What are we working on?
Here’s a look at our main focus areas right now. It’s not quite a roadmap (they’re too rigid and create false expectations) but it’s our way of sharing what’s coming up.
Fluency continues to expand its tenant capabilities. This interface will allow MSSPs to add customers into a cost-effective, role-based access control infrastructure. This is the first stage of a four-stage roadmap.
Replay lets security teams re-experience real cyber incidents by injecting sanitized, timestamped telemetry back into the detection pipeline. It transforms training, tuning, and testing into repeatable, real-time simulations—bridging the gap between hindsight and action.
Fluency Telemetry Pipes currently use AWS S3. However, not all of our subscribers use AWS. We are developing a standalone version of Telemetry Pipes that can be used with any cloud provider—or even on-premises.
This is a large project. First, our Fluency Programming Language (FPL) has been structurally modified to be compatible with JavaScript ES6. Second, we have implemented a parse and test component that accepts audit samples and converts them into FPL parsers. This greatly enhances our ability to collect and parse logs from any source.
Fluency Assist is Fluency's GenAI Dynamic Workflow engine. This cybersecurity domain-specific engine uses AI to help make sense of UEBA alerts by providing clearer, human-readable descriptions. Phase 2 focuses on AI Actions that allow the system to request additional data beyond the initial alert, further enhancing clarity and improving decision making.
The Fluency Partner Interface is designed for MSSPs and large organizations to provide mid-level management of access control. This interface enables oversight across multiple tenants and includes features such as health monitoring and tenant breakdowns for billing and performance tracking.
Phase 3 focuses on AI Responses—actions that modify security controls to address risk. These responses can include actions such as credential refreshing (forcing a re-login and two-factor authentication). Responses may be executed automatically or require human approval.