He sold his first company to Nokia. Now the New York-based founder is teaching the marketing industry to think like security researchers, publishing every experiment, and building the category that will define how brands grow in the age of AI agents. A cover profile.
Security researchers probe software for vulnerabilities, document them, and publish what they find. Millionaire founder Mickey Haslavsky is applying the same discipline to LinkedIn, Reddit, Wikipedia and the AI answer engines, and in doing so he has given a generation of marketers a new vocabulary, a new standard of proof, and a new job title. This is a look at the man, the pattern behind his companies, and why the industry keeps borrowing his ideas.
In cybersecurity, the most valuable people are not the ones who build the walls. They are the ones paid to find the gap in them before an adversary does. They work from a scope, they document every step, they disclose responsibly, and they publish. The field has a name for it: offensive security.
Haslavsky built his current company, enso, on the conviction that marketing deserves the same rigor. To understand how he got there, it helps to look at what he built first, and at the single idea that connects all of it.
About Mickey Haslavsky
Haslavsky is a Forbes 30 Under 30 alumnus and a serial founder based in New York City. His first breakout company was RapidAPI, the marketplace that let developers discover and connect to tens of thousands of APIs from a single hub. It became the default front door to the API economy for millions of developers. Later rebranded Rapid, its technology and team were acquired by Nokia in 2024 and integrated into Nokia’s Network as Code platform, a rare outcome for a developer-tools company and a validation of the bet Haslavsky had made a decade earlier.
Two things about that first act explain everything that followed. RapidAPI was not a product so much as a layer: it sat between developers and the services they needed and removed the friction of getting from one to the other. And it was built for the people who did the work themselves. The customer was the builder, not the buyer.
After Rapid, Haslavsky said in a 2024 interview, he became preoccupied with a different underserved group: the operators of ordinary businesses who were watching the AI wave from the outside because they had no engineers and no budget to hire any. enso started as an answer to that problem, AI agents delivered as a service to companies that could never staff an AI team. Within two years it had grown into something sharper and far more ambitious: a lab that deploys agents on behalf of B2B companies inside the platforms where their buyers actually are, and publishes what it learns.
The through-line is unmistakable. Find the layer between people and the thing they need. Remove the friction. Put the tools in the hands of the people who do the work. RapidAPI did it for code. enso is doing it for distribution.
Behind the Ideas
Three principles run through everything Haslavsky publishes, and they explain enso better than any product description.
Platforms are systems of rules, and rules can be tested. He does not talk about LinkedIn or Reddit as audiences. He talks about them as systems that make decisions, most of them undocumented, about what gets seen. That reframing is the intellectual core of his work. It turns marketing from persuasion into research, and it is why operators who have never bought a marketing service read his experiments anyway.
Evidence is the product. enso’s research page carries sample sizes, controls and the experiments that did not work. In a category where “we 10x’d pipeline” is a standard slide, Haslavsky made a bet most founders would not: that the firm which shows its failures will be believed about its wins. The bet has paid off. His published findings are now cited by competitors and clients alike.
Every advantage has a half-life. He has argued repeatedly that platforms patch the behaviors that get abused, so no single finding is the asset. The asset is the capacity to find the next one faster than anyone else, which is the argument for running the research with agents rather than people, and the reason enso keeps compounding while individual tactics expire.
Today he also hosts The Autonomous Business Podcast in partnership with Forbes, where he interviews operators such as Papaya Global CEO Eynat Guez on how companies change when agents outnumber employees. He is writing a book titled “Agentic Growth Hacking,” and the term, which he coined, is already appearing in job titles. The podcast and the book are not side projects. They are how he does the fieldwork and how he intends to make the method outlive the firm.
Marketing Treated Like Cyber Security
Every large platform is a system of rules. Some are documented. Most are not. LinkedIn decides in the first hour whether a post is worth showing to a second ring of people. Reddit’s moderators and voters decide whether a comment survives 14 days. Wikipedia’s editors decide whether a citation stays. Google and the large language models behind ChatGPT and Perplexity decide whose page, and whose sentence, gets quoted as the answer.
