> For the complete documentation index, see [llms.txt](https://descipher.gitbook.io/doc/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://descipher.gitbook.io/doc/overview/how-it-works.md).

# How It Works

1. **User Query Submission:**
   * A researcher enters a research topic or question into a simple search bar, much like a Google search.
2. **Task Distribution via the Orchestration Engine:**
   * A central orchestration engine receives the query and distributes specific tasks to each of the five AI agents.
   * The engine handles API calls, manages data flow, and ensures that each agent works on its designated area.
3. **Data Collection:**
   * Each agent connects to its relevant databases and sources:
     * The Research Agent gathers articles.
     * The Patent Agent scans patent records.
     * The Collaboration Agent identifies experts and potential partners.
     * The Funding Agent finds financial opportunities.
     * The Legal Agent checks for compliance and regulations.
4. **Synthesis Module:**
   * After each agent completes its task, the outputs are sent to a synthesis module.
   * Powered by GPT-4, this module combines the individual data points into one clear, actionable report, highlighting insights, trends, and recommended next steps.
5. **Final Report Delivery:**
   * The complete report is delivered back to the researcher, providing a holistic view of the research topic with recommendations and potential areas for further exploration.

**Technical Considerations:**

* **Orchestration:** The engine ensures efficient communication and task management among the agents.
* **Scalability:** Designed to handle multiple queries simultaneously, the platform can grow with user demand.
* **Flexibility:** The modular design means that additional agents or features can be integrated in the future as needed.