Many companies want to automate internal workflows with AI. Support tickets, bug reports, documentation search, customer questions, and developer tasks can all be partially automated.

In this guide we will build a simple but powerful AI workflow orchestration system using open-source tools:

Open WebUI – internal AI interface for employees
n8n – workflow orchestrator
Qdrant – semantic search database for AI knowledge
PostgreSQL – structured data storage

This stack lets you build things like:

AI-assisted support ticket triage
Internal company AI assistant
Automated bug reporting workflows
AI answering questions using company documentation

Everything here can run on one Linux server.

1. What We Are Building

Imagine a SaaS company workflow:

A customer submits a support request.
The system automatically:
- analyzes the request with AI
- searches company documentation
- creates internal tasks
- suggests a reply to support staff

2. Server Requirements

For a small company setup:

Recommended server:

Ubuntu 22.04
16 GB RAM
4 CPU cores
100 GB disk

Install Docker first.

Bash

sudo apt update
sudo apt install docker.io docker-compose -y

3. Docker Compose Setup

YAML

version: "3"

services:

  postgres:
    image: postgres:15
    environment:
      POSTGRES_USER: ai
      POSTGRES_PASSWORD: ai123
      POSTGRES_DB: workflows
    volumes:
      - postgres_data:/var/lib/postgresql/data
    ports:
      - "5432:5432"

  qdrant:
    image: qdrant/qdrant
    ports:
      - "6333:6333"
    volumes:
      - qdrant_data:/qdrant/storage

  n8n:
    image: n8nio/n8n
    ports:
      - "5678:5678"
    environment:
      - DB_TYPE=postgresdb
      - DB_POSTGRESDB_HOST=postgres
      - DB_POSTGRESDB_DATABASE=workflows
      - DB_POSTGRESDB_USER=ai
      - DB_POSTGRESDB_PASSWORD=ai123
    depends_on:
      - postgres

  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    ports:
      - "3000:8080"
    volumes:
      - openwebui_data:/app/backend/data

volumes:
  postgres_data:
  qdrant_data:
  openwebui_data:

Start everything:

Bash

docker compose up -d

4. Access the Tools

Open WebUI: http://SERVER-IP:3000

n8n dashboard: http://SERVER-IP:5678

5. Add an AI Model

Open WebUI needs a model backend. The easiest option is to run Ollama on the host.

Install Ollama (local model server):

Bash

curl -fsSL https://ollama.com/install.sh | sh

Run a model:

Bash

ollama run llama3

6. Storing Knowledge in Qdrant

AI becomes much more useful when it can search company knowledge.

Examples of documents to store:

Product documentation
Support procedures
API documentation
Internal manuals

7. Creating the Workflow in n8n

Open the n8n dashboard. Create a new workflow.

Example automation: Webhook → AI Analysis → Knowledge Search → Store Ticket → Suggest Reply

Step 1: Webhook trigger
POST /support-ticket

Step 2: AI analysis
Send ticket text to the AI model.

Step 3: Knowledge search
Query Qdrant for similar documentation.

Step 4: Save ticket
Insert structured data into PostgreSQL.

Step 5: Suggested reply
The AI combines ticket content and documentation results to generate a draft response.

8. Example AI Prompt

Text

You are a SaaS support assistant.

Customer message:
{{ticket_text}}

Relevant documentation:
{{qdrant_results}}

Write a clear support reply.
Do not invent information.

9. Real Company Use Cases

AI documentation assistant

Employees ask questions about internal systems.

Automated support classification

Tickets automatically routed to the correct team.

Developer bug report generation

AI extracts logs and creates developer tickets.

Customer onboarding assistant

AI answers new customer questions using documentation.

10. Why This Stack Works

Open WebUI – Human interface to AI
n8n – Automation and orchestration
Qdrant – AI knowledge memory
PostgreSQL – Structured business data

11. Final Advice

Start simple. Do not try to automate everything at once.

A good first project is: "AI suggests replies to support tickets."

Once that works, expand to:

Ticket routing
Knowledge search
Developer workflows

This is exactly how many companies introduce AI into real business operations. Small automation steps, built on a solid foundation.