When prompting Meltwater Copilot, you have two ways to guide its response—each serving a different purpose.

Reference a Saved Search for User Defined Results

Referencing a Saved Search ensures Meltwater Copilot pulls from a defined dataset, giving you consistent results every time you use that prompt. This is ideal when you want your outputs—such as numbers, insights, and mentions—to align with Explore. It’s especially useful for reporting, recurring workflows, or when validating against known criteria.

Reminder: if you use dynamic criteria, such as “the last 7 days,” the data will reflect this, just as it would in Explore. E.g. Results for the prompt “What is the top news coverage for Meltwater from the last 7 days” may vary if you run it in the morning and then the afternoon.

Explore New Topics Without a Saved Search

When a Saved Search isn’t included, Meltwater Copilot will craft its own search parameters at the time of the prompt using a combination of Boolean logic and vector-based techniques. This allows for more flexibility and creativity in how results are surfaced, making it perfect for researching new topics, uncovering trends, or exploring areas you haven’t yet defined.

Because these backend searches are dynamically generated, results may vary slightly between sessions or differ from those in Explore if a Boolean query is manually crafted.