RAG Ecosystem
RAG ecosystem is huge and each day there are more and more tools, frameworks, technologies and databases poping up. The main separation line falls between the Build and the Buy motions.
| BUILD (technologies, tools) | BUY (ready-made products) |
|---|---|
| Graph Databases | RAGaaS |
| Vector Databases | Enterprise Search |
| Frameworks | Enterprise Search on top of LLMs |
| LLMs | |
| Memory Layers |
If you want to build, you have a list of tools and tehnologies to use and put together. If you want to buy complete solutions and apps, there's another list that is ready. Tradeoffs are apparent: no matter what they say, complete solutions and apps are not going to be custom to your needs and the your underlying data will live somewhere else. On the other side, if you build, your proprietary data will be local, on prem, and you will be able to customise your apps and Agents as you see fit.
Let's dive into the list.
Build
Graph Databases
Vector Databases
LLMs
Frameworks
Memory Layers
Buy
RAGaaS
RAGaaS or RAG-as-a-Service are horizontal equivalents to vertical semantic search and unstructured data transformation companies. They offer the ability to query unstructured data sources like PDFs, extract data, and input results into a more structured database or system of record.
Enterprise Search
Enterprise Search companies offer an end to end solution. They offer semantic querying for indexing and retrieving conceptually relevant documents in order to better manage organization-wide knowledge and break down enterprise data silos. It's a black box product. Take it or leave it based on what they built and envisaged.
Enterprise Search on top of LLMs
In a classic sense, LLMs are standalone B2C products for Q&A to consumers. But, some have added on more capable and serious RAG offerings on top of their basic B2C products.