Aramco Ventures, the investment arm of Saudi Aramco, has co-led a $20 million seed funding round in US-based artificial intelligence startup Twin1 AI, which develops AI “twins” designed to replicate aspects of individual knowledge workers and handle routine office tasks.
The funding round was also co-led by Bessemer Venture Partners and Tribeca Venture Partners. The California-headquartered startup plans to use the investment to expand hiring in San Mateo and London, increase go-to-market activity and further develop its core technology.
Founded in 2025 by Lewis Z. Liu, Tom Cahn, Jonathan Budd and Huiting Liu, Twin1 AI develops digital twins using information from a user’s emails, meetings, documents and workplace systems.
The AI workers can operate within platforms including Slack, Microsoft Teams, Outlook, Gmail, Google Drive and SharePoint, where they can answer questions and carry out approved actions on behalf of users.
The founding team previously worked together at Eigen Technologies, and several investors in the earlier company have returned to back Twin1 AI.
For Aramco Ventures, enterprise governance and data control were key considerations behind the investment. Daniel Carter, Managing Director of Aramco Ventures US, said deploying AI within highly regulated enterprises requires more than standard security and compliance measures, adding that Twin1’s founders understand how large organisations operate and how AI can be deployed within them.
Twin1 AI has built an enterprise Model Context Protocol (MCP) server that allows other AI agents and internal tools to access a digital twin’s context. The platform also uses six layers of rules-based and AI-based permission controls governing the information each twin can access and share.
The company offers software-as-a-service, single-tenant and private-cloud deployment options. It describes its approach as “sovereign AI”, aimed at giving enterprises greater control over their data and reducing dependence on a single AI model or infrastructure provider.
Twin1 said it has worked with partners across the legal, financial services and energy sectors for more than a year. Its customers include Linklaters, Orrick, Dechert, Customers Bank and Aegis Energy. Orrick also participated strategically in the funding round.
The startup says customers have reported that its technology can automate between 30% and 50% of communications work performed by knowledge workers. However, the company has not provided a baseline, measurement period or methodology for the figure.
Chief Executive Lewis Liu said the technology is designed to amplify individual expertise rather than standardise it, arguing that AI should preserve an employee’s judgement and voice when carrying out knowledge-based work.
The investment highlights a growing focus among enterprises on the governance and permissions layer surrounding AI agents. As organisations deploy AI systems that can access corporate data and act on employees’ behalf, controlling what these systems can see and do is becoming a central consideration alongside their technical capabilities.
The issue is particularly relevant for highly regulated sectors such as banking, energy and government, where data privacy and security concerns can slow AI adoption. Technologies offering granular user permissions and controlled data hosting could give organisations greater visibility over how AI systems operate within regulatory frameworks.
For enterprises in markets such as Saudi Arabia and the wider region, where data protection and sovereignty requirements are increasingly important, the ability to demonstrate practical controls over AI access could become an important factor in technology procurement.