform. First, a document-AI engine that classifies, splits and extracts from unstructured documents — policy submissions, claims files, lending packages, KYC and onboarding bundles. Second, an "AI Hub" launched in 2023 that exposes LLM-powered content understanding as a gallery of reusable apps and turns documents into natural-language-queryable objects (partnered with OpenAI GPT models). Third, a low-code application layer — enterprises can assemble extraction, validation, orchestration and human-in-the-loop review into named business applications, then deploy them on Instabase's runtime. This platform-plus-apps model is why PitchBook described the 2023 round as funding an "app-building tool suite" rather than a pure IDP vendor.

Funding and scale. Cumulative disclosed funding is approximately $277M across four priced rounds: a $3.75M seed from Greylock and NEA (August 2015); a $23.2M Series A led by Andreessen Horowitz (June 2017); a $105M Series B led by Index Ventures with Spark Capital, Tribe Capital, SC Ventures (Standard Chartered's venture arm) and Glynn Capital (October 2019); a $45M Series C led by Tribe Capital at a $2B valuation (June 2023); and a $100M Series D led by the Qatar Investment Authority (QIA) with Greylock, NEA, a16z and Index Ventures all back on the cap table (January 2025). That $277M+ capital base materially exceeds Rossum's ~$104M but sits below Hyperscience's ~$289M. Reported 2024 revenue was $40.8M on a 265-person team per Latka, though this figure is unaudited and should be treated as indicative rather than authoritative.

Analyst position. Weaker than the two direct comparables. Instabase is a Representative Vendor in the 2024 Gartner Market Guide for IDP Solutions but — based on public coverage as of April 2026 — was not positioned in the 2025 Gartner Magic Quadrant for IDP Solutions, where ABBYY, UiPath, Tungsten and Hyperscience are Leaders and Rossum is a Challenger. The Magic Quadrant absence matters: for an enterprise buyer running a formal RFP, "Representative Vendor in a Market Guide" is a lower tier of validation than "positioned in a Magic Quadrant," regardless of quadrant. Instabase has been covered by TechCrunch, Forbes, PitchBook and CNBC around its funding rounds, which supplies press legitimacy but not category leadership.

Insurance footprint — handle with care. The one public insurance carrier case study is AXA UK commercial-lines underwriting, where Instabase automates extraction from broker submissions (emails, slide decks, Excel, Word) to free up underwriter capacity. InsTech London's vendor profile references Instabase being "in discussion with most of the largest brokers and insurers" and the vendor's own insurance page references life-insurance underwriting and claims use cases, but no second named carrier, no reinsurer, and no named broker case study surfaced in April 2026 public search. A Phidea reader should treat Instabase's insurance narrative as genuine but thin in public references — ask for direct carrier calls, not the case-study library.

Where it sits in old to new to AI. Legacy carriers and broker back offices run ABBYY FlexiCapture, Kofax (now Tungsten), or offshore data entry. Hyperscience is the "modern" rung. Rossum is the AI-native rung built as a purpose-built IDP. Instabase is the "platform" variant of the AI-native rung — broader horizontal scope (invoice, KYC, lending, claims, submissions, content understanding), deeper financial-services pedigree outside insurance, and a Series D war chest from a sovereign lead. The trade-off is real: if the buyer wants a configured IDP that ships opinionated ACORD and MRC accelerators, Rossum or Hyperscience are closer to the use case. If the buyer wants a general-purpose applied-AI platform on which insurance, lending and onboarding apps live side by side — and has the implementation capacity to build those apps — Instabase is the more strategic bet.