The Company Gaming That AI Agents Can Out-Hack the Internet’s Biggest Platforms

A little-known startup called enso is drawing serious attention in marketing and venture circles this year, not because it sells a new app, but because it treats growth itself as a science experiment, one run entirely by fleets of autonomous AI agents.

According to The Next Web, enso doesn’t operate like a typical software company. There’s no dashboard to log into, no free trial, and no self-serve pricing page. Instead, clients hire the company the way they might hire a security firm to test their defenses, except here the target isn’t a network, it’s the algorithms behind the world’s largest platforms. enso’s agents run structured experiments to figure out how those systems decide what gets shown to people, then the company publishes what it learns, successes and failures alike.

From Marketplace to Research Lab

enso wasn’t always built this way. The company originally launched as a marketplace for AI agents, raising a $6 million seed round backed by the venture firm NFX. That version of the business didn’t stick. enso scrapped it and rebuilt itself entirely around growth research, a pivot serious enough to justify a subsequent $25 million raise. Most startups would rather forget an abandoned first attempt. enso instead documents it openly on its own website, treating the reversal as part of the story rather than something to hide.

The company’s founder has coined a term for what the business now does: agentic growth hacking. It’s a callback to an older idea. Growth hacking, as a concept, dates to 2010 and described the scrappy tactics that built companies like Airbnb and Dropbox before platforms closed the loopholes and the term faded from relevance. enso’s pitch is that those loopholes never disappeared, they just require machines fast enough to find them before platforms notice and shut them down.

Selling Outcomes, Not Software

The bigger idea behind enso connects to a shift some venture investors have started calling “service as a software.” Traditional software businesses sell access: a login, a set of features, and the hope that a human will use them productively. This newer model skips that step and sells the finished outcome directly, produced largely by AI systems working without a person managing each task. Instead of a features demo, the company shows the actual result and the data behind it, including experiments that didn’t pan out.

That’s a meaningful departure from how most marketing technology gets sold. It also happens to fit a moment when two of the biggest channels marketers have depended on for years are both breaking down at once.

Two Broken Channels

Paid advertising is the first casualty. Digital ad costs have climbed sharply this year, driven partly by how easy it has become to launch a new product and immediately start competing for the same audience. As more companies bid for the same limited attention, the price of a single click keeps rising, in some competitive categories reaching many times what it cost a decade ago.

The second casualty is organic content and search visibility. For roughly two decades, the standard playbook was to publish useful material, rank for relevant search terms, and convert readers into customers over time. That model assumed people would actually read the content. Increasingly, they don’t. AI-powered answer engines now summarize the best available sources and hand users a direct response, often without sending any traffic back to the original page at all. The content still exists. The audience it was built to reach is going elsewhere.

Ad agencies, meanwhile, are shrinking even as overall advertising budgets grow, a sign that companies are cutting agency spending line by line rather than eliminating it outright. Businesses that once paid for a team’s hours are increasingly satisfied with AI-generated drafts they can produce themselves in minutes.

Where enso Is Positioning Itself

enso’s bet is that the next competitive battleground isn’t search rankings or ad auctions, it’s whether a brand gets mentioned by name when an AI system answers a buyer’s question directly. The company’s research is aimed specifically at that surface, and its early work there gives it a head start most competitors don’t have yet.

Several factors could support enso’s growth from here. The company has effectively created its own category with “agentic growth hacking,” which tends to attract investor interest precisely because there’s no existing product to compare it against. Each experiment the company runs also builds out an internal map of how major platforms make decisions, information that becomes more valuable as those platforms continue changing their rules. A recent hire, a VP of Creative, hints that the next phase of the business may pair its research-heavy approach with more traditional brand and creative work.

The Risks Ahead

None of this comes without real exposure. Everything enso’s agents uncover exists, by definition, because a platform didn’t intend for it to be exploited, and platforms retain full authority to reclassify a clever tactic as abuse once they spot it happening at scale. A fleet of automated agents is a far more visible target than a single employee testing an idea quietly.

There’s also the question of whether the company’s methods can outlast dependence on one founder’s judgment. enso appears to be addressing that concern directly, through a forthcoming book, public interviews, and open-source materials meant to formalize its approach so it doesn’t rely entirely on one person being in the room.

Growth hacking was widely considered a dead concept a few years ago, largely because humans couldn’t keep pace with how quickly platforms changed their rules. enso is wagering that machines can. Whether that bet pays off at the scale investors are hoping for remains to be seen, but for now, it’s a company worth watching closely.