aixa rizzo

The tweet that started it.

Not a deal, not a brief. Peanut fixed my everyday problems with payments abroad, so I posted it.

No hook, no planning. Three transfers, a bad rate, and the app that ended it. People started asking for the link.

I interviewed the founder.

The referrals kept coming, so I went to the source.

I wanted to know how they built it, so I sat down with the founder and recorded the conversation.

The people who joined because of me.

Each dot is one person. 130+ people moved their money to an app because someone they follow uses it. No ad spend, no giveaway. That's what a personal brand does when the product is actually good.

So they hired me.

The job: the lineup and the hooks. And we played the rage bait: I staged my own rejection from the card, and the timeline took it.

Months of shilling the app for free, then the card launches and I don't get one. That was the bit. The rejection tweet alone did 12.2K views.

The launch needed more voices than mine.

I found the creators, matched each one to an angle, briefed them, and ran the whole activation. No scripts, no copy-paste captions. They posted in their own voice, because the point was that it shouldn't look like a campaign.

For years I built everyone's brand but my own.

Then I wanted a job in crypto, and everyone told me the same thing: the way in is a personal brand. So I started posting.

In under a year I was signing brand deals and getting paid to post. People ask how it happened, and I think it's the advertising. A post is just an idea that has to land fast, in a small format, and make someone feel something.

Brand deals.

The receipts.

Apparently, when I complain, I go viral.

My three most viral posts. Not sponsored, not boosted, over 2.5 million views between them.

Conference coverage.

When the writing works, they hand you a mic.

They asked for a social media manager. They needed a business model.

I came in to run the Instagram and post about a gin festival. It took about two weeks to see the actual problem: the whole thing was built to sell tickets, and tickets were never going to pay for it.

So I went after the people with the money.

I needed the biggest names in the industry. No warm intros, no agency. I opened LinkedIn and started messaging the people who ran marketing at the brands I wanted.

Those cold messages turned into the sponsorships that anchored the festival: Bombay Sapphire, Schweppes and Fever-Tree, three brands that alone covered 30% of the total budget. Under them, I grew the exhibitor base to 50+ gin brands through B2B partnerships.

The people buying tickets were never the ones who'd pay for this.

A festival like this gets funded by the brands, not the door. They pay to put their gin, their tonic and their name into thousands of hands for one night. The real customer was the industry.

I wrote the business model. So they made me a co-founder.

I went from running social media to Commercial & Marketing Director, because the plan that turned the festival into a real company was mine.

That meant owning the commercial side: sponsorships, the exhibitor pipeline, event sales, planning, and the accounting behind all of it.

Then we filled La Rural.

The second edition took over La Rural in Buenos Aires: more than 30 distilleries, 63 gins to taste, food, four DJs and Los Tipitos live. A rebrand of the first event, which had already pulled a 3,000-person crowd at the Hipódromo de Palermo.

The rest of the job.

Beyond the model and the sponsors, I ran the commercial and marketing operation.

Advertising, social and email campaigns that doubled ticket sales. And a team of 25 across brands, social media and content.

Hired for growth. Promoted to product marketer.

The job was growth and email marketing. Within a few months it turned into product marketing, because every experiment kept pointing at the same thing: the emails, the ads and the product weren't talking to the same person.

Nobody inside agreed on who the customer was.

Marketing was writing for one persona. Sales was chasing another. Product was building for a third.

We needed a sharper story than "design tool".

Venngage was going up against Canva. Competing template-for-template was a fight you lose slowly. The wedge was data visualization: the one thing Venngage did better than a general design tool. So that became the story. Then we tested it.

Data visualization, about things people cared about.

I set up a UGC content engine. The videos pointed straight at what the audience was already watching: Taylor Swift's Time cover, celebrity marriage timelines, the Taylor and Travis story. Same product, told through pop culture instead of spreadsheets.

The comments told us who Venngage was really for.

We were building for HR and ops managers. The people sharing these videos, and the people in the replies asking how to make one, were students. The ICP stopped being a guess.

