Loading page…
A B2B app that turns the mechanic's inspection into an estimate the car owner can approve from their phone — pre-sold as three 5K lifetime licences before a line of code existed.
about €20,000 in government grant money on top · an angel syndicate lined up a half-million-dollar round subject to accelerator graduation
"Three out of three said: yes, Patricia, take my money and build it — and build it faster." She was selling prototype number 26, and it was made in PowerPoint.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Donatas Jonikas tells the Auto Pixie story on stage at a software developers' conference, partly because it embarrasses the room. The founder, Patricia, is Irish, was over 60 when she started, and is not a developer. Her product is a B2B app for auto repair shops. The problem she went after is a bay-occupancy problem dressed up as a communication problem. A customer brings a car in, the mechanic inspects it and then has to stand there explaining why it costs what it costs. The customer says that sounds expensive, maybe they will get a second opinion — and meanwhile the car sits in the garage and the shop cannot work on anything else. Auto Pixie moves the explanation into an app: the mechanic sends the diagnosis to the owner's phone, brakes flagged red as compulsory, parts priced, labour priced, with the option to bring your own parts. The owner gets a fast yes or no, and the bay clears. Around that sit two more pieces. Invoicing and payment happen in the app, because a mechanic with oil to their elbows is not writing invoices. And it keeps customers coming back — birthday greetings, reminders for a scheduled check or a tyre change. Before any of it was built she interviewed 106 auto repair shops in five countries, some of the owners five or six times, and built 26 prototypes in PowerPoint. Prototype 26 was the one she sold.
Her customer list and her research list were the same list. She interviewed 106 auto repair shops across five countries, and she did it as a deliberately global exercise rather than an Irish one — she wanted an app that worked for mechanics anywhere, so she went and asked mechanics anywhere.
The quick-reference card — acquisition channels, replication playbook, risks & traps, and our verdict — is members-only.
Join the Unicorn Club — $5/moCancel anytime, in one clickPaste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: Told second-hand from a conference stage: Donatas Jonikas, a marketing consultant who coaches SaaS founders, is recounting the story of a founder he worked with, so every figure is creator-reported and nothing is documented on screen. The 106 interviews, five countries, 26 prototypes, three lifetime licences at 5K, the roughly 15K she held herself and the 30K build estimate all come from his telling. The currency of those figures is never spoken — only the roughly €20,000 government grant is explicitly in euros — so we report the licence price as a bare number. The accelerator is described as one of the world's top US accelerators but is not named, and the half-million-dollar angel investment is described as prepared and conditional on graduation, not as money received. We have not verified that the product shipped or what it earns today.
She priced the build before she priced the product. Her estimate was that it would take 30K to make, and she had roughly 15K of her own, so she needed 15K from the market.
The first loop was funding, not usage. Three paying customers converted into a government grant, the grant into credibility, the credibility into an accelerator place, and the accelerator into a half-million-dollar syndicate offer.
Twenty-six prototypes means twenty-five of them were wrong.
A B2B app for auto repair shops with three jobs, in the order Jonikas lists them. First, faster decisions. The mechanic performs the inspection and the app carries the result to the car owner's phone: what is wrong, which items are compulsory and flagged red, what the spare parts cost, what the labour costs, and whether the owner would rather bring their own parts. The explaining is done by the app rather than by the mechanic, in enough detail that someone who knows nothing about cars understands why an item matters. The shop gets a quick yes or no instead of a car parked in a bay while the owner shops for a second opinion. Second, zero paperwork. Invoices and payments run through the app. The framing is the mechanic's, not the accountant's — someone with their elbows in oil is not filling in invoices. Third, customer loyalty: birthday greetings, reminders that a car is due a check or new tyres, all aimed at bringing the same owner back to the same shop. Our read: the buying decision is the first job and the retention comes from the third. A shop manager will sign up because bays clear faster, which is a revenue number they can feel that week, and will stay because the reminders bring cars back. The invoicing is table stakes and the weakest part of the pitch, because incumbents in shop-management software already do it.
