The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
AI Assistant Using Your Business Knowledge Base — RAG on Your Documents
THE PROBLEM
Q&A bots break down when knowledge lives in documents: a 100-page manual has no "questions" to match, it can't fit into a prompt, and generic chatbots hallucinate instead of admitting what they don't know.
THE SOLUTION
A RAG (Retrieval-Augmented Generation) knowledge base: documents are split into meaningful chunks, embedded into a vector index, and the assistant answers from the right sections — by meaning, not keywords.
Any format as-is: PDF, DOCX, TXT, Markdown — 100+ pages is fine
Answers grounded in YOUR documents — it says "I don't have that information" rather than inventing
Source references — every answer shows which document and section it came from
Runs on your infrastructure — documents never leave your control
One command to re-index after updating documents — documented, no programmer needed
LIVE DEMO
Ask it about my products, pricing, deployment and support — it answers from 6 real documents with sources:
https://kb-assistant-demo.dvc-ai.workers.dev
The 'I don't have that information' line is the part most RAG builds skip, and it's the one that matters. How do you set the cutoff? On mine, a fixed similarity threshold broke once I filtered results by user role. Scores shifted and it refused questions it could answer.
Good catch — that's the classic failure mode with fixed cutoffs. Raw similarity scores aren't probabilities: they depend on the embedding model, chunk length, and question phrasing, so the distribution shifts whenever the candidate pool changes. Filtering by role is exactly that...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Challenges
View all
Related posts
AI Voice Agent for Real Estate Lead Booking
An AI-powered phone receptionist built for real estate agencies that answers inbound calls, qualifies leads in real time, and books showings directly to Google Calendar — all within the call itself, with no manual follow-up required.
The problem: Real estate agencies routinely lose leads to missed calls — after-hours inquiries, calls during showings, or overflow during busy periods. A slow callback often means the lead has already moved on to a competitor.
What it does:
Answers every call instantly, 24/7
Naturally qualifies the caller (buy/sell intent, area of interest, showing vs. agent call)
Collects contact details and preferred timing conversationally
Checks real-time calendar availability before booking
Prevents double-bookings and duplicate entries automatically
Confirms the appointment out loud before ending the call
Built with: Retell AI (conversational voice layer), n8n (workflow automation and business logic), Google Calendar API (scheduling)
Available for: Custom builds for real estate agencies, brokerages, or property management companies looking to stop losing leads to missed calls
Beautiful workflow by the way. My nicheeee
The Museum of the Impossible | Surreal Architecture in Motion
This film explores how a change in the rules of space can change the way we feel inside it.
A stone arch becomes weightless. Water hangs overhead. A staircase opens towards the sky. These impossible elements invite the viewer to pause and reconsider familiar materials and forms.
I connected the scenes through warm stone, blue light and slow camera movement. The restrained visual language gives each unusual event room to register and creates a sense of calm curiosity.
My work included concept development, visual direction, AI scene generation and animation through Higgsfield, editing, titles and an original instrumental soundtrack.
For galleries, digital art projects, design studios and architecture brands, films like this can introduce a creative direction, communicate the atmosphere of an imagined space or become part of an exhibition’s visual identity. They make an abstract idea easier to see and experience.
Have a spatial concept or an idea for an art project? Send me a message to discuss how we could develop it into a film.
20 seconds · Full HD · Independent AI portfolio concept featuring imagined architecture.
Your website is no longer just being read by people.
AI search engines are reading it too and they understand your website in ways surprisingly similar to how a good UX designer evaluates it.
@Framer's latest AEO guide highlights a few things that matter:
1. Clarity beats cleverness - If a page jumps between ideas, uses vague language, or assumes too much context, AI systems have a harder time understanding what the page actually means.
2. Structure matters - Clear headings, sections, summaries, definitions, and FAQs help AI break content into usable chunks. That is essentially good information architecture.
3. Context > Keywords - AI search is not simply matching keywords. It is trying to understand relationships between ideas, anticipate follow-up questions, and determine whether a page provides enough context to answer them.
This changes how I think about AEO.
You don't need to "write for AI." You need to build websites that are genuinely easy to understand.
For humans first. And increasingly, for machines too.
That means website strategy, content architecture, UX, SEO, and AI discoverability are becoming much more connected.
The interesting part?
Good UX is starting to become good AI optimization.
If your website is struggling with poor AEO results, book a free discovery call today!
