Have you AI purposefully misspell words and correct itself to seem more human. Just turn it on and set a percentage of frequency which will be random on each run - recommended 8% - 12% for the most human feel :) This will only work in production at the moment, not chat lab*

Bots ring busy. AI voice is out. Time to be honest: This is the 4th time AI voice is broken in past 20 days. First it was GHL outage - than it was buildship outage, than we had server capacity issue and now its probably SQL provider outage (or maybe even a 5th new case) Outages happen, but it has been too frequent to consider it normal. These outages define Assistable as an unstable system. As someone who has dedicated all time & resources to sell this solution, this de-motivates me to promote the product to anyone as I know for a fact it is an unstable product, I feel like I will have too many headaches when I have 5-10 voice clients. It would be great to hear stories from community members that have 5+ voice clients, how do you convince your clients to continue paying for a system with this many outages? I hope the bot down time will be shorter compared to previous outages. @Jorden Williams

Can anyone confirm if the text knowledge base is working properly. I have about 8-9 pages of plain text (markdown) tags but it doesn’t appear the assistant is using it. I have also tested in the playground and the results are terrible. I have used q&a before which was fine but never text on assistable. Should I do my own rag?

I've just deployed Cole Medin's Agentic RAG system for our real estate database, with some custom modifications worth sharing: What it does: 🔍 📊 Creates a dynamic knowledge base that auto-updates whenever source spreadsheets change 💬 Handles complex natural language queries across your data ("Find products with specific features" or "Compare items matching certain criteria") 📱 Powers customer-facing interactions through QR codes, websites, or messaging platforms 🔗 Returns precisely formatted responses with accurate links and only the most relevant details 📈 Works for any industry with structured data - real estate, e-commerce, inventory management, service catalogs, product specifications, or company knowledge bases What makes this system awesome: ✨ 🧠 It's super smart: It knows when to search for information, when to do math, and when to grab the whole document—just like knowing when to use different tools for different jobs 📚 It reads entire documents when needed: Instead of just looking at random pages, it can read the whole document to understand everything 🔢 It's great at math problems: It can calculate numbers from spreadsheets without making mistakes 🧩 It connects the dots: It can find information that's spread across different files and put it together—like solving a puzzle 📂 It handles lots of files at once: You don't have to upload one file at a time; it works with all your information together 💾 It saves space: It stores all your information in a clever way that doesn't waste computer memory Technical implementation: ⚙️ 🛠️ Built on Assistable using Cole's n8n workflow template 📝 Modified with custom SQL queries for property-specific lookups 🤖 Using ADA-002 for embeddings and GPT-4.0 for responses (Claude Opus seemed slower) 🗃️ Supabase backend with 4000 token chunks and zero overlap 🗣️ Text and voice response formats with separate prompts for each Challenges faced: 🚧 🔗 Ensuring links remain functional through vectorization 📊 Finding the right chunking strategy for property data

So, we have to head back home (14 hrs on the road) and we'd like to do it mid day. If I can cancel wednesday call, I will get voice KB and fine-tuning models / kbs done before Friday and we can do a full build with all the new stuff on Friday. Good deal or no? (you know ill do it anyways I just dont want to drive all night lol)