They don't have SaaS budgets. So we changed the motion.

The easy move was to keep chasing whoever had a credit card.

We met students where they already were: partnerships with universities and high schools, putting Venngage in front of them through the institution instead of asking a broke 20-year-old to expense a subscription.

Accessible data visualization. Nobody else had it.

At the same time, accessible websites stopped being optional. Charts and infographics are some of the hardest things on the web to make accessible.

So we built the accessibility features, and marketed that as a real reason to choose Venngage over the generic option.

I built the machine that made a security firm sell.

Marketing Operations Director at Coinspect, the crypto security firm that audits smart contracts. I built the B2B pipeline, ran the MetaMask partnership end to end, and translated the hard technical work into plain language people could actually buy.

Qualified pipeline grew 60% that year.

Revenue operations.

From first touch to signed audit, one pipeline.

The Wallet Security Ranking: security research turned into a public product.

Coinspect tests the security of crypto wallets and publishes the scores. I supervised the UX of the ranking and built the content engine around it.

MetaMask cites Coinspect on their homepage. That was the deal.

A partnership between the security firm and the biggest self-custody wallet in crypto: the security ranking, the writing, the co-marketing.

Two technical articles.

Security research translated for the people who decide: how wallet security keeps you safer against scams, and how the ranking methodology works.

Marketing a security firm means the receipts are public.

The launch was one night. The role was six months of building the product it launched.

I joined Sophon as Ecosystem Product Marketing Manager: positioning, copy, go-to-market, lifecycle and content for a consumer crypto app built from zero.

The launch dinner.

Personalized place settings with each name projected in light. The Sophon sky wrapping the whole room. A card holder with a personalized card for each guest, and a token to exchange in the app.

The gift was a camera. Because everyone in the room was a content creator.

No paid push.

The event went viral, no paid promotion. That was the design: every action and every piece of merch was designed content first, imagining the post it would produce before deciding the idea.

The merch kept getting called the best in the space. So we made one jacket.

The Aura Genesis experiment: one jacket, one auction, one leaderboard.

The jacket, physical plus NFT, went to the highest onchain bidder. Half of the final bid became a prize pool for whoever the community decided had the most aura: every user got three upvotes and three downvotes, and the final rankings set each share.

The rest of the job.

Everything from the app's first onboarding words to every campaign around it.

The launch narrative, written down.

Long-form explainers translating the deep tech, zkTLS, the Social Oracle, lifestyle crypto, into language people read. Written by me, published by Sophon.

The brand, on screen.

Beyond the announcement, the brand lived on film: the launch commercial and the merch drop, both built so a crypto app could feel like culture.

Social reports are all numbers. They tell you the content performed, not the reaction it left.

Google Analytics in one tab, Kaito, Sprout Social, each with its own login and its own bill. I put an agent on all of it.

Below is one day of a market's timeline. Every dot is a post. The sweep separates it: four small clusters you'd stop scrolling for, and the big pile you wouldn't.

Four sweeps into one ranked brief.

It sweeps on a loop, groups the posts into narratives, drops what the brand doesn't care about, and ranks what's left into one brief.

Four hundred posts in. Three worth acting on.

One real day for one brand, names stripped. The agent read four hundred posts and kept the three that needed a decision.

One read, one real finding: this account grows on engagement, not reach.

Beyond the daily triage it compiles the periodic read: what reached, what landed, what to do. Everything below is one real window from one brand. An example of the findings it hands you, not a law.

It watches your competitors too.

Point it at a list of competitor accounts and it tracks what they launch, what angles they push, and which of their posts get replies.

Nobody has time to read a long report. It ends in a to-do list.

These four are from one brand's week:

Everyone sprays and prays. What gets replies is a personalized message.

The same tools surface the same leads. Personalized messages get replies, but the research behind them is time-consuming, so nobody does it.

Prospects post their pain points. The same ones you already fix.

