This is a submission for the Runner H "AI Agent Prompting" Challenge
What I Built
I built an AI-powered Research Assistant using Runner H that automates the entire literature review process. The agent:
🔍 Searches arXiv, PubMed, and IEEE for newly published papers
🧠 Filters results by recency (7 days) and relevance (85%+)
✍️ Extracts core contributions, methodologies, and limitations
📄 Compiles clean, publication-ready summaries in Google Docs
📢 Instantly posts the final report to Slack for easy team access
Whether you're a researcher, student, or startup founder, this on-demand agent saves hours and delivers high-quality insights in just one prompt.
Demo
📹 Watch the full workflow in action:
🖼️ Screenshots:
1. Generated Research Summary in Google Docs
2. Slack Notification with Summary Link
How I Used Runner H
The Full Prompt: "You are my Research Assistant. First, ask me what topic I would like the research summary to focus on. Once I provide the topic, confirm it by saying something like “Great, you’ve chosen: [Topic]. Should I proceed?” After confirmation, search arXiv, PubMed, and IEEE for new academic papers related to that topic, filtering for those published in the past 7 days with relevance score >85%. For each paper, extract the core contribution in one sentence, list three key methodologies used, and identify any limitations. Compile the findings into a Google Document titled “[Topic] Research Summary – [Date]”, formatted with clear sections for the title, contribution, methodologies, and limitations for each paper. Finally, share the link to the Google Doc in the Slack #research channel with a brief summary message".
🛠️ What the Agent Does
This prompt powers an autonomous research assistant that handles the entire literature review process from a single input. Here's how the workflow breaks down:
1. Dynamic Topic Collection
The agent starts by asking the user to provide a research topic, this keeps it reusable across fields and disciplines.
2. Intelligent Search + Filtering
It searches arXiv, PubMed, and IEEE Xplore for relevant academic papers published in the last 7 days, then filters out any that score under 85% relevance based on semantic similarity to the given topic.
3. Structured Information Extraction
For each papers, the agent extracts:
- A core contribution summary
- Three key methodologies
- Any stated limitations by the authors
4. Formatted Report Generation
The agent automatically creates a Google Doc with organized sections, making it ready for publishing, team review, or academic citation. Each paper is clearly broken into Title, Contribution, Methodologies, and Limitations.
5. Slack Integration
The final document is posted directly into a dedicated Slack channel (#research) so the team or user can access the latest findings instantly.
Use Case & Impact
This solution is perfect for:
- 🧪 PhD students and academic researchers conducting topic reviews
- 💼 Startup founders and execs keeping up with trends in their domain
- 📊 Product, growth, and AI teams needing technical insight without drowning in PDFs
- 🎓 Students working on dissertations, term papers, or proposals
🔍 Why It Matters
- Manual literature reviews are time-consuming and error-prone.
- Searching across multiple databases and extracting meaningful insights takes hours.
- Most people don’t have a system for filtering by relevance, and often rely on keyword matches alone.
This agent delivers:
- 🚀 Speed — cuts down research time by 80%
- 🎯 Precision — filters irrelevant results with semantic analysis
- 🧘🏾♀️ Convenience — compiles everything and drops it in Slack, hands-free
- ♻️ Reusability — works across any research domain with zero code changes
It's a scalable way to support better, faster, and more informed decision-making in research-driven environments.
Social Love
I’d love to be considered for the Community Champion prize category, this project is for researchers, students, and anyone trying to save time with AI!
Top comments (0)