Partner Story · in conversation with Alexandra, founder of Laexa Digital. This is a written interview submitted through Submit Your Story. The claims below are her own, presented as they were given to us. No payment was involved.
Alexandra, founder of Laexa Digital. Photo: courtesy of the founder.
Alexandra's first business had nothing to do with software. She was a nail technician. Her days were appointments, clients, nail products and long hours of work that had no connection to code.
"If someone had told me a few years ago that I would be building an international digital marketplace, I probably would have laughed," she writes in the story she published on her own site.
Today she runs Laexa Digital, a platform built in Romania with a marketplace for digital products and two AI subscriptions that are live and on sale. She still does not call herself a programmer. She built all of it anyway.
A planner, a question, and a first website
Like many people who find their way into the online economy, Alexandra started by selling what she knew. She created digital resources for the beauty industry: nail guides, and a beauty planner with daily motivation and content ideas for people working in salons. It was her first real experience of turning an idea into a product.
Then came the part nobody warns you about. The product was ready. She still needed somewhere to sell it.
Every platform she looked at took a cut or a monthly fee for a storefront, and the question that would change everything arrived almost by itself.
"Why should I keep paying for someone else's storefront when I could try building my own?"
She had, she says, no idea how ambitious that question would become.
Her first website was BeautyBrain.ro, a store for her own products. She did not write the code herself. She used ChatGPT as a step-by-step guide: asking, following instructions, making mistakes, and working out what went wrong. Then she built a site for a client, and another for her partner's car detailing business. None of it made her a developer. All of it taught her that technology was not a world she had to stay outside of.
Watching other creators struggle
While the websites were teaching her how things worked, Alexandra was also learning digital marketing and sharing what she learned on TikTok. Her account grew to nearly 35,000 followers, and she began offering mentoring and account audits to other creators.
Those conversations kept circling back to the same problem. People were learning how to make digital products. Selling them was another matter entirely. They paid for subscriptions, built storefronts, posted every day, and still struggled to reach anyone who might buy.
"Many platforms offer a storefront," she says, "but creators still have to attract customers, manage their content, organize their work, and pay for multiple subscriptions to different tools."
That is the gap Laexa Digital was built to close.
Three layers: a marketplace and two AI tools
The first layer is the marketplace. Creators open their own branded store and sell e-books, courses and templates, with 0% commission on paid seller plans and payments through Stripe. At the time of writing it lists 27 products from independent creators, most of them from Romania, priced from about €8 to €54. Anyone can open a store.
The second layer is Brainy, an AI personal assistant for people who run a business alone. It keeps tasks, calendar and notes in one place, connects up to three email inboxes, searches the web, writes a morning report of what needs attention today, and sends reminders to Telegram or WhatsApp. It costs €9 a month.
The third is FlowPost, an AI assistant for social media. It helps plan content, writes captions, generates images and schedules posts across Instagram, Facebook, TikTok, YouTube, Pinterest and X, so a whole week can go out on time without the founder touching it. The Solo plan is €19 a month and Business is €39. Both tools together, as the Laexa Hub, cost €42 a month with a 7-day free trial. All the plans and limits are published up front.
A fourth product, AI Hosting, is still in development. The idea is to let people create and train AI agents for customer service, bookings and documents, and then sell those agents to other businesses on a subscription.
Building something she didn't know how to build
A website with step-by-step guidance was one thing. A multi-vendor marketplace was something else. Suddenly there were user accounts, vendor permissions, payments, subscriptions, integrations and dashboards, all of which had to work together.
Alexandra moved from ChatGPT to the AI coding tools that write and run code directly: OpenAI Codex and Anthropic's Claude Code. She says she has worked with OpenAI's GPT-5.6 Sol and GPT-6 and with Anthropic's Claude Fable, Opus and Sonnet, and that Fable has been especially useful for the hardest coding tasks.
"Sometimes a task that sounded simple turned into hours of troubleshooting," she writes. "Sometimes fixing one problem revealed three more."
But the most interesting thing she did was not choosing a tool. It was refusing to trust one.
"Codex gave us a NO-GO"
An AI assistant, Alexandra noticed, can produce a solution that looks convincing while hiding a problem inside it. Something works at first glance and breaks something else. A technical decision needs a second opinion she was not qualified to give.
So she built a bridge between Claude Code and Codex, so the two can talk to each other while working on the same project.
"Claude Code can work on implementing a feature while Codex independently reviews the changes, identifies issues, and recommends corrections. If the review reveals problems, the implementation goes through another round of improvements."
She uses the bridge in two ways. For tasks with many decisions, she asks the two assistants to collaborate from the start. For simpler ones, Claude builds first and then asks Codex to review. And sometimes the review comes back with a verdict. Her favourite message, she writes, is one Claude has returned more than once:
"Codex gave us a NO-GO, and he's right. He found two issues I missed."
"I love those moments," she says. "Not because I enjoy finding errors, but because the disagreement is useful."
It is the same principle any good team uses: the person who writes the code is not the person who checks it. Alexandra simply staffed both roles with AI. She still tests every result from a user's perspective and makes the final decisions, especially where users, data or payments are involved.
She is also clear-eyed about the limits. Two assistants agreeing does not make the answer right. They can miss the same issue, misunderstand the same requirement, or agree on code that still needs testing. She treats the bridge as an extra layer of review, not a replacement for security checks or professional expertise when it is needed.
"This workflow has made it possible for me to tackle projects that once seemed completely beyond my abilities."
Three real customers, and why she counts them out loud
Laexa is young, and Alexandra is direct about it.
The first paying customer came in July 2026. Since then there have been three paying customers in total, with two subscriptions running at the same time for several months. Today there is one active paying subscriber and a small number of people on the free plan.
None of it came from advertising. She has not spent anything on paid promotion so far, and she does not yet know how much faster Laexa could grow with a marketing budget.
"These aren't impressive numbers by startup standards, and I'm not going to pretend otherwise. But they're real, and they matter to me," she says. "I would rather share three real customers than claim hundreds I don't have."
Her biggest challenge now is not building more features. It is getting Laexa in front of the people who could use it, and earning their trust.
What AI changed, and what it didn't
Alexandra does not believe AI turns anyone into a software engineer. She still has a lot to learn, she says, and does not pretend to understand every line of code behind her platform.
"Using AI doesn't eliminate mistakes or difficult decisions. It gives me a way to approach problems I wouldn't previously have known how to solve."
What it gave her, in her words, was the ability to start. Instead of being stopped by everything she did not know, she could ask, experiment, make progress and learn on the way. She could turn an idea into something real without first mastering a whole new profession.
For anyone thinking about building a product without a technical team, that may be the most useful part of her story. AI did not make the work disappear. It made the work possible.
"Perhaps the most important change wasn't technical. It was realizing that I no longer had to abandon an idea simply because I didn't know how to build it."
What comes next
Alexandra's priority now is growth: more creators on the marketplace, more entrepreneurs using Brainy and FlowPost, and finishing AI Hosting. The longer goal is a place where people can build and train AI agents and businesses can find the agents they need.
Laexa is based in Romania but open to users anywhere. She would especially like to hear from independent creators, small businesses, partners and communities that help people start using AI in practice.
Her own account of the journey, and of the Claude Code and Codex bridge, is on the Laexa Digital press page. The platform is at laexadigital.com.
---
This is a Partner Story: a written interview with a founder building with AI, submitted through Submit Your Story and published free of charge. Statements about the company's products, customers and results are its own. Building something with AI? Tell us about it.
Would you try this product?
