November 18, 2025

We are often asked, what is needed to setup and pilot our RheinInsights Retrieval Suite. Therefore this blog post describes the steps which are needed for a successful RheinInsights Retrieval Suite integration in your environment.

Scope

Of course, you need to have an understanding of what you want to achieve. Do you want to try Copilot with third party sources? Do you want to implement your own bot application or an agent? What is its purpose?

Based on this, you need to define the actual pilot scope:

1. Which source system do you want to index?

2. Which search engine do you want to use?

3. And where should answers be surfaced - in a bot framework, in Copilot, intranet or in the web?

Image: Alex from the Rock – stock.adobe.com

Project Approach

1. Provide the technical prerequisites (see below)

   • Responsible: IT team

2. Download the latest Retrieval Suite release from our customer portal and acquire a trial license

   • see RheinInsights Customer Portal

   • Responsible: project team

3. Deployment and Configuration of the Retrieval Suite

   • see RheinInsights - Documentation - Deployment

   • [Responsible: project team]

4. Configure your connectors -

   • see RheinInsights - Documentation - Managing Connectors

   • Responsible: project team, content source owner for accounts, search engine owner for access

5. Initial crawl

   • Responsible: project team

6. Configure your search UX

   • see RheinInsights - Documentation - Search Experience and Query Pipelines and RheinInsights - Documentation - MCP, Agents and Bot Integrations

   • Responsible: project team, IT team

7. Review the quality of the results

   • Responsible: project team, business owner

8. Fine-tuning if needed

   • Responsible: project team

9. QA and UAT

   • Responsible: project team, business owner

10. Staging towards production

    • Responsible: project team

Technical Prerequisites

1. Provisioning of the host machine or container
   see RheinInsights - Documentation - Virtual Machine / Docker Container.

2. Database (optional)
   see RheinInsights - Documentation - Database Setup and Configuration

3. Provisioning of the search engine
   see RheinInsights - Documentation - Search Engines

4. If vector search should be used

   1. Provisioning of an embedding
      see RheinInsights - Documentation - Vectorizer and Embeddings

5. In case a search UX or bot integration is needed then

   1. Provisioning of an LLM
      see RheinInsights - Documentation - Managing Large Language Models

   2. Provisioning of the bot prerequisites
      see RheinInsights - Documentation - MCP, Agents and Bot Integrations