search

deepset-ai / haystack

:mag: AI orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited for building RAG, question answering, semantic search or conversational agent chatbots.
https://haystack.deepset.ai
Apache License 2.0
17.55k stars 1.91k forks source link

Investigate and refine insights report bot #7740

Closed mrm1001 closed 2 months ago

mrm1001 commented 5 months ago

In business intelligence or insights research, it is common to need self-driven agents that have access to multiple tools to fulfil a scoped task.

User story

Build a business intelligence agent that has access to multiple tools to generate information to later create a report (outside of the scope of this agent). Example of the tools: search_property_database, search_property_web, do_map_search, crawl_web_page. In this case, the agent might need to use to same tool multiple times (e.g. if they need to do web search followed by crawling web pages) or different tools (e.g. if they first do a search on the property database search_property_database and use the web search_property_web if the results are not good enough).

This agent needs to be able to be used in a multi-agent architecture. The actual multi-agent architecture is outside the scope of this issue.

Outcome

The outcome of this issue would be a colab or notion page with either pseudo-code or a diagram explaining how this problem could be solved from an end-user point of view, without going into implementation details.

julian-risch commented 2 months ago

Let's close this and focus instead on new tool abstractions as part of haystack-experimental. We're tracking the work on a notebook here https://github.com/deepset-ai/haystack/issues/8312