Key Data at Your Fingertips: RAG, AI Agent, Knowledge Bank
Katya Dominguez
The problem
A government team spent hours hunting for answers across internal documents, reports and procedures. Every answer also had to follow the team's SOPs.
What I built
Structured ingestion: I curated the team's documents and SOPs, cleaned them and indexed them.
Retrieval (RAG): the system finds the passages that answer each question.
Grounded answers: the AI answers from the team's own material.
A Voiceflow assistant, where staff ask their questions day to day.
Ai Knowledge Bank: retrieval-augmented generation (RAG).
Tools
Voiceflow, OpenAI and n8n, with retrieval-augmented generation (RAG).
The result
Search time fell 50%.
The team saves 15+ hours every week.
SOP compliance went up.
My role
I designed the knowledge architecture and the retrieval, built the Voiceflow assistant and its integrations, tested the answers and handed the system over.
Next step
Private knowledge systems are a Custom Production AI System. Start with the Free AI Systems Fit Call and bring the questions your team asks most.
Like this project
Posted Oct 22, 2026
Search time fell 50% and staff save 15+ hours a week. I built a private RAG knowledge bank in Voiceflow that answers a government team from its own SOPs.
