I see Atlassian is focussed on creating many AI tools but I miss the connection of existing external data, like documents/manuals, datasets.
This is especially relevant for companies that have to deal with tons of red tape and API external specs. Most of them are not public, or at least not available to recent LLM.
So I think Atlassian should build a simple RAG tool that can index PDF/office documents, then the rovodev agent should query the RAG providing the results to the LLM context.
[User Input] ──┐
│
[Embed & Search Vector Store] ←─── Local RAG (client)
│
[Format Prompt with Retrieved Context]
│
[Send Prompt to LLM API]
│
[LLM Response Output]
This would definitely be a USP in the crowded market of coding agents
