Activity
Mon
Wed
Fri
Sun
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
What is this?
Less
More

Memberships

Nexis AI w/Jason

488 members • Free

Assistable.ai

3.1k members • Free

AI Accelerator

16.5k members • Free

No-Code Nation

3.1k members • Free

41 contributions to Assistable.ai
WhatsApp AI Restrictions in 2026 — Anyone Know More?
I’ve seen some news about WhatsApp planning to restrict the use of general-purpose AI on the WhatsApp Business API starting in 2026. Does anyone here know more details about this change or how it might affect AI-powered workflows? I’m specifically wondering if business use cases like appointment confirmations, rescheduling, and follow-ups will still be allowed. Any insights, documentation, or experiences would be greatly appreciated. Thanks! https://www.whatsapp.com/legal/business-solution-terms/preview?lang=en
0
0
Claude rocks for prompt engineering and RAG optimization
I was struggling with a prompt and RAG knowledge base for days. Decided to use Claude to help me diagnose and rebuild it, within 30 minutes, I fixed a problem that had stumped me for days. What made the difference: - Claude excelled at structuring data for RAG retrieval - Caught edge cases I missed in my pricing logic - Created both markdown AND JSON formats without asking - Spotted ambiguities in my original prompt that were causing issues Takeaway: Different AI tools have different strengths. For structured data work, prompt engineering, and systematic problem-solving, Claude definitely gets my vote.
0 likes • Oct '25
Interesting, Christian — I’ve been struggling with something similar. In my case, I haven’t been able to get reliable retrieval results from the Knowledge Base, especially when trying to extract pricing data. We’ve tested different chunking and embedding setups — full-text, file mode, JSON, Markdown — but none gave consistent results. When you say Claude helped “structure data for RAG retrieval,” do you mean you used Claude just to structure the Knowledge Base, or that you actually used Claude as the model powering the RAG assistant? Would love to understand how you approached it.
How to switch from one assistant to another in the same conversation?
I’m not sure if this is a very obvious question, but I haven’t managed to figure it out. Does anyone know if there’s a way to switch from one assistant to another within the same chat conversation ? I’ve tried using automatic tags, but they don’t seem very reliable. What I’d like is to transfer control depending on the need — for example, moving from a sales bot to a marketing bot, or vice versa. Is there a more stable way to achieve this?
0 likes • Sep '25
@Sean McGinnis Thanks Sean, just to clarify — are you referring to the general prompt or inside the tools’ prompt? Because in the general prompt I’ve never tried it, I didn’t even think it could work. I’ll test it. The truth is, the tagging tools have never worked reliably for me, so I’m still not sure if there’s a more stable way to handle this.
0 likes • Sep '25
@Michael Kittinger thanks Michael, I tried doing it that way, but the problem is that if the assistant has label A (for example, original), and we tag it with label B (which should remove label A and activate label B), the issue is that the assistant’s first response is always based on label A, not B. So the problem is that the assistant doesn’t switch immediately to the new label.
Feature request: use custom fields in tools & see tool call values in logs
I’ve noticed two things that would be great improvements: 1. Custom fields – right now they can’t be used inside tools, only in the prompt. It would be amazing if tools could also access those values to send as parameters. 2. Logs visibility – when a tool call is made, I can only see what was sent through the webhook. It would be super helpful if the logs showed the exact values that were sent with the tool calls directly inside Assistable.
0 likes • Sep '25
@Harry Stokes No, for chat. If you put the variable {} in the tool it doesn’t recognize it. As prompt from the main prompt it is a lottery.
FlowBuilder
Hi, I am aiming to use the new FlowBuilder's test messaging to test out the prompt I have built out. I ran into the initial error of: 401 Incorrect API key provided: sk-proj-********************************************************************************************************************************************************5qgA. You can find your API key at https://platform.openai.com/account/api-keys. I have created an API Key and added it to my WorkSpace that shows OpenAI is connected. Is this the issue? Right now, OpenAI no longer gives out the old-style sk-... legacy keys for new projects or accounts — all keys are now project-based (sk-proj-...). Any feedback on how to get this working would be greatly appreciated. Thank you!
FlowBuilder
0 likes • Sep '25
I tried too. Faced same problem.
1-10 of 41
Felipe Ortiz
2
6points to level up
@felipe-ortiz-8706
Aprendiendo de Automatización con AI

Active 6d ago
Joined Apr 30, 2025
Powered by