Dev POC #1

Strategic Themes Prototype:

• Nx + next.js + MUI + Vercel AI SDK
• Ollama + Llama 3.1-8B
• Drawer (SAFe, Knowledge base, Ticket system)
• Main page area (Strategic Themes, OKRs)
• Chat area (open-webui like chat interface)

Strategic Themes + Chat with docs

Build something minimal with LLM + agent. Say, for Strategic themes. Possible features:

• possiblility to get help with how to formulate strategic themes
  • a chat interface (with a sidechat on the right side of the screen), like chat with docs.
  • The helper might be trained on SAFe articles and other data (recommended by SAFe documentation?) about strategic themes and OKRs, based on top of vector search and vector DB. (Check vercel example for this).
• possibility to generate strategic themes based on some data input
  • AI agent (built on top of one of AI frameworks)
• possibility to validate existing strategic themes against SAFe guidelines
  • • AI agent (built on top of one of AI frameworks)

Components

• sidechat
  • that comes along any core SAFe functionality, that comes in the left side of the window? And the side-chat in the right side of the window? Like how Pythagora works in vscode or zed.
• vector search
• vector DB
• ai agent

Tools

• kotaemon (opens in a new tab) ?

  • An open-source RAG-based tool for chatting with your documents.

• RAG with Vercel AI SDK (opens in a new tab) ?

  • RAG Chatbot Guide (opens in a new tab)
  • Vercel AI SDK
  • OpenAI (local LLM?)
  • Drizzle ORM (opens in a new tab)
  • postgresql (opens in a new tab) + pgvector (opens in a new tab)

• haystack (opens in a new tab) / mindsdb (opens in a new tab) ?

  • haystack github (opens in a new tab)
  • mindsdb github (opens in a new tab)

• firecrawl (opens in a new tab) or an opensource alternative? custom solution?

Examples

Train Local LLM with your own data

• "I want Llama3 to perform 10x with my private knowledge" - Local Agentic RAG w/ llama3 (opens in a new tab)

Build chatbot with AI framework (minds db)

• How to Build a Domain Expert Chatbot Using Langchain & MindsDB | Harrison Chase & Patricio Mardini (opens in a new tab)

Examples (Chat with docs)

• LangChain’s Harrison Chase on Building the Orchestration Layer for AI Agents | Training Data (opens in a new tab)

• How to Build a Domain Expert Chatbot Using Langchain & MindsDB | Harrison Chase & Patricio Mardini (opens in a new tab)

• The RIGHT WAY To Build AI Agents with CrewAI (BONUS: 100% Local) (opens in a new tab)

• Let's Build: Automate ANYTHING With AI Agent Teams (Step-by-Step) (opens in a new tab)

• PrivateGPT 2.0 - FULLY LOCAL Chat With Docs (PDF, TXT, HTML, PPTX, DOCX, and more) (opens in a new tab)

• Graph RAG UI: Powerful Chat with your Docs! (opens in a new tab)

• AnythingLLM Cloud: Fully LOCAL Chat With Docs (PDF, TXT, HTML, PPTX, DOCX, and more) (opens in a new tab)

• How To Build Your Own AI Chat with Docs (PDF, TXT, HTML, PPTX, DOCX, and More)! (opens in a new tab)

• PrivateGPT 2.0 (opens in a new tab)

• Chat with Docs using LLAMA3 & Ollama| FULLY LOCAL| Ollama RAG|Chainlit (opens in a new tab)

• Multi-modal RAG: Chat with Docs containing Images (opens in a new tab)

• RAG + Langchain Python Project: Easy AI/Chat For Your Docs (opens in a new tab)

• GPT4All - Chat with your Private Documents with LocalDocs Plugin (opens in a new tab)

• ULTIMATE Llama 3 UI: Chat with Docs | Open WebUI & Ollama! (Part 2) (opens in a new tab)

• How To Install Quivr 🏹 - Chat With ANY Document from ANYWHERE! (PDF, CSV, PPT, TXT) (opens in a new tab)

• Private Chat with your Documents with Ollama and PrivateGPT | Use Case | Easy Set up (opens in a new tab)

• AnythingLLM: Fully LOCAL Chat With Docs (PDF, TXT, HTML, PPTX, DOCX, and more) (opens in a new tab)

• PDF To Chat: Chat with your PDF Documents on the Web For FREE! (opens in a new tab)

• How To Install PrivateGPT - Chat With PDF, TXT, and CSV Files Privately! (Quick Setup Guide) (opens in a new tab)

• PrivateGPT 4.0 Windows Install Guide (Chat to Docs) Ollama & Mistral LLM Support! (opens in a new tab)

• Chat With Documents Using ChainLit, LangChain, Ollama & Mistral 🧠 (opens in a new tab)

• How to Build an AI Document Chatbot in 10 Minutes (opens in a new tab)

• How to chat with your PDFs using local Large Language Models [Ollama RAG] (opens in a new tab)

• How to install and use PrivateGPT 2.0 Locally | Chat with Docs ( TXT, HTML, DOCX , PDF and more ) (opens in a new tab)

• Retune to make a live chatbot, chatgpt, chat with docs, chat with pdf files and more using re:tune (opens in a new tab)

• ChatGPT With Your Docs | Full Tutorial WITH Code Examples (opens in a new tab)

• PrivateGPT is INSANE! 🤯 Chat with your Docs, TXT, PDF and more 🚀 100% LOCAL (Full Tutorial) (opens in a new tab)

• AnythingLLM - Chat with Any Docs with full Privacy|Runs Offline|FREE LOCAL LLMs + NO Code (opens in a new tab)

• LangGraph + Adaptive Rag + LLama3 Python Project: Easy AI/Chat for your Docs (opens in a new tab)

• AnythingLLM Free AI for Fully LOCAL Chat with Docs | Save Your Money (opens in a new tab)

Future

Personas ?

What would persona (s) for this task be? What roles according to SAFe are eligible to formulate strategic themes (OKRs) ? This could probably taken outside of the scope of POC, for the sake of speed and simplicity.

Possible feature

Once and if other parts of the app become available, it is possible to connect their data to the strategic themes helper and vice versa in different ways. One of the things - track the OKRs progress based on lower level items (features, enablers, tasks, etc.), and provide visibility and intelligence about those. Summarize on each respective hierarchy level.