Signal Forward Insights
One source of truth, many lenses. Insights is the reporting layer we build over your own data: role-specific views for the CFO, the division leader, and the physician, computed directly from source data elements, so every number traces to its origin and every access is audited.
The leadership view of the physician-productivity lens: organization KPIs computed from charge lines, and the AI-written period story beneath them, every number tied to its run. All values shown are fabricated demo data.
Not another rollup of a rollup
Most reporting is built on someone else's aggregation: a vendor dashboard or a separate analytics database, consumed as pre-rendered extracts. It inherits that layer's opacity and flaws, and it does not port when systems change.
Insights computes from the atomic rows your warehouse already holds. wRVU is CPT times units times the CMS RVU table. Collections and close readiness come off the same charge line the productivity view reads. Access comes from the appointment lifecycle. The lineage is honest, and the layer is portable.
It also settles arguments. The physician sees the same math the CFO sees, which is what lets a compensation or productivity conversation land.
Source the data once. Every lens builds on it.
Each view is a lens over one canonical data model rather than a re-ingestion. The productivity and quality lens serves physician enterprise leadership. The collections and close-readiness lens shown here serves finance, off the same charge line, so finance and operations argue from one number instead of two extracts.
New lenses (access and front office, surgical operations, referral flow) are built the same way: a metric pack, a role table, and a view over the model your team already trusts.
The AI writes the story. The engine computes every number.
Insights carries an AI-written narrative layer that stays inside the same discipline as everything else we build. The engine computes the metrics and detects the exceptions with versioned rules; the AI only writes the language around them. Every numeral in a narrative is validated against the metrics it cites before it is stored, and each narrative carries its audit record.
Shown here on the exception table: the engine finds the exception, the AI explains it, and every number in the explanation ties to source. No AI runs at read time, so the layer your governance review cleared is the layer that serves.
Scoped to the reader, audited by design
Role-scoped by default
A physician sees their own scorecard. A division leader sees their division. Leadership sees the organization. Scope is enforced server-side, default-deny, not hidden in the front end.
Every access audited
Each query is logged with who asked, what they saw, and under what scope. Patient-level drill-in is gated to a minimum-necessary capability. It is a reporting layer a governance and security review can clear.
Yours, in your environment
Built inside your tenant with your identity provider, against your warehouse. You own the model, the metric definitions, and the code, and they keep working after we hand off.
Insights shares one discipline with Provenance and the rest of what ships with our work: numbers computed from source through documented, versioned logic you can inspect and keep. Provenance shows you what the model did. Insights shows you what the business is doing.
Ask for the walkthrough
The working demo runs on synthetic data for a fictional orthopedic group. Tell us which numbers your leadership team argues about, and we will show you the lens that ends the argument.