Two prices, deliberately different. The three founding shops bought lifetime licences at 5K each, all future updates and upgrades included. Everyone after them pays a monthly or yearly subscription. That split is the whole pre-sale mechanic: the lifetime licence is expensive enough to fund development and generous enough that an early adopter is being paid, in effect, for taking a risk on a product that did not exist. The second revenue line was non-dilutive and came because of the first: about €20,000 of government grant money, plus a contract for government investment. Public funding bodies respond to paying customers, and three signed licences is a far better application than a deck. Our read: the currency of the 5K, 15K and 30K figures is never spoken on stage — only the €20,000 grant is explicitly in euros — so treat the licence price as a ratio rather than a number. What is copyable is the ratio itself: a lifetime licence priced at roughly one sixth of the total build budget, sold to enough customers to cover the half you cannot fund yourself. Lifetime pricing is a debt you take on, and three of them is a manageable debt; thirty would have been a business model.
The research and validation stack was PowerPoint. Twenty-six clickable prototypes, built by a non-developer, shown to mechanics to demonstrate what happens when you tap a screen. Jonikas notes this was before AI tools existed to do it faster, which is the point of telling the story to a room of developers in 2026 — the bar for demonstrating a product before building it has only dropped since. The build budget was estimated at 30K, of which she held about 15K and pre-sold the other 15K. The estimate was wrong, and Jonikas says plainly that it cost a lot more than that at the time. The €20,000 grant absorbed part of the overrun. Our read: the cost overrun is the most instructive number in the story, because it is the normal case rather than the exception. A first-time non-technical founder underestimating a vertical B2B app with payments, mobile delivery and multi-country requirements by a factor of two or more is entirely predictable, and the reason it was survivable is that the funding sequence kept producing new money. Build the same way today and the stack is a mobile-friendly web app rather than native, with payments delegated to a processor and the inspection report as the only genuinely custom surface. The margin question is not gross margin — vertical SaaS has plenty — it is whether a shop's subscription outruns the support load of teaching non-technical mechanics a new workflow.
Copy the sequence exactly: interviews before prototypes, prototypes before code, a signed pre-sale before a build budget. One hundred and six interviews across five countries and repeat visits to the same owners is a lot of work, and it is the reason prototype 26 sold to three out of three rather than one out of ten. Copy the pre-sale agreement structure Jonikas teaches alongside the story, because it is what makes a stranger comfortable paying for vapour: the customer pays only when you deliver, install and train their people, the technical specification is an appendix to the agreement, and a missed deadline or a missed spec releases them from the obligation. The risk sits with the founder, which is the only reason the signature happens. Copy the three forms of commitment. Money is the strongest, time given to co-creation is second, and a reputation lent through a warm introduction is third. Anyone who offers none of the three is not a customer, however enthusiastic the meeting was. Our read: change the lifetime licence. It worked here because it was capped at three and priced at roughly a sixth of the build. Sell ten of them and you have sold your best customers' future revenue away. The modern equivalent that keeps the upside is a founding-member rate — a permanent percentage discount on a subscription rather than a permanent exemption from one.
Shop-management software is not an empty category. Repair shops already run systems that handle invoicing and scheduling, and the incumbent's answer to Auto Pixie is to add a customer-approval screen. The defensible part is the quality of the explanation sent to the car owner, which is a content and UX problem rather than a technical one, and that gap closes if the founder stops obsessing over it. The cost overrun is the second risk, and it already materialised: the build cost considerably more than the 30K she budgeted, against three fixed-price lifetime licences that could not be repriced. A pre-sale locks your revenue while your costs stay variable. Adoption is the third. The buyer is a small owner-operated business whose staff are mid-repair when they are meant to be using your software. If the inspection flow adds a minute to the mechanic's day rather than removing five, the app goes unused and the subscription lapses. Our read: the funding path carries its own risk. A half-million-dollar syndicate conditional on graduating a US accelerator points this company toward venture-scale expectations, while the underlying business — a subscription sold shop by shop to garages across five countries — grows at the speed of a field sales process. That mismatch is how a healthy vertical SaaS ends up judged as a failed rocket.
The four founder playbooks — first customers, first dollar, the flywheel, the silent stretch — are in the Unicorn Club, on this idea and on all 1,435 others. The story, the numbers, the sources and every later section stay open to everyone.
Join the Unicorn Club — $5/moCancel anytime, in one click