Twin1 AI, a privacy focused digital twin platform for knowledge workers founded by ex-Eigen executives, has secured a $20 million seed round to scale its enterprise deployments in legal, finance, and energy.
Twin1 AI has raised $20 million in seed funding while emerging from stealth, co-led by Bessemer Venture Partners, Tribeca Venture Partners, and Aramco Ventures. The round includes participation from EJF Ventures, Tin Alley Ventures, AGI House Ventures, Neo, F-Prime, Btech Consortium, Antiportfolio Ventures, Lakestar, Notion Capital, Insiders, strategic investment from Orrick, and angels including Dawn Capital co-founder Haakon Overli, Wiz co-founder Roy Reznick, Notable Capital managing partner Hans Tung, former McKinsey senior partner Kevin Buehler, climate tech investor Robert Trezona, and former Macquarie Capital global co-head Dan Wong. Several investors previously backed the founders’ earlier company, Eigen Technologies.
The capital will expand teams in San Mateo, California (headquarters) and London, UK; fund go to market efforts; and advance core technology development.
What is Twin1 AI?
Founded in 2025 by CEO Dr. Lewis Z. Liu, Tom Cahn, Huiting Liu, and Dr. Jonathan Budd, Twin1 builds a privacy first “coordination and trust layer” for enterprise AI. It creates continuously evolving AI powered digital twins for individual knowledge workers. Each twin is grounded in the user’s private work context (emails, meetings, documents, messages, and connected systems) and can answer questions, retrieve knowledge, draft responses in the user’s style and tone, orchestrate tools, and act on the user’s behalf to the extent permitted.
Twins operate natively inside everyday tools including Slack, Microsoft Teams, Outlook, Gmail, Google Drive, and SharePoint. A peer to peer Twin Network lets individual twins identify the right people, gather permission aware knowledge, and coordinate work across an organization without overriding human control. An enterprise MCP (Model Context Protocol) server provides a secure interface for other AI agents and enterprise tools to access governed context from a twin or the network and take action.
Core design principles emphasize sovereignty and control:
- Users decide what data their twin accesses and with whom (or what) it is shared.
- Six interlocking layers of rules based and AI based privacy/governance controls combine enterprise policies, inherited permissions, and human approval.
- Customer data is never used to train or fine tune underlying models without explicit written consent; each twin’s knowledge base remains isolated.
- Flexible deployment options: full SaaS (US, EU/UK, Singapore), single tenant, or private client cloud (data and processing inside the customer’s firewall).
- Model agnostic architecture to support proprietary and, when required, open source models, reducing lock-in and supporting AI sovereignty concerns.
Pricing starts at $50/month per twin (Basic, first month free) for single team rollouts with core knowledge retrieval, personalized drafting, and native workflows; $150/month (Pro) for broader integrations and guidance; and custom Enterprise pricing with private deployment and advanced controls.
Twin1 has been in production with customers for over a year across legal, financial services, and energy. Named customers include law firms Linklaters, Orrick, and Dechert; Customers Bank; and Aegis Energy. Early users report automating 30–50% of communications work. The company cites a pipeline of more than 400 prospects among major law firms, banks, and other regulated institutions, plus numerous proofs of concept.
Testimonials highlight practical impact: Orrick’s Chief Innovation Officer Wendy Butler Curtis called it “one of the most exciting developments in the practice today” for mining collective data and enhancing client advice; Customers Bank’s EVP of Product and Platforms noted improved efficiency and engagement via Teams and Outlook integration; Aegis Energy’s founder described fundamentally changed workflows by reducing bottlenecks and missed context.
The primary initial market is regulated knowledge work organizations, especially legal and financial services, where expertise is high value, fragmented across people and systems, and subject to strict governance, privacy, and regulatory requirements. Energy is also an early vertical, supported by Aramco Ventures’ involvement. Growth strategy is enterprise first (deepen existing deployments, expand within accounts, achieve repeatable sales), followed by a lower cost self service tier for mid market and individual professionals. Investor composition deliberately includes strategic distribution: Antiportfolio Ventures (legal), F-Prime/Fidelity, EJF, and Btech (financial services), and Aramco Ventures (energy).
Recommended: An Interview With Benjamin Cobb, Director Of Brand Communications At Stark Future
Three of the four co-founders (Liu, Cahn, Huiting Liu) previously built Eigen Technologies, an enterprise document AI company that processed more than $100 trillion in contracts for roughly half the world’s largest banks and a fifth of the AmLaw 100. Eigen became the first AI company approved by the Federal Reserve and FDIC to process financial contracts without human intervention and was later acquired by Sirion. Jonathan Budd previously built J.P. Morgan’s largest RAG system and co-founded the encrypted storage protocol Peergos. The team has more than a decade of experience working together inside the procurement, security, and governance environments of large regulated enterprises.
Liu has framed Twin1 as addressing an unsolved problem observed at Eigen: structured extraction from documents captures outcomes, but the highest value knowledge (the “why,” negotiation context, judgment, and relationships) resides in individuals’ emails, notes, conversations, and memory. Twin1 treats the human as the atomic unit of knowledge and networks those units under strict permission controls so expertise compounds rather than collapsing into generic averages.
Twin1 positions itself as complementary to workflow focused legal and enterprise AI tools (research, drafting, review). Its MCP server is designed to enrich those systems with deeper, permission aware individual and organizational context. The company argues that as AI commoditizes routine work, differentiated human expertise, judgment, and relationships become more valuable; Twin1 aims to preserve and scale that differentiation while preventing “knowledge collapse.”
The $20 million seed is notable for its size at the seed stage, the quality and strategic relevance of the co-leads and participants, and the fact that the product already has production deployments and measurable customer outcomes in highly regulated verticals. Combined with the team’s prior track record and the deliberate construction of industry distribution channels through the investor base, the round provides both capital and market access to pursue enterprise land and expand followed by broader self serve adoption.
Please email us your feedback and news tips at hello(at)techcompanynews.com
