Information retrieval with GraphRAG
Motivation
Discover the generative power of Microsoft's GraphRAG for your information retrieval. Customers as well as employees can benefit from quick and most importantly precise and comprehensive responses generated by large language models. In our workshop, we will guide you to identify where your business benefits most in using a GraphRAG approach compared to a usual RAG approach when querying language models.
This workshop lays out the importance in retrieval of contextual information in a short amount of time. Finding results most closely to the words used in a query asking a chatbot does not always brings the most accurate responses. We want to fine tune your chatbot by using a GraphRAG approach in order to deliver comprehensive results to your customers or employees depending on your business needs.
The primary goal of this workshop is to showcase the benefit of improved information retrieval that is individualized to your business needs.
What we bring
20 years of experience with software architecture and custom software development projects for many of the world’s largest enterprises has generated the expertise needed for PRODYNA to help you design and develop applications in the Azure Cloud.
We are a Microsoft Solutions Partner for both Digital & App Innovation and Data & AI and are a member of the Cloud Native Computing Foundation (CNCF), an open-source organization which maintains many of the innovative AI frameworks and projects used in data science projects today. With this offering we will assist you with:
- Understanding the features and capabilities of Microsofts’ GraphRAG.
- Understanding the use-cases in which GraphRAG is beneficial and outperforming other information retrieval techniques.
- Demonstrating how customized chatbots can be integrated into your existing business workflows and applications to achieve quick wins.
- Understand how to generate business value with information retrieval.
What you need
All you need are people interested in information retrieval (think “Chat GPT”) and how this can be even more improved in quality and speed related to your business information. This should be non-technical persons that know and understand your daily business. Persons from your IT are welcome, but not necessarily required.
- Complex business information, or a business process.
- 1 day of continuous time.
What you get
PRODYNA will guide you and your employees through the following program:
Introduction
(45 min)
Introduction from PRODYNA:
- Setting the stage, expectations, and goals for the day.
- Introduction and explanation of what GraphRAG is and how and why custom information retrieval solutions can benefit of such implementation.
- Overview of today’s scenarios.
Identify use cases
(120 min)
Split into work groups:
- Identify where your business information is most needed and define the most important topics.
- Highlight where information is complex and contextual responses are required, where a quick response brings the most benefit.
- Mark where visualization could be beneficial.
Information retrieval with GraphRAG vs RAG
(120 min)
Work in groups:
- Explore results from vector embeddings versus vector similarity on (a) general example(s).
- Create a knowledge graph based on your business data.
- Query your business data and highlight the information used in the knowledge graph.
- Identify topics of disperse information that is required to produce an accurate answer.
Hypothesis
(45 min)
Group discussion:
- Present the findings from the workgroups.
- Create one, or more hypotheses for using a customized chatbot using vector similarity to retrieve business related information (i.e. if employees can ask a chatbot instead of corresponding with multiple other coworkers we gain precision and save x hours in delivering …, or: if customers receive precise information about our products we can reduce multiple calls in our customer center)
- Next steps
Quick facts
- Duration: 1 day
- Up to 12 participants
Benefits
- A quick understanding of information retrieval using GraphRAG.
- An understanding of how to customize a chatbot using GraphRAG and provide a more precise and comprehensive tool to answer any business-related question.
- Identification of potential use cases, or steps in your business workflows where an GraphRAG approach could generate business value for your organization.
Contact us
for more information.
David Wainwright
