Create README.md

Author: mwhalin

samples/agent-catalog/msft-agent-samples/foundry-agent-service-sdk/magentic-one-agent/README.md

@@ -0,0 +1,83 @@
+# 🧠 ResearchFlow Agent
+
+This agent orchestrates a multi-step research workflow using AI, helping users gather, synthesize, and structure knowledge from complex sources—ideal for analysts, strategists, product teams, and technical writers.
+
+---
+
+## 💼 Use Cases
+
+- **Market & Competitive Research**: Automate the collection and summarization of competitor data, trends, and product insights.
+- **Technical Literature Review**: Extract structured summaries from technical documents and API documentation.
+- **Customer Deep Dives**: Combine customer feedback, CRM notes, and meeting transcripts into coherent insight packs.
+- **Strategic Briefing Packs**: Generate polished summaries, outlines, and visual insights for leadership-ready deliverables.
+
+---
+
+## 🧩 Tools & Capabilities
+
+This agent is built with **Azure AI Agent Service** and typically integrates:
+
+- **File Search** to extract and cross-reference details from uploaded PDFs, Markdown, and structured datasets.
+- **Summarizer Agent** to generate section-level summaries from technical or dense input.
+- **Planner Agent** to break large prompts into logical subtasks and route them to specialized sub-agents.
+- **Progress Tracker or Router Agent** to manage task completion and coordination across a multi-agent graph.
+
+The agent orchestration is defined using `.agent` and `.fdl` configuration files.
+
+---
+
+## 🧠 Architecture Overview
+
+![Architecture Diagram](assets/architecture-researchflow.png)
+
+---
+
+## ⚙️ Setup Instructions
+
+### Prerequisites
+
+
+---
+
+## 💬 Example Agent Interactions
+
+**User**: Can you generate a competitive landscape analysis on cloud-native agent orchestration tools?  
+**📥 Agent Response**: Planner routes task to summarizer + file search, returning a 3-part overview (market size, players, differentiators)
+
+---
+
+**User**: I uploaded three analyst reports—can you extract the trends relevant to SMB adoption and turn it into a leadership summary?  
+**📄 Agent Response**: File Search → Summarizer → Final summary document with key stats, quotes, and predictions.
+
+---
+
+**User**: What's the breakdown of feature gaps between our platform and Company X?  
+**📊 Agent Response**: File Search identifies platform docs; Summarizer creates comparative feature tables.
+
+---
+
+**User**: I'm prepping a 2-minute pitch—summarize these 10 pages of product reviews into themes I can speak to.  
+**🧠 Agent Response**: Sentiment clustering and summarization into 3 key talking points with example quotes.
+
+---
+
+## 🛠 Customization Tips
+
+- **Add Custom Tools**: Plug in web search or vector DB search to complement file-based knowledge.
+- **Tune Agent Routing**: Adjust the planner or agentRouter logic to improve task delegation.
+- **Control Summarization Style**: Modify the summarizer agent’s system prompt for narrative, factual, or bullet-style summaries.
+- **Export as Report or Slide Deck**: Add downstream tools that turn outputs into formatted PDFs or PowerPoint slides.
+
+---
+
+## 📁 Declarative Files Included
+
+- `ResearchFlow.agent` — main orchestration agent
+- `Summarizer.agent` — task-specific summarizer
+- `LedgerPlanner.agent` — planner for prompt decomposition
+- `progressManager.agent` — optional tool for managing intermediate steps
+- `DeepResearchAgent.fdl` — agent graph and tool configuration 
+
+---
+
+