About: The Software Was Never the Hard Part

We started Cyber Spark in 2023. We built software for people. Someone came with an idea, we built it, they paid, we moved on to the next one. I was proud of that work.

Here is what I never did. I never checked what happened next.

Building got easy

Last year, building software got easy. Really easy. A person with a laptop and a ChatGPT subscription could get a working app in two weeks. The thing we sold was not special any more.

That scared me for about a month. Then it gave me something I had never had: time to think.

Where the businesses went

So I did the thing I had never done. I went back and looked at the businesses we had built for. Where were they now?

Most were gone.

Some died while we were still building. Money ran out, the idea changed, the second version never came. I expected that. But the ones we finished, the ones that actually launched, died in a way I did not expect.

They died quietly. The app worked. The website was live. And nobody came. Nobody searched for them. Nobody found them. I typed the name of one of them into Google and got nothing. Not page two. Nothing.

That was the moment for me. The founder had spent everything on the build. They thought the build was the hard part. So did I. We were both wrong. The hard part was being found.

Two kinds of people

So I asked: who fixes that? I found two kinds of people.

SEO agencies. They have patience. They will write content for months, fix profiles, chase links. But most of them are not engineers. When a page renders wrong, or the structured data quietly says the wrong thing, they can see something is off but they cannot tell you why. The good ones, who can, cost a lot.

And developers. Us. We know exactly why the page renders wrong. But SEO is slow, it is boring, and you cannot tell if it worked for months. No developer I know wants to do it. I did not want to do it.

One kind of person has the patience. The other has the knowledge. A business needs both, and usually pays twice or gets half.

Engineers get bored

So we tried something. We took our own engineers and put them on search. Not marketing people. Engineers.

They hated it at first. Then the thing happened that always happens when engineers get bored. They saw the same problem for the third time and built a tool so they would never have to look at it by hand again.

The crawler blocked by a template nobody read. The phone number on Google that did not match the website. The structured data that passed every validator and described the wrong company. Each one became a check. The checks piled up. At some point it stopped being a script and became a product.

Then the question changed

People stopped only asking Google. They started asking ChatGPT “who is a good accountant near me”. And the answer did not include our clients.

Being found now means being found by an AI too. That is why the check is called what it is called.

The deal

We gave the thing a name and a home: this site. A brand new domain. No links, no history, nothing. On purpose.

Here is the deal I am making, in public. We will use only the methods we tell clients to use, on this domain. Every month I will post the numbers on the blog. Visits, links, how many people ran a check. When a month goes badly, I will say so. Most of our client work is under NDA, so I cannot show you that. This one, I can.

If it works, you can watch it work before you pay anyone. If it does not, you will know that too.

What we sell, plainly

The check is free, and it stays free. It needs no account, only an email address to send the results to. If it finds things you want to fix yourself, the report tells you what and where.

If you want a deeper look, a full check (the Deep Scan) is also free — one each month with a free account, then £9.99 for another, or Pro at £14.99 a month for 15. It crawls more pages, renders them like a browser, and pulls in the outside sources the essential scan skips.

If you want us to do the fixing, that is the work we do now. Search engineering for businesses that have already built the thing. We will tell you honestly if you do not need us. Plenty of sites do not.

The small print

I am Muntazir, SEO and AI search consultant for small businesses. I get you found on Google, ChatGPT and local search — technical SEO, local SEO, GEO. Based in London.

If your business is good but hard to find on Google or in ChatGPT, that’s the problem I fix.

I’m a software engineer who works on search. I help small businesses, founders and agencies find out why they aren’t showing up, then fix it in the site itself.

For years I built websites and apps for clients. The software worked. What I never checked was what happened after launch. When I finally went back, a lot of those businesses had gone quiet. Not because the product was bad, but because nobody could find them.

That problem is getting bigger. People still search, but more of them now ask ChatGPT, Perplexity or Gemini, and the answer names someone else.

What I work on:

• Technical SEO: crawling, indexing, rendering, speed
• Structured data that describes the right business
• Local presence: Google Business Profile and consistent listings
• AI search visibility (GEO): whether AI assistants mention you, and why not
• Fixing it in the code, not just writing a report about it

I come at SEO as an engineer rather than a marketer. When a page renders wrong or a template blocks a crawler, I can tell you why and fix it.

Those checks turned into a product: AI Visibility Check (aivisibilitycheck.io), a free scan of how search engines and AI assistants see your site.

Based in London, working with businesses across the UK and remotely.

Want to know where you stand? Run the free check at aivisibilitycheck.io, or send me a message. If it turns out you don’t need me, I’ll tell you.

Cyber Spark is the trading name of Cyber Spark IT Services LTD, 71-75 Shelton Street, Covent Garden, London, WC2H 9JQ, company number 14891489. Email info@cyberspark.uk and a person replies. If you want to know exactly how the check scores a site, the maths is on the methodology page.