This founder built an AI agent that does around $2M ARR

^{A week in AI is like a year in other industries. I hope these issues become your weekly source of AI information, inspiration, and ideas. If we haven’t met before, I’m Amanda Smith. I write about AI and the fascinating folks who are building in this brave new world.}

This week's founder spent two decades in performance marketing before building an AI system that changes how teams manage paid media. And they’re running big brands and budgets on it. Another category leader crushing it outside the spotlight.

Subscribe | Past essays | Let’s collaborate | Insane Media

🤝 This edition is brought to you by EasyLlama

Think you’re compliant? You might want to check.

Turns out, most organizations don’t struggle with access to compliance training; they struggle with knowing what’s actually required.

And just one missed compliance training puts your entire org at risk. Unfortunately, these missing training requirements aren’t always obvious, so by the time you catch them, it might already be too late.

Want to see where you stand right now? EasyLlama can help. Their Compliance Grader gives you a personalized report in two minutes, with a clear picture of what training your org is legally required to have and what’s recommended.

Answer a few questions to see what training you need.

_{We partner with a select group of brands that we use or admire, allowing this newsletter to remain free and independent. Reach out for your campaign here.}

Seva Ustinov has spent the past two decades in performance marketing – now he’s built an AI agent to automate 90% of his daily work. He’s doing it for brands with big budgets of $500K+ per month. 

After successfully scaling a paid marketing agency to a team of 100, working on large-scale paid media across big tech brands, Ustinov saw just how taxing it was to manage massive amounts of data – analyzing dashboards and allocating budgets. 

In 2004 when the new generation of AI models emerged, Ustinov realized he could build something that could plan, remember context, and evaluate their own work. 

“Our initial go-to-market was very founder-led. Before writing the product, I built a prototype of the entire concept in Lovable and used it to make the idea tangible. Then, I hired 7 senior engineers on a part-time basis and gave all of them the same task, working independently. After six weeks, three of them had produced tangible results, and they became our initial AI team,” Ustinov said. 

“For validation, I went directly to growth leaders I already knew from the performance marketing community. I explained what we were building and offered to connect the agent to their accounts in co-pilot mode for a month in exchange for honest feedback. What surprised me was that they were willing to pay around three times more than they had previously paid for analytics dashboards.” 

Ustinov learned the founder has to be part of the product demonstration. People want to talk to someone who has experienced the problem themselves. He said they’re keeping that principle as they scale Plurio. 

Plurio is built for marketing teams where paid acquisition is the main growth channel. With thousands of ad entities that run at the same time, multiple products and geos, and where delayed conversions, complex funnels and large amounts of advertising data make manual optimization completely ineffective.

For example, for one of their clients (Finom), Ustinov said they combined full-funnel attribution with conversion predictions for campaigns where conversions were delayed. On Meta, they tracked frequency and moving-average CAC by creative to identify when an ad was becoming saturated, while Cost per Sign Up acted as a leading signal in markets where conversion data was delayed.

“The result was a system that could react to these signals before the final conversion data caught up. From November to March, Finom increased new clients by 38%, and reduced CAC by 33%.” 

But there’s nuance to all this. You can’t simply connect an agent to an ad account and assume it will always make the right decisions. Success comes down to quality of data, agent rules, and having humans at the review stage. “An agent might decide to pause an ad because the numbers look bad, even though that ad is actually bringing in customers with high lifetime value,” Ustinov explained. 

In 2025, Ustinov moved Plurio’s marketing, sales, customer success and development teams into a shared AI workspace. For example, an agent can analyse emails and call transcripts with a potential customer, research the company and draft a response in about 100 seconds. This has given them a real example to show customers. 

Using agents to run Plurio hasn’t stopped them from hiring humans though. They’re currently hiring a Head of Performance Marketing for a cushy $120-180K. 

Externally, the brand focuses on LinkedIn, webinars and events, and producing research. Marketing a technical product should focus on making the value tangible, show what the product can do, and real results. 

Ustinov recently released a researched report that reviewed 17 of the fastest-growing sales-led AI companies – analyzing more than 500 primary sources to understand what’s driving their growth. 

“Some of the clearest patterns were that successful AI companies tend to lead with a very specific ‘wow’ use case, demonstrate value on the customer’s own data, build trust early, and turn successful outcomes into proof for the next sale.” 

While Ustinov is bullish on removing marketing’s routine work, he also sees a new challenge. 

“As more marketing work is automated, companies will need people who know how to work with AI and manage these systems. They will need to understand what the AI agent is doing, set the right rules, and make sure the quality of its decisions stays high.”

If one of these stories stuck with you, I’d love to hear which one.

Speak soon,
Amanda

P.S. If you want to get a founder feature about your own story, reply to this email. If you’d like to reach our newsletter audience (founders, creators, and marketers), click the button below.

If you’re new here, I’m over the moon you’ve joined us! To help me craft content that’s actually useful (and not just noise in your inbox), I’d love it if you took 1 minute to answer this quick survey below. Your insights help shape everything I write.

Dive into our other newsletters - where psychology meets the creator economy, e-commerce marketing, and Human resources.