The Launch Desk

Why I'm Building in Public

A former Reuters correspondent and B2B marketer starts vibe coding two AI products and writing about every step. Here is the plan, the reasoning, and what readers of The Launch Desk will get.

I am learning to build software by talking to it, and I am going to show you every step.

For seven years I was a correspondent at Thomson Reuters. I interviewed CEOs, filed on deadline, and learned to write a first line a trader would stop for. For the five years after that, I ran brand and content for companies in fintech, AI, analytics, and payments. I grew a newsletter from zero to 73,000 subscribers. I took a category term nobody had heard of to No. 1 on Google in two countries. I tied content to more than $1.5 million in pipeline.

I have spent twelve years telling other people's launch stories. This is the first time I am telling my own, and I am going to tell it as it happens.

What changed

Until recently, a marketer with a product idea had two options: learn to code, or raise money to pay someone who could. Most of us did neither, and the idea stayed a slide.

AI coding tools changed the arithmetic. I can now describe a product in plain English and watch a working version appear in an afternoon. The code still needs judgment, testing, and patience. The barrier to a first version, the one that kept people like me on the sidelines, has collapsed.

That shift rewards something experienced practitioners have in surplus: context. I know which problems marketing teams lose sleep over because I have lost the sleep myself. Domain experience tells you which product to build. AI tools now let you build it yourself.

What I am building

A GEO scanner. It runs the questions buyers ask through ChatGPT, Perplexity, Gemini, and Google's AI answers, then tracks how often a brand shows up and what the machines say about it. I have been doing this by hand every month for the brands I work on. It is slow. A tool should do it.

An app that helps people stop doom scrolling. This one is personal. I build audiences for a living, so I know exactly how feeds are designed to hold attention. I want to build something that hands a little of that attention back.

Why in public

Three reasons.

Accountability. Writing a weekly update makes it much harder to quietly abandon a project in month two. You will know if I stall. So will I.

Credibility. I write about AI launches. The most convincing proof I can offer is a launch of my own, with the numbers visible, including the ugly ones.

Distribution. I believe the AI products that win from here will be the ones with an audience waiting for them. I would be a hypocrite to build mine any other way.

What you will get from The Launch Desk

One essay a week. Some will be playbooks from my own work: category creation, launch narratives, AEO and GEO, newsletters. Some will be build logs: what I shipped, what broke, what it cost, and what the first users said. All of them will follow the newsroom rule I still live by: lead with the news, cut what is not working, and check the facts twice.

If you are a founder launching an AI product, a marketer figuring out where you fit in this new world, or a professional wondering whether you can build something of your own, this is for you. It is early. Pull up a chair.

Stay on deadline,

Derek

Frequently asked questions

What does building in public mean?

Building in public means sharing the progress of a product or business openly as it happens, including decisions, metrics, and mistakes. Founders use it to earn trust, gather feedback, and grow an audience that becomes the first wave of users and customers.

Can a non-engineer build an AI product?

Yes. Vibe coding tools such as Claude Code, Lovable, Bolt, and Replit let people describe software in plain English and generate working applications. Non-engineers can build and validate a first version, then bring in engineers to harden it once users respond.

What is Derek Francis building?

Derek is building two products in public: a GEO scanner that measures how visible a brand is inside AI answer engines, and a mobile app that helps people stop doom scrolling.