Good marketing shouldn't be a luxury. Why I built SMARTYPANT AI.

The idea for SMARTYPANT AI didn't come from a business plan. It came from a conversation with a friend who runs a small local business.

She was doing everything right — good product, fair prices, customers who liked her. But almost nobody knew she existed. When I asked about marketing, she said what I'd hear repeatedly over the following two years: "I don't have time for that, and I can't afford someone who does."

I'd heard versions of it before. I'd said it myself. But something about hearing it that plainly made it stick.

I built something advanced. Nobody knew it existed.

Before SMARTYPANT AI I built dudu AI — a conversational assistant that, for its time, could do a remarkable amount. Deep search and live web search. Price comparison across the web. End-to-end shopping inside the conversation, so you could describe what you wanted, compare options and buy without leaving the chat. And customisable agents that could be pointed at real work: screening candidate CVs, marking exam papers against a rubric, whatever a user needed built.

This was genuinely ahead of most of what was available. I know because I built it, and I know what it took.

I had no marketing team. I had no distribution. I couldn't convince enough people to even try it. Some did — and some of them used it for exactly what it was built for, including the shopping. But "some people" isn't a business.

So I did what a technical founder does when faced with a marketing problem: I treated it as an optimisation problem. I read about SEO. I wrote content. I studied social media strategy. I tried at least twenty different tools and platforms.

I did rank number one on Google. For "AI dog leash." Which, it turns out, is a phrase almost nobody types into a search engine.

Zero customers from that ranking. Zero.

dudu AI is closed now. Not because the product failed — because I never solved the part where people find out you exist. That distinction has stayed with me, because it's the difference between "I built the wrong thing" and "I built the right thing and lost anyway."

Then I watched it happen to everyone else

After that I spent a couple of years supplying eco-friendly packaging to restaurants and cafés across Germany. Hundreds of visits — family pizzerias, urban cafés, catering kitchens. Real businesses, real customers, real margins.

What I mostly did, though, was sit across tables from other small business owners.

Nearly every one of them had an Instagram account. Nearly none of them posted. When I asked why, the answers were consistent enough to be predictable: no time, no idea what to post, tried it for a month and stopped.

These weren't people who didn't understand marketing. They were on Instagram personally. They could tell me which competitor did it well and why. They knew exactly what they should be doing.

They were just out of hours, and content production loses to dinner service every single time.

That's when it stopped being my problem and became the problem. I'd lived one version of it — advanced product, no distribution. They were living another — good business, no visibility. Same failure, different shape.

The expensive truth

Here's what nobody tells you when you start a small business: marketing is expensive, and it's expensive in a way that scales badly for small operators.

A social media manager has a median total pay of roughly €45,000 a year in Germany and around $71,000 in the US, per Glassdoor. In Germany, statutory employer contributions add roughly another fifth on top — before equipment, recruitment, or any of the other costs of employing someone. Agencies charge accordingly. Freelancers are cheaper but still meaningful money for a business turning over a few hundred thousand.

For a large company those are line items. For a café, a salon, a small manufacturer, they're the difference between hiring a second person and not.

So the small business does it themselves. And doing it themselves means doing it badly, not through incompetence but through arithmetic: content creation competes with running the business, and running the business wins every time. So the Instagram account goes quiet in March and gets remembered in October.

Which produces the outcome I find genuinely unfair. How well a business markets itself often says more about its available time and budget than about the business. The best pizza in a neighbourhood loses to the one that can afford to be seen. The most careful supplier loses to the loudest.

What I built

I'm an engineer with a doctorate and more than a decade working on AI systems. I kept looking at this problem — mine and theirs — and thinking: this is repetitive, pattern-based, and constantly requires the same categories of decision. That's the shape of a problem machines are good at.

Not "helping" with marketing. Not another tool for a queue of tools — I'd already tried twenty of those. Actually doing the work.

SMARTYPANT AI reads your website, understands your products and voice, creates content — images, video, platform-specific copy — decides when to post, publishes across your connected channels, and learns from what happens. It also audits your website and fixes what's holding back your visibility in search and in AI answers.

The distinction I care about is between a tool and delivered work. A tool gives you a capability and leaves you to use it, which is precisely what failed me at dudu. What I wanted was the recurring execution you'd normally hire an agency to perform — at a price a café can pay.

Social Autopilot covers the content side and starts at $149 a month. The website and AI-visibility work sits in Growth Autopilot. Both are on the pricing page, and both are deliberately, structurally cheaper than a person — not because the work is worth less, but because the whole point was to put it within reach of businesses that were priced out.

My own packaging business ran on it. Not as a favour — because I needed it, and because I wasn't going to sell something I wouldn't run my own company on. Those numbers are published, including the ones that aren't flattering.

What I've learned since

Two years of building this and talking to owners has taught me things I didn't expect.

The barrier is almost never understanding. Every owner I've met knows what good social media looks like — they're on Instagram themselves, they can tell you which competitor does it well. They aren't confused. They're out of hours.

Consistency beats brilliance. The businesses that grow aren't the ones with the cleverest campaigns. They're the ones that show up every week for a year. Humans are bad at sustaining that — not through laziness, but because something urgent always arrives. Software holds a cadence better, as long as somebody notices when it breaks.

Distribution is harder than marketing. This is the one I'm still working on, and the one dudu died of. Making good content is a mostly solved problem now. Getting it in front of the right people — across search, social, AI answers, marketplaces — is fragmented, changes constantly, and punishes anyone without a full-time specialist. Every part of it I can automate is a part a small business no longer has to lose at.

And it works for businesses that shouldn't work on social. A pizzeria near Stuttgart went from a 700-click weekly baseline to over 2,800 during an eighteen-day pilot — around 1,000 of which he'd already reached through his own manual posting. A B2B packaging supplier in Berlin revived an account dormant for 72 weeks: 21 posts and 4,000 views in three weeks. That company is now named by ChatGPT for its category, though I won't claim Instagram alone caused it — several things changed in that period. Neither is a business anyone would call Instagram-friendly.

What I'm actually trying to fix

I want the small business owner who makes something genuinely good to be able to compete on that, rather than on marketing budget.

Not because marketing shouldn't matter — it should, it's how people find you. But because the current arrangement means the ability to afford marketing decides who gets found, and that has nothing to do with who deserves to.

That's the whole thesis. Everything I build is downstream of it.

Being honest about where this is

This is a work in progress and I'd rather say so.

The social side is live and running for paying customers. The website audit and fixes are live. AI visibility tracking is live. Market monitoring and adaptation are being built. Some channels aren't supported yet — TikTok and Facebook among them.

Nothing here guarantees results. What it does is remove the reason most small businesses never try: the production burden. Whether that turns into customers depends on your product, your market, and a hundred things software doesn't control.

But you'll be visible while you find out, which is more than most small businesses manage.

If this sounds like you

If you're running something good and losing to businesses that market better, I built this for you. If you've abandoned three social media accounts because life got busy, I built this for you.

And if you've ever ranked first on Google for something nobody searches — I've been there. It's not a skill problem. It's a time problem, and time problems have engineering solutions.

You can try it free: two sample posts and a full visibility audit, no card required. If it's useful, keep going. If not, you've lost fifteen minutes.