Back office ops · Production
The case for using structured and semi-structured data in generative AI
The problem
Most discussions of generative AI assume unstructured data, but organizations also hold large amounts of structured and semi-structured data in SaaS applications and operational databases that they want to query with AI.
Workflow diagram · grounded in source
1
Extract and vectorize table text
ai_action
“The existence of free-form text in your tables gives you an opportunity to turn structured data into unstructured data. You can directly extract and vectorize the contents of text-rich fields from tables. Depending on the additional fiel…”
2
RAG chatbot answers questions
ai_action
“build a retrieval-augmented generation (RAG) chatbot to consolidate its knowledge base, answer questions about its operations”
3
Natural language to query
ai_action
“Business intelligence platforms increasingly leverage generative AI to convert natural language into queries or scripts”
4
Reports and metrics produced
output
“produce charts, tables and metrics as needed for reporting”
Reported outcome
(not stated)
Reported stack
Fivetranvector databases
Source
https://www.fivetran.com/blog/the-case-for-using-structured-and-semi-structured-data-in-generative-ai
Read source ↗Frequently asked questions
What did this team achieve with this AI workflow?
(not stated)
What tools did this team use?
Fivetran, vector databases.
How is this back office ops AI workflow structured?
Extract and vectorize table text → RAG chatbot answers questions → Natural language to query → Reports and metrics produced.