Back
Roast
Posted

Roast my idea

I am building LlamaSQL: "AI-powered chats with your database". Basically, you can analyse your data using natural language. You put in a natural language query, it will first generate the SQL query for it based on the database schema, validate & auto-correct it and execute it to get the results. And finally, a graph will be generated for the results.

The above is just the MVP, and I have been doing evaluations to get each step of the process perfect, which includes query rewriting, related tables inference, intent detection and more.

Users will be able to share these chats either publically or with teammates. They will be able to build a dashboard and pin the charts there. They will be able to manually edit the SQL query as well.

Shareability and dashboard along with integration like Slack will allow this to be used by anyone and save a lot of time for them.

This is a very crowded space and products like Metabase & Amplitude and add Natural Language to SQL very quickly, if not already. 

Please share your honest feedback! 

WhatsApp Image.jpeg 95 KB


It's a good idea but many startups already do this (like a lot)

I would figure out at least one thing that makes you a bit different for your goto market

I understood maybe two words of this, but I had ChatGPT explain it to me like I'm 5. I think I understand it a little better (maybe), and if I do, then I think it's a pretty neat idea! I didn't even know something like this already existed.

Do you have a defined market for this?

If so, I bet you could get even more specific with features to serve that specific market's needs in a way that similar products that went broad wouldn't be able to do during the first iteration then expand to other markets as you gain traction. Bonus: Your copy would be insanely specific, so any marketing campaign you do would be that much more effective.

Like for example, you could target the education niche then get more specific with online homeschooling companies so they can track things like student progress and churn rate, then you could expand into other micro-niches within the education market.

I'm writing content for a homeschooling education company right now, so that's why that came to mind first lol. Any underserved market would work, though, especially if you can create hyper-specific solutions for them that don't yet exist in other products. (I'm acting as if I know education is an underserved market for this. For all I know, it's the most served one 🤣)

Thanks for you feedback @cat . I figured that I still do not have enough clarity on the GTM, so will pause my efforts for now. Building the POC anyways got me to a good start point of picking up any different idea now.

I am exploring few more ideas.