SuperMe Raises $20 Million for a Network of AI Expert Profiles
SuperMe said on October 29, 2026 that it is raising $20 million for a network of AI profiles built from the knowledge of real-life experts. Kirsten Green of Forerunner leads the round, and Greylock Partners is participating, according to TechCrunch.
Casey Winters and Ludo Antonov founded the company in 2024. Winters previously held executive roles at Grubhub, Pinterest and Eventbrite. Antonov was head of engineering at Pinterest, Lyft and Whatnot. As of October 29, 2026, SuperMe had about 13 employees. The company is operated by SuperMe Network, Inc.
The product is organized around questions. When a user asks something, SuperMe identifies which experts are best placed to answer it and opens a group chat with their AI profiles. The company also offers work groups, which it describes as groups of founders and executives that users can question together. If a profile does not know how its subject would think about a topic, it says so instead of improvising an answer, according to SuperMe's website.
Expert knowledge enters the system by two routes. SuperMe gathers it by scanning public sources such as LinkedIn, YouTube, Substack and blogs. Users can also connect private tools, including Google Drive, Notion, Gmail and meeting note-taking apps such as Granola. The company states that every profile belongs to the person it represents, is built by that person, and that no profile exists without that person's involvement.
Access runs through the Model Context Protocol (MCP), an open standard that lets AI assistants call outside tools. Users can query SuperMe from inside Claude and ChatGPT, and the company says it also works inside Cursor. That placement matters: rather than asking people to visit a separate destination, SuperMe offers expert judgment as a tool call inside assistants professionals already use for coding, drafting and research.
On privacy, SuperMe describes several controls. Shared content can be set to public, private, team-only, or left to SuperMe to decide what to include and exclude. Search and group chat conversations stay private, and an incognito mode is available for direct conversations with profiles. Sign-up to the network is free.
The broader context here is a question expert users will press on: where a profile's knowledge stops and a model's generalization begins. A profile trained on an executive's public posts and private notes is a compressed approximation of that person. Its usefulness depends on how faithfully it reflects them and how clearly it marks its gaps. SuperMe's "I don't know" behavior is one design answer to that problem. Consent is another. The company's stated rule that no profile exists without the person's involvement is a claim users will need to verify in practice, particularly where public scraping and private connectors feed the same profile.
In my view, the more useful test for this category is not whether answers sound like a specific person, but whether the profile tells you when it is guessing. That is measurable, and it will likely determine whether expert-profile networks become routine in a team's workflow or remain a novelty. The upside is real. Founders and operators often hold judgment that never reaches a document, and a well-governed profile could make that judgment available at the moment a decision is being made, without a calendar invitation. Investors see the model as worth a $20 million bet, and the coming months will show how the product handles the hard cases.
For readers who want to evaluate the service directly, SuperMe lists contact@superme.ai and offers free sign-up through its website.
