> For the complete documentation index, see [llms.txt](https://help.dscout.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.dscout.com/dscout-ai/dscout-mcp/dscout-mcp-best-practices.md). # Dscout MCP best practices {% hint style="info" %} **Note:** The Dscout MCP is currently in closed beta, meaning only certain users have access and the feature will undergo changes prior to general availability. For more information, contact your Dscout Account Director. {% endhint %} As with any AI tool, it's important to keep in mind what an MCP excels at as well as where it may be prone to missteps. To help guide your work with the Dscout MCP, we've compiled a list of best practices for you to follow: ### Ground with context Start your chats by offering context about your research. Describe what your goals are or what you're hoping to learn. This can help kickstart your conversation with the AI, and it'll be able to offer you guidance if you get stuck or aren't sure what to look into next. ### Break complex work into chunks Sending one long prompt with multiple requests gives AI more chances to head in a direction other than what you intended. Instead, break your requests into logical chunks. This will give your conversation and workflow natural checkpoints for you to keep an eye on what's being done. Don't put all your eggs in one basket (at least not one held by a robot). ### Be specific when possible Don't let the AI guess. If you're looking for data from a specific study, clearly state that study's title, preferably in its full form. Abbreviations can often be misinterpreted, especially if you've got a few Dscout studies or projects with similar names. When drafting, tell the tool what you want it to do with as much detail as possible. Adding a question? Tell it where you want the question to be placed in the study, any wording preferences you might have, and anything else that's relevant to your specific request. ### Start broad before homing in When querying your data, start broad to ensure you and your AI tool are on the same page. For example, prompt the AI to locate the study or studies you want to analyze. Once you've confirmed it's looking at the correct data, then ask more specific questions. ### Always fact check Any AI tool can hallucinate or misinterpret your requests—and they'll often do it with confidence. So, we recommend occasionally verifying the AI's sources. If it's giving you participant quotes, pop into your study to verify they're accurate and not being taken out of context. ### Preview before launching After the MCP creates or modifies a study, open it in Dscout and review it yourself before going live. Check that questions read the way you intended, recruitment criteria and incentives are correct, and nothing was skipped or misinterpreted. ### Let us know what feels off If you run into something that doesn't feel quite right, we want to know! Reach out to so we can help and make any improvements where possible. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.dscout.com/dscout-ai/dscout-mcp/dscout-mcp-best-practices.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.