Modern business is data-driven. Decisions are based on data, there are a number of large consulting firms who do nothing else than data analysis, and presentation of insights based on those data.
Typically though, the insights are answers to natural questions that businesses have – such as
- which department has the highest net profit, highest costs?
- which product line would be the most profitable if I increased prices?
- in which country/region should I focus my sales activities?
- …
To get these answers today, the process looks like follows:
- humans need to understand the questions from the business
- humans need to understand data that exist in the client organization
- humans need to calculate results based on those data
- humans need to interpret those results to the business
Now, this is extremely costly, and very often disappointing – since the business is not getting the answers they are looking for.
This idea is about creating a platform, using modern chatGPT-like tools, to skip all that, and have business people asking this platform their question in natural language, and getting natural language responses back.
Update July 2024
- There are platforms that offer this functionality – Databricks Assistant, Microsoft Copilot, etc.
- However, what is still a good opportunity is creating a friendly UX, that is suited for SMB, and behind the curtains leverages for example https://python.langchain.com/v0.1/docs/use_cases/sql/ – to translate the business questions to SQL
0 Comments