@PrajwalTomar_: An AI agent ran a real company's Meta Ads account for 31 days. No human touched it. The guy handed it $1,500, full cont…
Summary
An AI agent autonomously managed a real company's Meta Ads account for 31 days, writing ads, optimizing budgets, and only faltering when a human intervened. The experiment highlights data ownership and open-source tools as key enablers for AI-driven advertising.
View Cached Full Text
Cached at: 07/31/26, 04:52 PM
An AI agent ran a real company’s Meta Ads account for 31 days. No human touched it.
The guy handed it $1,500, full control, then walked away and published the entire log.
It wrote the ads, killed the losers, and raised its own budget on day 12 when one ad hit $1.29 a lead.
The single worst crash of the whole month? The one time a human stepped in.
Everyone is arguing about whether AI can write good ad copy, and missing the real unlock: data ownership. The whole setup pipes your Facebook, Stripe and analytics into one warehouse, so the agent can see which exact ad made money, then acts on it every 48 hours.
Agencies charge tens of thousands a month for a slower version of this loop.
The wild part: it’s just two open source tools and a cloud box. Anyone can copy this tomorrow.
(full breakdown in the article below)
Similar Articles
I replaced my old job with an AI agent
The author recounts building an AI agent using GPT-5.5 Codex and Meta Ads MCP that automates the entire media buying process, from research and copywriting to ad deployment and reporting, replacing the role of a traditional media buyer for less than $20/month.
@PrajwalTomar_: I genuinely did not expect this to work this well. I hired an AI employee at the agency this week. It runs on an open-s…
A user shares surprise at how well an open-source AI agent works for client research using Hyperagent, which lets you swap between models like Kimi K2.6, GLM 5.2, DeepSeek V4, and Qwen 3.7.
I let an AI agent run my company's social media unattended. Here is the full run, failures and all.
The author shares the full results of an AI agent's first unattended run managing their SaaS company's social media, highlighting four failures that it gracefully recovered from without human intervention.
@PrajwalTomar_: Honest take: the part that got me was watching it work. It opened a browser, clicked through the site like a real analy…
The author built an AI employee using Hyperagent and open-source models like Kimi K2.6 to automate client research, generating pre-call briefings. Multiple models are available for swapping.
@METR_Evals: Could an AI company lose control of its own agents? To find out, Anthropic, Google, Meta, and OpenAI let us (1) test th…
METR published its first Frontier Risk Report, assessing the risk of AI companies losing control of their own agents. The report involved testing the best internal models from Anthropic, Google, Meta, and OpenAI with chain-of-thought access and reviewing non-public information about capabilities and alignment.