I’m looking for some guidance on our next steps for an AI-powered workflow we’ve developed. We’ve been using a basic setup to handle customer inquiries in a traditional industry. It’s been working great with our test group, but now we want to scale up.
Our main goals are:
Move away from our current platform
Build our own system
Start selling to more clients
Add new features easily
We’re debating between these options:
Use a low-code tool like Flowise or n8n
Build a custom backend with Node.js or Python
Some other approach we haven’t thought of yet
We like the idea of low-code tools so our non-tech team can help with prompts. But we’re not sure how well they’d handle real traffic.
Has anyone used Flowise or n8n for production? Any good or bad experiences? If you switched from a low-code to a custom solution, did you keep any parts of the old system?
I’d really appreciate hearing about what setups are working well, especially for managing state and integrating with email and calendars.
hey there, i’ve used flowise for a small project and it was pretty good for quick setup. but when we tried scaling, it got kinda sluggish. we ended up building our own backend with node.js which gave us more control.
if ur non-tech team needs to tweak stuff, maybe keep some flowise components for them? just make sure ur core system can handle the load. good luck with ur project!
Having worked on similar projects, I can offer some insights. While low-code platforms like Flowise or n8n are excellent for rapid prototyping and involving non-technical team members, they often struggle with scalability and complex customizations.
For your goals, I’d recommend a hybrid approach. Start by building a custom backend with Node.js or Python. This gives you the flexibility and control needed for scaling and adding new features. However, don’t discard low-code tools entirely.
Use your custom backend for core functionalities and heavy lifting, but integrate a low-code platform for specific workflows where quick iterations are beneficial. This way, your non-tech team can still contribute to certain aspects without compromising on performance or scalability.
Remember, the key is to design your architecture with future growth in mind. Ensure your custom solution has well-defined APIs that can interface with various tools and services as your needs evolve.
I’ve been down this road before, and I can tell you from experience that the choice between low-code platforms and custom builds isn’t always straightforward. We initially started with a low-code solution (n8n in our case) for rapid prototyping and getting our non-technical team involved. It was great for quick iterations and testing ideas.
However, as we scaled, we hit some limitations. Performance became an issue with higher traffic, and customization for complex workflows became cumbersome. We ended up transitioning to a custom Node.js backend, which gave us more control and better performance.
That said, we didn’t completely abandon the low-code approach. We kept using n8n for certain workflows, especially those managed by our non-tech team. It’s now part of a hybrid system where the core functionality runs on our custom backend, but we use n8n for rapid prototyping and simpler integrations.
My advice? Start with a low-code platform to validate your concept and get quick wins. As you grow, be prepared to transition critical parts to a custom solution. This approach gives you the best of both worlds – quick start and scalability when needed.