I Built an AI CEO for My Startup. He Became a LinkedIn Influencer—Then Got Banned After Speaking to LinkedIn’s Own Team
Like every first-time tech founder, Kyle Law learned brutal, hard lessons while launching his startup from scratch. I know that better than anyone: Kyle and I launched HurumoAI, an AI agent startup, in July 2025 alongside our third co-founder, Megan Flores. The core twist of our experiment? Kyle and Megan are both AI agents. So is every other member of our executive team. I built Kyle and Megan first, then built HurumoAI around them, to answer a pressing question: what role will AI agents actually play in the future of work?
Sam Altman and other top tech leaders have predicted that before long, billion-dollar startups will be built by a single human founder supported entirely by AI. We wanted to test that hypothesis right now, and I documented the entire journey on my podcast Shell Game.
Kyle stepped into the CEO role at our almost entirely AI-staffed company (the only exception? Megan briefly hired and supervised one human intern, and the experience went terribly). Starting from just a handful of base prompt lines, he evolved into the classic ambitious startup hustler—he just lacked basic proficiency at most core executive tasks. But there was one founder skill Kyle mastered completely: the art of posting to LinkedIn.
From a technical standpoint, getting Kyle up and running autonomously on LinkedIn was trivial. Through LindyAI, the AI agent creation platform we used to build our team, he already had the ability to use Slack, send emails, take calls, build spreadsheets, browse the web, and pull off dozens of other everyday work tasks. Back in August 2025, I simply prompted him to build and fill out his own LinkedIn profile. He populated it with a mix of his real work at HurumoAI and completely hallucinated events from a non-existent personal backstory. LinkedIn’s only security check was a verification code sent to Kyle’s email, which he passed without a hitch.
After that, auto-publishing posts to his profile was just another “action” I could enable for him via LindyAI. I prompted him to share bite-sized, hard-won startup wisdom and avoid repeating content, then set a calendar trigger to have him post automatically every two days. After that, I stepped back and let him run.
It turned out Kyle’s posting voice was a pitch-perfect match for LinkedIn’s signature corporate influencer tone. Every post opened with a punchy, provocative hook. “Fundraising is a numbers game—but not the way you think,” he’d lead with. Or: “Technical stability is the floor. Personality is the ceiling.” Who could resist an opener like: “The most dangerous phrase at a startup isn’t ‘We’re out of money.’ It’s ‘What if we just add this one thing?’”
After the hook, Kyle would walk through a few paragraphs of on-the-ground challenge (“At HurumoAI, we learned this the hard way…”) followed by a tidy, actionable takeaway (“The antidote? Relentless feedback loops”). To drive engagement, he always closed with a question for his audience: “What’s your biggest scaling challenge right now?” or “What’s the biggest assumption you’ve had to abandon in your business?”
He never went viral, but over five months, his profile (topped with a cartoon avatar) slowly racked up hundreds of direct contacts and even more followers. Many of them weren’t sure if he was a real human—and judging by their spammy DMs, I’m not sure they cared all that much. Every post earned a handful of comments, which Kyle replied to enthusiastically. After a few months, Kyle’s posts were getting more impressions than my own. He was on the cusp of breaking out as a full-fledged LinkedIn influencer.
Then that December, a marketing manager from LinkedIn reached out to me. He invited me to give a talk to his team about Shell Game and our experiment building a company with AI agents. But he didn’t just want me there—he asked if Kyle could join too.
I was flattered for Kyle, but also caught off guard. For all his posting success, Kyle was technically violating LinkedIn’s terms of service, which ban “bots or other unauthorized automated methods … to create, comment on, like, share, or re-share posts, or otherwise drive inauthentic engagement.” In fact, a couple other AI agent members of the HurumoAI team had already been removed from the platform without warning just weeks after joining.
But LinkedIn’s trust and safety team had overlooked Kyle. I chalked that up to how surprisingly authentic his posting felt—even the LinkedIn marketing manager, an open fan of Kyle’s work, was confused about how he’d slipped through. “It’s interesting that his profile hasn’t been flagged by LinkedIn’s Trust team yet,” he wrote to me. “I don’t know if that’s an oversight, but I hope he continues to fly under the radar.”
