RedFlarePlaybooks
A playbook defines how RedFlare analyzes a conversation. It specifies what RedFlare detects, how interactions are evaluated, how scores are calculated, and how the overall results are generated.
Whenever you upload or process a conversation, you select a playbook. This selection determines the analysis performed on the and influences the insights displayed across the platform, including Conversations , Dashboards, Alerts, and Agent Views.
What a playbook decides
Every playbook bundles three kinds of instructions:
- Features to detect. The things RedFlare measures and pulls out of the call, such as sentiment, pace of speech, talk time, and other signals. These become the metric cards and insights you read on a conversation.
- Events are pre-defined AI models that identify notable moments in a conversation, such as interruptions, escalations, or other key interactions. RedFlare surfaces these events in the Events tab, allowing you to quickly navigate to the most important parts of the conversation.
- Rubrics to Score are pre-defined scorecards that RedFlare uses to evaluate conversations. Each rubric defines the evaluation criteria for a specific aspect of the interaction, such as compliance, communication, or process adherence. Together, the selected rubrics generate individual parameter scores and contribute to the overall weighted score displayed on the Scorecards tab.
On top of these, a playbook also drives the summary and key takeaways RedFlare writes for each call.
How playbooks shape your insights
The selected playbook determines which features, events, and rubrics are applied during conversation analysis, directly influencing the insights generated. A comprehensive playbook provides richer analysis with more detected events, detailed sentiment timelines, and additional scorecards, while a simpler playbook focuses on a smaller set of insights.
If an expected insight is not available for a conversation, it is likely because the playbook used to process that conversation did not include the corresponding feature, event, or rubric.
Two calls can look very different simply because they were run through different playbooks. To compare like with like, process them with the same playbook, and reprocess a call whenever you want a different set of insights, events, or scores.
Where you pick a playbook
Playbooks are not a menu you visit on their own. Instead, you choose one in the moment you bring a call in or run it again:
- When uploading, the Select a Playbook step lets you pick the playbook that will process your new recordings.
- When processing or reprocessing existing calls, you choose a playbook in the processing panel before you confirm.
If your calls come from more than one integration, you pick a playbook separately for each one before processing. See Conversations for the full upload and processing steps.
Statuses and versions
Each playbook carries a status that tells you whether it is ready to use:
- Active playbooks are live and available to process calls.
- Draft playbooks are still being put together and are not ready for everyday use.
- Inactive playbooks have been set aside and are no longer in active rotation.
Playbooks also have versions. As a playbook is refined over time, its version moves forward, so you can tell which iteration was applied to a given call. When you reprocess a call, RedFlare uses the current version, which is why reprocessing can refresh or change the insights you see.
The name "Playbooks" is what you will see in the app. A playbook lists how many features, events, and rubrics it includes, so you can tell at a glance how much analysis it will run on your calls.