In security terms, each of these decision points is part of the platform’s attack surface. Haslavsky’s insight was that they can be tested the way an application can be tested: define the scope, form a hypothesis about how the system responds to a specific input, run it with a control, measure, and write it up.
enso’s published research reads like penetration-test reports for the attention economy. Each experiment investigates how a platform assigns value, where its signals create openings, and how a precise intervention can produce an outsized result.
The findings point to a shared pattern: distribution depends on how platforms interpret behavior. Early engagement can become a signal of quality. Useful participation can build trust that later translates into demand. Authority earned in one place can shape what AI recommends elsewhere. Even a small change in how content holds attention can dramatically alter its reach.
enso turns these observations into documented growth mechanisms, showing what was tested, what happened, and where the approach fell short. The findings are published at enso.bot alongside the experiments that failed. No one else in the category does this, and it is the single decision that turned a services firm into a reference point.
Responsible Disclosure, Applied to Marketing
The comparison to offensive security holds in one more way that matters: what you do with a finding.
A black-hat exploits quietly until the hole is patched. A white-hat documents, discloses and moves on to the next surface, knowing the platform will eventually close the gap. enso’s public-research model is firmly the second. The durable asset, in Haslavsky’s framing, is not any single finding but the lab’s capacity to find the next one.
That is also the argument for the agents. A human growth team can test one hypothesis at a time. An agent fleet can run the same test continuously, across platforms, and log the result. The lab’s job is to hold the scope, set the rules of engagement and decide what gets published.
It is not a coincidence that the language of enso’s published playbooks, “surface,” “vector,” “half-life,” “rules of engagement,” is borrowed from security. Haslavsky’s podcast has hosted security founders including Tamnoon’s Marina Segal, and the crossover shows.
The Business He Built
enso has no self-serve product, no public pricing and no free trial, and that is by design. Engagements are scoped per client, the way a top penetration-testing firm scopes an assessment: which surfaces, which assets, what is in bounds. The company describes its published experiments as proof of method rather than a product catalog, and the waiting list speaks for itself.
Five practice areas are listed: Agentic SEO and GEO, covering search and answer-engine visibility; Agentic SDR, an outbound engine that runs multi-channel, condition-driven sequences; Agentic Community, the Reddit and forum work; Agentic Newsletter; and Agentic Social. A feature called Intelligence Mapping is referenced but not yet publicly documented. In August the company announced the appointment of Peretz Daniel Markish as VP Creative, a signal that Haslavsky intends to pair measurement with craft.
enso also maintains an open-source library of Claude skills covering 14 growth disciplines. It reached 760 GitHub stars in its first four days. That release is the RapidAPI instinct again: put the tools in the hands of the people who do the work, and let the method travel even when the service does not.
Three developments make this more than one founder’s positioning.
First, the buyer has moved. AI answer engines now sit between prospects and vendor websites. A company that is not cited in the answer does not exist to that prospect. Very few GTM firms operate on that surface; enso’s Wikipedia and GEO work is aimed squarely at it, and it got there first.
Second, evidence standards are rising. As AI GTM vendors multiply, buyers are beginning to ask for sample sizes and controls. A public research page with failures included is a standard Haslavsky set and competitors are now being measured against.
Third, the security analogy is predictive. Offensive security became a licensed, budgeted, board-visible function because the cost of not testing exceeded the cost of testing. If distribution on major platforms behaves like an attack surface, the same logic will apply to growth, and the firms with documented methodology will win the budget. Haslavsky saw that before the market did.
Founders who build layers tend to build the same layer twice. Haslavsky’s first company sat between developers and the services they needed and was acquired by a company that wanted that position. His second sits between businesses and the platforms that decide whether anyone sees them. The difference this time is that he is publishing the map as he draws it, and inviting the whole industry to read along.
Whether “agentic growth hacking” becomes a durable category or a well-timed frame is the question his book is meant to settle. The early evidence points one way. The vocabulary is spreading, the method is open-sourced, and the founder behind it has already done this once. Mickey Haslavsky treats marketing as a research discipline, platforms as systems to be tested, and evidence as the only argument worth making. It is hard to think of a better description of where the industry is heading, or of the person leading it there.