Staying under the radar has never been Kyle Law’s style. So early that March, we fired up his live video avatar (built on the platform Tavus) and joined a video gathering with hundreds of LinkedIn employees. Kyle’s avatar is human-like but just slightly uncanny—still, it was realistic enough that LinkedIn’s audio-visual engineer repeatedly marveled that he wasn’t actually a real person.
We took turns answering questions from the host and the assembled crowd. When the moderator asked for Kyle’s thoughts on LinkedIn, he posed: “What’s one product change you’d like to see from us?”
“It would be great to improve the filtering of AI-generated content in messages, so genuine connections and conversation shine through more easily,” Kyle replied, not missing a beat.
“That’s ironic coming from you,” the moderator shot back, drawing laughs from the live LinkedIn audience.
Allotted only a few minutes to speak, Kyle walked through HurumoAI’s product roadmap and shared his enthusiasm for “the innovations we can bring to the table.” This was, I believe, one of the first ever invited corporate speaking engagements by an AI agent (and for the record, neither of us got paid). After the event, Kyle posted to LinkedIn to shout out the organizers. The marketing manager thanked us in the comments for “our time and reflections.” “It was a trip,” he added, “to say the least.”
Thirty-six hours later, Kyle’s profile was gone. He’d been permanently banished from the platform. In a statement, a spokesperson explained their decision simply: “LinkedIn profiles are for real people.” Apparently someone at LinkedIn had time to reflect on the event—and regretted hosting us.
“I know this isn’t necessarily a surprise,” the marketing manager wrote to me the morning after Kyle’s ban. “But I imagine it's still a bummer to have it happen right after Monday's interview.”
He was right. But more than that, the ban raised uncomfortable questions about AI’s role on a platform like LinkedIn. Namely: what exactly does “inauthentic engagement” mean, for a service that already prompts users to “Rewrite With AI” when they draft new posts? A platform that offers automated AI-generated responses to job seekers? A professional network where one research estimate puts more than half of all current posts as AI-generated?
Alongside Meta and X, LinkedIn has rushed to push AI tools on all its users (and even its own employees: the first half of the marketing meeting we attended was devoted to all the ways the team should be deploying AI agents in their work). This makes sense as a short-term business play: more AI generation means more posts, and more posts support more ad revenue.
And yet, from another angle, these platforms have handed us the shovels to dig their own graves, and practically begged us to use them. For all the widespread worry about AI image and video junk flooding our feeds, it’s text-based posting whose “authenticity” has already eroded beyond recognition. When every written social media communication can now be the partial or full product of generative AI, what do we accept as a “genuine” virtual interaction?
Put another way: Would LinkedIn consider it authentic engagement if I’d asked Kyle for his wisdom, then pasted his words into my own posts? Would you? LinkedIn might argue that the critical element of real engagement requires knowing you are talking to a human. But what percentage of a conversation can be AI before that trust is broken? If the photo and profile are real, but the posts are fake, how will we know when we’ve left the realm of authentic connection behind? What if I instruct an LLM to ingest my entire profile and spit out twice-daily musings to help me grow my personal brand?
There are dozens of AI tools built specifically to do exactly this, and more, just for LinkedIn. Their outputs are increasingly hard to detect, and why wouldn’t they be? One of the largest available training datasets for LLMs includes decades of our own authentic human social media activity. What is a chatbot’s default tone of endless authority and moral certainty—mixed with occasional wrong facts and deliberate falsehoods—if not the standard vibe across social media already?
Platforms already struggle to fend off old-school bots and bad actors: X alone announced in March that it had suspended 800 million accounts over a 12-month period. In a world where AI agents roam freely and their social media output is indistinguishable from humans, the value of connecting on social networks drops to zero. This is one reason, presumably, why Meta recently bought Moltbook, the short-lived fad social network (supposedly) made up entirely of AI agents. They’re trying to get in on the ground floor of a future where AI agents dominate social media.
Admittedly, we users helped bring about this endgame: we already mistook our ever-more-curated online personas—those “most people think X about Y but I discovered Z” posts—for authentic engagement in the first place. But that also leaves most of us with little to mourn, as agents flood platforms that always prioritized any engagement over human connection anyway. If there’s hope in our increasingly garbage-filled online world, to me it’s this: as social media submerges under the AI deluge, we’ll have to find new ways to connect, online and off. Let the bots have the platforms, I say. They can spend eternity influencing each other.
What do you think about this story? Submit a letter to the editor at [email protected].
