I’ve built https://ai2sql.io/ , an AI-powered tool that converts natural language into SQL queries, which has reached $9K MRR. We use GPT-4 for the core functionality and launched by focusing on a niche: automating SQL generation for non-technical users. Our revenue model is subscription-based with tiered pricing for different levels of query complexity and usage.
Challenges included refining the accuracy of generated queries and handling edge cases for diverse databases. My advice: start small with a specific problem LLMs can solve, iterate based on user feedback, and focus on delivering tangible value in your niche.
To promote AI2sql and reach the target audience, I used a combination of:
1. Building in Public: Shared the product development journey, updates, and milestones on https://x.com/mustafaergisi to build transparency and engage early supporters.
2. Community Engagement: Actively participated in relevant online communities to provide value and gain organic exposure.
3. Content Marketing: Created helpful, relevant content to attract and engage the audience.
4. Networking: Leveraged personal and professional connections to spread the word.
Consistent interaction and showcasing progress helped build trust and attract the right audience.
LLM costs are indeed a significant part of our operating expenses. While I can't share exact percentages, I can say that they constitute a meaningful portion of our revenue. We continually optimize our usage and explore cost-effective strategies to manage these expenses while maintaining the quality of our service.
New features for an existing product. Most of them were switched on globally when we launched them.
Which LLM(s) are you using (e.g. GPT-4, Claude, open source models)?
We started using language models before ChatGPT and friends came out. We built our models from scratch, and still use and update them.
What's your revenue model?
Saas
How much are you making (if you want to share, ballpark figures are fine)?
We don't charge extra for features that use ML, it's an implementation detail that our users don't care about. So, I am not sure what % of the whole they contribute to. Its not the majority.
What challenges have you faced?
"mlops" was a real challenge in the early days, and we built so many bespoke distributed systems to figure it out. Eventually we did, but finding the right balance between building a performant system and a system that doesn't explode in cost was tough. There is off the shelf things that do this these days, but they can be expensive.
Any advice for others looking to enter this space?
Solve problems customers are willing to pay for. Some of the features we have dreamed up have been total flops. Many of the features we have developed (ml or otherwise), while the add value for the customer and add to the moat, they aren't valuable enough that our users are willing to pay extra for it as a standalone product.
I don't have any books specific to MLOps, just because they weren't out when I was building that system. All of the good practices from building resilient distributed systems apply. Designing Data-Intensive Applications is always a great read.
Some things that have were notable:
Model pipelines tend to be flakier than other pipelines you have. They are much more complicated, and it can be easy to hit a resource limit if you aren't careful, or have a unhandled exception accidentally kill a pipeline 10hrs into it.
Avoiding those outright is obviously the best path, but that can be easier said than done.
One thing that we found really helpful was creating an error record in a database for every piece of data that failed to get processed, where it failed in the pipeline, etc. Retries, and alters were easy to tack on after that.
I’ve been using an AI chatbot powered by GPT-4 to manage customer service tasks for my small business. The chatbot handles basic queries, FAQ responses, and routing, and it’s been a great time-saver. The service operates on a freemium model, where the base features are free, and you can pay for more advanced capabilities like analytics and integrations.
One challenge I've noticed is ensuring it works seamlessly across different business needs while staying easy to use. For those exploring LLMs, starting with a targeted niche and iterating based on real user feedback has been essential in maximizing the tool’s efficiency.
I’ve built https://ai2sql.io/ , an AI-powered tool that converts natural language into SQL queries, which has reached $9K MRR. We use GPT-4 for the core functionality and launched by focusing on a niche: automating SQL generation for non-technical users. Our revenue model is subscription-based with tiered pricing for different levels of query complexity and usage.
Challenges included refining the accuracy of generated queries and handling edge cases for diverse databases. My advice: start small with a specific problem LLMs can solve, iterate based on user feedback, and focus on delivering tangible value in your niche.
Hey love your site and it looks amazing.
I’m genuinely curious to ask this question.
Why would someone use your site compared to just using claude or chat gpt
How did you promote the product to reach your target audience?
To promote AI2sql and reach the target audience, I used a combination of:
1. Building in Public: Shared the product development journey, updates, and milestones on https://x.com/mustafaergisi to build transparency and engage early supporters.
2. Community Engagement: Actively participated in relevant online communities to provide value and gain organic exposure.
3. Content Marketing: Created helpful, relevant content to attract and engage the audience.
4. Networking: Leveraged personal and professional connections to spread the word.
Consistent interaction and showcasing progress helped build trust and attract the right audience.
What % of your reveneue goes towards LLM costs?
LLM costs are indeed a significant part of our operating expenses. While I can't share exact percentages, I can say that they constitute a meaningful portion of our revenue. We continually optimize our usage and explore cost-effective strategies to manage these expenses while maintaining the quality of our service.
Your site and offering looks great. Thank you for the feedback!
Thank you for the kind words! I'm glad you liked the site and offering. Let me know if you have any questions or need more information—happy to help!
What does your product/service do?
Ad tech
How did you develop and launch it?
New features for an existing product. Most of them were switched on globally when we launched them.
Which LLM(s) are you using (e.g. GPT-4, Claude, open source models)?
We started using language models before ChatGPT and friends came out. We built our models from scratch, and still use and update them.
What's your revenue model?
Saas
How much are you making (if you want to share, ballpark figures are fine)?
We don't charge extra for features that use ML, it's an implementation detail that our users don't care about. So, I am not sure what % of the whole they contribute to. Its not the majority.
What challenges have you faced?
"mlops" was a real challenge in the early days, and we built so many bespoke distributed systems to figure it out. Eventually we did, but finding the right balance between building a performant system and a system that doesn't explode in cost was tough. There is off the shelf things that do this these days, but they can be expensive.
Any advice for others looking to enter this space?
Solve problems customers are willing to pay for. Some of the features we have dreamed up have been total flops. Many of the features we have developed (ml or otherwise), while the add value for the customer and add to the moat, they aren't valuable enough that our users are willing to pay extra for it as a standalone product.
Thank you for providing a detailed response. It's helpful!
> There is off the shelf things that do this these days, but they can be expensive.
Could you share some names for someone inexperienced in MLOps to do some research?
https://www.kubeflow.org/
Kubeflow is the one I see used the most (self hosted). The public clouds also have products targeting these workloads.
Would love to read an MLOps lessons learned or approach you had or if you recommend any particular books that really hit the nail on the head
I don't have any books specific to MLOps, just because they weren't out when I was building that system. All of the good practices from building resilient distributed systems apply. Designing Data-Intensive Applications is always a great read.
Some things that have were notable:
Model pipelines tend to be flakier than other pipelines you have. They are much more complicated, and it can be easy to hit a resource limit if you aren't careful, or have a unhandled exception accidentally kill a pipeline 10hrs into it.
Avoiding those outright is obviously the best path, but that can be easier said than done.
One thing that we found really helpful was creating an error record in a database for every piece of data that failed to get processed, where it failed in the pipeline, etc. Retries, and alters were easy to tack on after that.
presumably someone is making money by spamming HN with links to substack articles that are LLM-generated ads
I’ve been using an AI chatbot powered by GPT-4 to manage customer service tasks for my small business. The chatbot handles basic queries, FAQ responses, and routing, and it’s been a great time-saver. The service operates on a freemium model, where the base features are free, and you can pay for more advanced capabilities like analytics and integrations.
One challenge I've noticed is ensuring it works seamlessly across different business needs while staying easy to use. For those exploring LLMs, starting with a targeted niche and iterating based on real user feedback has been essential in maximizing the tool’s efficiency.
Thank you AI.
[dead]
[dead]