We made $325k in 2023 from AI products, starting from 0, with no-code, no funding and no audience

I met my co-founder in late 2022 after an introduction from a mutual friend to talk about how to find contract Product Management roles.

I was sporadically contracting at start-up at the time and he had just come out of another start-up that was wiped out by the pandemic.

We hit it off, talking about ideas, sharing what other indie-hackers were doing, and given GPT-3’s prominence at the time, we started throwing around ideas about things we could build with it, if nothing else, just to learn.

I should caveat, neither of us were AI experts when starting out, everything we learned has been through Twitter and blogs, my background is as an accountant, and his a consultant.

Here’s how it went since then:

​

Nov 2022 (+$50)

- We built a simple tool in around a week using GPT-3 fine-tuning and a no-code tool (Bubble) that helped UK university students write their personal statements for their applications

- We set some Google Ads going and managed to make a few sales (~$50) in the first week

- OpenAI were still approving applications at the time and said this went against their “ethics” so we had to take it down

​

Dec 2022 (+$200)

- We couldn’t stop coming up with ideas related to AI fine-tuning, but realised it was almost impossible to decide which to pursue

- We needed a deadline to force us so we signed up for the Ben’s Bites hackathon in late December

- In a week, we built and launched a no-code fine-tuning platform, allowing people to create fine-tuned models by dragging and dropping an Excel file onto it

- We launched it on Product Hunt, having no idea how to price it, and somehow managed to get ~2,000 visitors on the site and make 2 sales at $99

​

Jan 2023 (+$3,000)

- We doubled down on the fine-tuning idea and managed to get up to ~$300 MRR, plus a bunch of one-time sales and a few paid calls to help people get the most out of their models

- We quickly realised that people didn’t want to curate models themselves, they just wanted to dump data and get magic out

- That was when we saw people building “Talk with x book/podcast” on Twitter as side projects and realised that was the missing piece, we needed to turn it into a tool

- We started working on the new product in late January

​

Feb 2023 (+$9,000)

- We started pre-selling access to an MVP for the new product, which allowed people to “chat with their data/content”, we got $5,000 in pre-sales, more than we made from the previous product in total

- By mid-February, after 3 weeks of building we were able to launch and immediately managed to get traction, getting to $1k MRR in < 1 week, building on the hype of ChatGPT and AI (we were very lucky here)

​

Mar - Jul 2023 (+$98,000)

- We worked all the waking hours to keep up with customer demand, bugs, OpenAI issues

- We built integrations for a bunch of services like Slack, Teams, Wordpress etc, added tons of new functionality and continue talking to customers every day

- We managed to grow to $17k MRR (just about enough to cover our living expenses and costs in London) through building in public on Twitter, newsletters and AI directories (and a million other little things)

- We sold our fine-tuning platform for ~$20k and our university project for ~$3k on Acquire

​

Aug 2023 (+$100,000)

- We did some custom development work based on our own product for a customer that proved pretty lucrative

​

Sep - Oct 2023 (+$62,000)

- After 8 months of building constantly, we started digging more seriously into our usage and saw subscriptions plateauing

- We talked to and analysed all our paying users to identify the main use cases and found 75% were for SaaS customer support

- We took the leap to completely rebuild a version of our product around this use case, our biggest to date (especially given most features with no-code took us <1 day)

​

Nov - Dec 2023 (+$53,000)

- We picked up some small custom development work that utilised our own tech

- We’re sitting at around $22k MRR now with a few bigger clients signed up and coming soon

- After 2 months of building and talking to users, we managed to finish our “v2” of our product, focussed squarely on SaaS customer support and launched it today.

​

We have no idea what the response will be to this new version, but we’re pretty happy with it, but couldn’t have planned anything that happened to us in 2023 so who knows what will come of 2024, we just know that we are going to be learning a ton more.

​

Overall, it is probably the most I have had to think in my life - other jobs you can zone out from time to time or rely on someone else if you aren’t feeling it - not when you are doing this, case and point, I am writing this with a banging head-cold right now, but wanted to get this done.

A few more things we have learned along the way - context switching is unreal, as is keeping up with, learning and reacting to AI. There isn’t a moment of the day I am not thinking about what we do next.

But while in some way we now have hundreds of bosses (our customers) I still haven’t felt this free and can’t imagine ever going back to work for someone else.

Next year we’re really hoping to figure out some repeatable distribution channels and personally, I want to get a lot better at creating content/writing, this is a first step!

Hope this helps someone else reading this to just try starting something and see what happens.