Companies write down what hurts, in job ads, announcements and reviews. Including the problem your clients pay you to fix. You just have to be reading.

You give it your ideal customer profile. It finds the companies, the contacts and the proof.

The ICP becomes a scoring rubric. From there it finds companies with live buying evidence and drafts a first message that opens with the proof.

It finds the trigger, then offers the solution.

A trigger tells you something changed. The pain read tells you what to say about it. Before it writes a word, it analyzes the company's own public information, the careers page, the homepage copy, the reviews, the tech they list, and infers the specific problem they're living with right now.

So the opener isn't "hey, I like what you do, let's connect synergies."

The first sentence is a verified fact, or the message doesn't send.

A cold template could be sent to anyone. This one gets taken apart and rewritten:

Ten researched leads from one URL. Same day.

A real run. Input: the URL of a UK managed IT provider, nothing else.

No guessing allowed.

Every company is scored against the rubric, 0 to 100, and tiered. If a fact can't be verified, the field stays blank.

Good replies take hours of scrolling before you write a single word.

Most of the work is the hunt for the right tweet to reply to. So I automated the hunt, that part I built fast. Making the replies actually good is what this project became about.

Sounding human took months.

The first versions wrote replies that read like slide titles. Smart-sounding, empty, and everyone scrolled past them. Closing that gap took calibration, kill-lists, and rewrites.

Every reply has to survive a filter built from the ones that failed.

A specific set of kill-checks, each catching a way a reply goes empty. The demo runs a thesis-slop draft through the gate.

You cannot save a weak target with better wording.

When a reply sounds bad, the problem is almost always the tweet it answers: nothing specific to react to. So every target clears seven checks before a reply is written.

An open timeline becomes a short list of sharp, human replies.

By default, the agent drafts. A human posts.

The voice is a set of rules, so when a brand wants a different tone, I turn a dial.

Retuning is fast. Personality up or down, product mentions more or less frequent, banned words swapped, humor on or off. A brand asks for drier, and drier ships the next day.

There is a version that posts on its own. I taught it to behave like a human.

Automated posting gets accounts flagged because bots act like bots: instant, perfect, always on the exact same schedule. So I humanized it. Here is one being written.

Ghostwriting takes interviews, drafts, and months.

But everything you need is already public: every pattern, every tic, every belief is sitting in what they already wrote.

What you post on social media says more about how you think than you realize. The casing, the punctuation, the jokes you repeat, the things you never say. Analyze enough of it and you can rebuild the voice, and a good part of the person behind it. I turned that analysis into a system.

A free reading nobody asked for.

This is how I sell the service. I pick who I want to work with, run their handle through the agent, and send them their reading. No pitch, no deck. Nobody can ignore a page that says something true about them.

The reading covers who they are, what they believe, the eras they've moved through, and the thing they don't say out loud.

It finds the voice your audience rewards most.

Even the one you barely use. This chart is from my own reading.

This is what it wrote about me.

Every section, from my own reading.

This is that reading, section by section, with what each one does and what it found. Every prose line has to run a named psychological technique, and every claim needs a receipt.

The last page is three tweets you'd have written.

Every reading ends the same way: three posts in your exact voice, ready to publish, projected against your own best-performing cluster. Not advice. The actual work, done. Your voice handed back to you, sharper.

The build, one level deeper.

The short version of the architecture doc. How the clone gets built, how it learns a market, and why the psychology isn't vibes.

A hundred freelancers, nine pods, nobody whose job is posting.

Podcasts, launches, events, tutorials, founder takes. All of it needs distribution, and the voice changes depending on who's writing that day.

So I built the intern. Now anyone in the network can ship sounding like Myosin, the posting runs itself, and a human stays in the loop to approve every post and calibrate the agent.

The team was skeptical. Our clients use it now.

The whole loop, the part that normally lives across six tools and three people's heads, running on the rules I gave it: what makes a clip, where to cut, how the caption reads.

There are a lot of clipping tools out there, and the team had tried them. This one replaced the one we paid for.