How a user-problem statement made us restart an already 'finished' product

Hey there,

I've recently posted in the past about an AI product I've been working on (https://www.trimmr.ai).

The silence and lack of activity is because I realised something about the product that we made which I'm sure a lot other AI product builders felt: the idea is overdone in an already saturated market.

I want to explain the process the team and I went through without peddling the product too much, so if it is a bit excessive I apologise in advance! I tried writing it without the product, but the context helps in explaining how a user-problem statement helped us create a new feature which separates us from the crowd.

Here is a TL;DR or it without too much product peddling:

1. How to identify a user problem statement: Take any email that comes through with respect, and write it down somewhere as it'll come in handy one day. If they're emailing you, it's a good sign.
2. Be aware of the silo: Thinking independently is incredibly important, but it's easy to fall into group think which can be dangerous and limit the products potential
3. Don't be afraid to split from the crowd: Major players in the market had so much purchasing power over us that we couldn't be a cheaper, better (maybe biased lol) offering as our search was being brought by them (you can see the exact day they started advertising our term on Google Analytics).
4. Bootstrapping can give you more freedom: I often felt inadequate that these companies would post their successes while we were just starting out, but being bootstrapped meant we have no obligations. If we fail, we're going to fail spectacularly.

Deep dive

I love trimmr.ai and all the hardwork that's gone into it. But with competitors like Opus Clip and Get Munch receiving venture backing it felt like we were banging our heads into rusty nails and being bullied in a competitive, but quickly evolving space. The rest of the founders and I were close to giving up, relying solely on Google Search to generate paying customers (which to be fair was going really well).

We had to take a breather because it wasn't easy. We all stepped away for a bit, until one day a stray Jira ticket came through. It simply said "how can i automate this to social media". There were probably 20 spam tickets but this one stood out to me. I asked "how can you automate this to social media?" or better yet "how can we automate this?".

People under estimate the effect of being siloed. You're in your own mind, closed off to possibilities. Our team really is amazing, but we truly thought we had a finished product and were just working on minor changes and product refinement.

We took the time to actually map out what we could use, but I realised I had previous work experience with Zapier to automate workplace actions.

From there, we had discussions on what the ideal flow could look like (I pressed hard on the fact that it had to be purely autonomous) and we created a flow that fulfilled this use case:

1. New or updated spreadsheet row in Google Sheets
2. Import YouTube Video in Trimmr.ai
3. Generate Short Idea in Trimmr.ai
4. Render Short in Trimmr.ai
5. Save video in Google Drive
6. Send email to user

We thought "if we can create this flow, we can entirely automate short-form content creation". From here there were tonnes of use cases and made us realise that we had now created an entirely unique offering that the big players had missed. People can:

• Upload their YouTube videos and have short ideas delivered directly to their phone after posting
• Create faceless content accounts and validate their potential, before going all in by creating a spreadsheet with YouTube links and posting them autonomously to Instagram
• Trim company content and create bite-sized videos for employees to watch, sent straight to a Slack channel
• Combine Trimmr.ai with other AI-products to create a fully-fledged automation machine

It wasn't easy, and was a hard pill to swallow at the start. We thought we had finished and the idea had been done. But we are determined to make this work.