AI you can audit.
Models you can defend.
A boutique healthcare data and AI consultancy for the organizations, and the counsel, that cannot deploy black-box AI. We use AI to build deterministic, auditable models that automate the decisions that have to hold up. The model reads and writes; the money, coverage, and risk decisions run as versioned, auditable code.
One craft, applied where the risk is highest
We use AI to build deterministic, auditable models for the decisions that carry the most risk in healthcare. Each one is engineered for the people who will challenge it: a CFO underwriting savings, a compliance committee, an auditor, or opposing counsel.
Revenue Cycle & Denial Prevention
Deterministic, auditable coding and denial-prevention models that lower cost-to-collect and recover revenue, built so a CFO can underwrite the savings and an auditor can trust them.
Accounting & Close Automation
Deterministic, auditable automation for the finance back office: physician compensation, the month-end close, and management reporting. Every dollar runs as versioned code that reconciles to source.
Operational Performance Improvement
Lean process engineering, labor and productivity, and throughput programs, led by a performance-improvement leader who built them at the nation's largest health system, AI-accelerated.
Value-Based Care & Risk Economics
CMS bundled-payment, ACO, and risk-contract economics, modeled directly from claims with a deterministic risk and acuity engine that sizes opportunity, risk-adjusts fairly, and targets the right partners.
Secure AI Enablement
HIPAA-eligible cloud AI inside your own walls, with authorized-view, no-copy access, so your team adopts modern AI without stalling in security and compliance review.
Litigation & Fraud Analytics
Privilege-grade forensic analytics for outside counsel, with defensible, reproducible methods engineered to withstand a Daubert challenge.
Built to be challenged
Most AI in healthcare and in litigation fails on the same point: a CFO cannot underwrite it, a compliance committee cannot clear it, and an expert cannot defend it under cross-examination, because it is a black box. We build the AI that passes that test.
- The LLM reads and writes. The decisions run as deterministic, versioned code.
- Documented assumptions, frozen reference tables, and false-positive validation.
- Auditability and human-in-the-loop are now a buying gate, and what survives a Daubert challenge.
- Senior people do the work, with a credentialed core team and a vetted specialist network.
The glass box is not a metaphor. It is something you open.
Three named capabilities ship with our work; the screens are the working demos, running on synthetic data. Each is built per engagement, delivered into your environment, and owned by you.
Provenance
The audit surface
Every model we build ships with Provenance: a surface where your team watches each run, inspects every check, and traces every number back to source. Your team operates it and owns it.
Signal Forward Insights
The reporting layer
One source of truth, many lenses. Role-specific views computed directly from your source data, scoped to the reader, with every access audited. The AI writes the story; the engine computes every number.
Signal Forward Groundwork
The data foundation
The reporting is only as good as the foundation under it. Groundwork is the governed data foundation we build in your cloud: landed, conformed, and computable, with every pipeline run validated and recorded. Built on your platform. Owned by you.
Groundwork lays the foundation, Insights reads it, Provenance proves it, and that is why the AI on top can be trusted. See how they compose →
Three ways in
Advisory Services
Principal-led strategy, fractional leadership (1-3 days/week), or focused consulting sprints (3-4 weeks) on the decisions that matter most.
Delivery Services
AI-accelerated, team-based execution and transformation with SLAs and principal oversight. Senior people do the work, no junior hand-offs.
AI Governance & Compliance
SOC 2 readiness, HIPAA-compliant AI deployment, model risk management, and governance frameworks for responsible adoption.
New to AI and not sure where to start? Our fixed-scope AI Strategy Workshop separates fund-now operational analytics from longer-horizon innovation and gives leadership a governance and ROI story to act on.
Results that hold up
Representative, anonymized results from client work. Some figures are modeled or small-sample; we are precise about that in the detail.
Plus a privilege-grade forensic methodology that risk-tiered a large provider population for outside counsel and held its tiering through a post-delivery refinement, and a deterministic E&M rule engine that fixed 29% of baseline rule-application errors.
From the decision to a model you own
Frame
Understand the decision that has to hold up, and who will challenge it.
Build
Use AI at build time; ship deterministic, auditable models and rules.
Validate
False-positive checks, within-dataset calibration, documented methodology.
Transfer
You own the model. We document, train, and hand off.
Validate and Transfer happen in the open. Through Provenance, your team sees the model run, confirms each check, and takes ownership of a model it can keep auditing long after we hand it off.
Trusted by
From Fortune 5 to rural critical-access
About Daniel Heacock
Founder and Principal of Signal Forward. I started the firm to do what its name says: cut through the noise in healthcare data to find the signal that moves a decision, then make sure it holds up where it counts.
Before Signal Forward I led global analytics at Optum and worked in HCA Healthcare's Clinical Services Group. We build AI you can audit: the model reads and writes, and the money and risk decisions run as deterministic, auditable code that a CFO can underwrite and an expert can defend.
Senior people do the work here, backed by a credentialed core team and a vetted specialist network. From the nation's largest health system to rural critical-access hospitals.
“Better data, better decisions, better outcomes. Everything else is just noise.”
Ready to deploy AI you can actually defend?
Tell us about the decision you need to automate, and who has to be able to trust it. We will review and reach out within 24 hours.
Advisory, delivery, AI governance, or a fixed-scope AI Strategy Workshop, tailored to your stage and goals.