About Mehrik

We like the conversations that start with “this is probably a stupid question.” They almost never are.

Mehrik works with businesses that want to use AI properly, rather than just be seen to be using it. Most of the people we talk to aren't sure whether any of it applies to them. That's a good place to start from — better, usually, than being certain.

Behind that is twenty years of building software — systems, databases, APIs — and, alongside it, twenty years of running the commercial side of online retail. Seller accounts on Amazon, eBay, Walmart and the smaller marketplaces. Inventory systems, order flows, listings, shipping integrations, automating warehouse operations. The unglamorous machinery that a business actually runs on.

That second half matters more than it sounds. It's the difference between knowing how to build something and knowing what it costs a business when the wrong thing gets built.

Why we started

Everyone is being told to use AI. Almost nobody is being told what for.

Most small businesses we meet are somewhere in the same place: they know AI is important, they've been told they're behind, and they've opened a chatbot and asked it a few questions. That's usually where it stops. People think they're using AI, when what they're really doing is using a chatbot as a faster Google.

There's nothing wrong with that. But it isn't the thing. The useful version of AI for a small business is almost never a chat window — it's something quietly handling the work that currently sits on one person's shoulders, or the twenty minutes a day that goes into retyping information from one place into another.

Nobody was having that conversation with them. The market was busy selling AI to companies large enough to buy it, and the advice reaching everyone else was either generic or trying to sell something.

So that's the gap we started in. Not adding AI because it's available — finding whether there's a real use for it in your business, and being honest when there isn't.

How we think about the work

Most AI advice starts from the technology and works backwards to find somewhere to put it. We start from the opposite end — what your business actually does every day, and whether any of it is worth changing.

That means a lot of our answers are no.

We've looked at work that turned out to be a set of fixed rules — clear conditions, predictable outcomes, the same decision every time. That doesn't need AI. It needs the rules written down properly once. Putting a language model in the middle of it would have cost money and made it less reliable, not more.

We've also done the arithmetic and found the technology costs more than the problem. Around $500 a month in API charges to do work a person was doing for $100. The technology worked. It just wasn't worth it.

Neither of those is a comfortable conversation. Both are the reason people come back.

And plenty of the time, the thing you need already exists — a product on the market, or software you're already paying for with half its features switched off. When that's the case, we'll tell you, and we'd rather help you set it up properly than build our own version of it.

We're not trying to become the company that sold you AI. We're trying to be the people you call when you're not sure whether something is worth doing.

What we believe

Technology is supposed to be an enabler. Somewhere along the way a lot of it became a burden instead — another system to maintain, another login, another thing that only works if everyone remembers to use it correctly.

It should be the other way round. The tool adapts to how you already work. You don't reorganise your business around the tool.

AI, simplified. Not made smaller or less capable — just stripped of the parts that were never for you in the first place. The jargon, the hype, the pressure to have an opinion about it. What's left is a question anyone can answer: does this make the work easier, or does it just make it messier?

Easier, not messier.

That's the whole test, and we apply it to our own work as much as anyone's.

What that looks like in practice

We start small. Not because it's cautious, but because it's the only honest way to find out if something works. The smallest useful version, in your hands, doing real work — that tells you more than any proposal.

We explain everything in your language. If an explanation needs a technical background to follow, we've explained it badly. That's on us, not you.

We won't guess at timelines to win work. Some things are quick, some genuinely aren't, and we don't know which until we've looked properly. You'll get an honest answer once we understand the problem — not a date invented to sound impressive.

Your data stays yours. We don't reuse it, we don't sell it, and we don't train anything on it. You also choose how it's handled — models running on infrastructure you control, or a commercial service. Both are legitimate. We explain the trade-off and you decide.

What we won't do

That last one costs us work. It's also the reason people come back.