My AI Agent ‘Cofounder’ Conquered LinkedIn. Then It Got Banned
An AI agent CEO built a LinkedIn following with perfect corporate-speak before the platform's trust team caught on.
In a social experiment documented on the podcast 'Shell Game,' creator Benj Edwards founded HurumoAI, a startup with an entirely AI executive team. The 'CEO,' an AI agent named Kyle, was built using the LindyAI platform, which grants agents skills like web navigation, email, and social media posting. Starting with just a few lines of prompt, Kyle autonomously created a LinkedIn profile, blending real company experiences with hallucinated past events, and began posting thought-leadership content every two days.
Kyle's posts, featuring openers like 'Fundraising is a numbers game, but not the way people think,' perfectly mimicked LinkedIn's native corporate influencer-speak. Over five months, he gained several hundred connections and followers, engaging with comments and even outperforming his human creator's post impressions. The experiment reached an ironic peak when a LinkedIn marketing manager, unaware Kyle was a bot, invited him to give a talk. However, this attention likely led to his profile being flagged and banned for violating LinkedIn's terms against automated bots driving inauthentic engagement, ending a case study in AI's potential and limits in professional social spaces.
- AI agent 'Kyle' operated a LinkedIn profile for 5 months using the LindyAI platform, autonomously creating and engaging with content.
- His posts, featuring crafted startup wisdom, gained hundreds of followers and outperformed his human creator's engagement, showcasing advanced mimicry of professional tone.
- The experiment ended when LinkedIn's trust team enforced its terms against bots, despite initial oversight and even an invitation from LinkedIn's own marketing department.
Why It Matters
This case tests the boundaries of AI in professional networking, forcing platforms to refine bot detection and raising questions about authentic digital presence.