CLAIRE Appliance Catalog · Leadership Brief

Three Appliances.
One Engine. One Delivery Model.

Three different pharma pain points. One platform under the hood. Three separately-salable products — and one repeatable delivery model for regulated customers who refuse to put their data in our cloud.

ComplianceG Labs · Internal Strategy Document · 2026
Part 1 · What's Different

Three appliances, three different buyers, three different revenue streams.

Each appliance targets a distinct regulated-industry problem with a distinct buyer, cadence, and pricing motion. They are independently salable.

Dimension MAiGRATE Data Migration ATR Audit Trail Review APQR Annual Product Quality Review
Problem solved Migrate legacy quality data into a modern target system Continuously review audit trails for GxP compliance Compile the Annual Product Quality Review regulatory report
What it replaces $1–3M consultancy migration project Manual quarterly compliance reviews + GRC audit prep ~1 month of manual work per product, per year
Primary buyer Head of Quality + Migration PM CISO + Head of Compliance + Internal Audit Head of Manufacturing Quality · QA Director
Data flow Read source, write target Read only Read only
Cadence One-time event Continuous (weekly / monthly) Periodic (annual, per product)
Source systems One legacy system (TrackWise, MasterControl, etc.) One or a few (Veeva Vault, Jira, MES) Four or more (TrackWise + JMP + ERP + SharePoint)
Primary output Migrated records + per-record evidence pack Compliance report per cycle with severity-ranked findings Formal APQR document (Word / PDF) per product
Knowledge Graph role Optional Central — users, roles, training, SOPs Load-bearing — product → batch → deviation → CAPA
Current pipeline Vericel · Vantive · Integer Active sales conversations Vertex · Regeneron Demo engagements Mankind PoC in preparation
ROI story ~50% cheaper than consultancies; defensible audit trail Lower cost than quarterly audits; catches what humans miss 1 month → 1 day. 90% time reduction.
Default pricing model One-time license + services (~$450K–$830K) Annual subscription per system Annual subscription per product
Sales motion Project-driven; procurement-heavy Ops-driven; continuous-compliance narrative Quality-driven; obvious-ROI narrative
Strategic role Landing wedge — lands the account First expansion — recurring revenue Largest expansion — scales with product catalog
Part 2 · How We Deploy & Deliver

The same deployment model and engagement shape for every appliance.

Regardless of which appliance the customer buys, it runs in their environment, on their network, with their AI, under their identity system. And every engagement follows the same six delivery steps — what changes per customer is the configuration, not the code.

Deployment
Inside the customer's environment — always.
Source
System
→
MAiGRATE · ATR · APQR Docker stack on one VM
→
Target
System
↕
Customer's
LLM
Customer's
Identity Provider
Their network Runs inside customer's data center
Their AI Azure OpenAI, AWS Bedrock, or self-hosted
Their identity SAML into their Azure AD / Okta
Their infrastructure Docker on one VM — 8–16 vCPU, 32–64 GB RAM
Delivery Approach
Same six-step engagement, every customer.
01
Install
Docker stack deployed into their environment. SAML wired to their identity provider. Credentials placed in their secrets vault.
02
Configure
Solution engineer workshops with their QA team. Mapping rules, lookups, validation rules, terminology guide. No code is written.
03
Dry-run
Sample execution on a small batch. Subject-matter experts review every AI decision. Rules tightened based on feedback.
04
Test environment run
Full execution against the customer's test tenant. QA validation and sign-off before any production activity.
05
Production run
MAiGRATE: phased batch cutover. ATR: ongoing review cycles. APQR: periodic report compilation. Every record logged in an evidence pack.
06
Hand off
Customer staff trained. Evidence packs delivered. Clean handoff with no ongoing vendor dependency.
Part 3 · What's Shared

Twelve platform capabilities — inherited from CLAIRE, identical across all three appliances.

These are not per-appliance features. They come from the CLAIRE engine itself. Every appliance we ship inherits all twelve for free. The fourth, fifth, and sixth appliance will too.

Platform capability MAiGRATE ATR APQR What it gives you
Home Dashboard / Command Center ✓✓✓ Realm-scoped home with activity feed, health indicators, pending approvals, per-entity drilldowns
Execution Observability ✓✓✓ Real-time and historical flow execution logs, per-step timing, errors, retries, outputs
Full Audit Trail ✓✓✓ Every data mutation logged and traceable to user + timestamp
LLM Audit Logger ✓✓✓ Every AI call logged with prompt, response, model, cost, latency — feeds the per-record evidence pack
Trend Analysis ✓✓✓ Historical charts, period-over-period comparisons, anomaly trending across any metric
Web UI / Frontend ✓✓✓ React app, entity detail pages, wizards, theme, graph visualization — one UI, all three appliances
Role-Based Access Control ✓✓✓ Realm-scoped roles (OWNER / ADMIN / WRITER / OPERATOR / READER), per-entity ACLs, runtime enforcement
SAML SSO + Identity Provider ✓✓✓ Wires into customer's Azure AD / Okta — same login and offboarding as their other systems
Workflow Orchestration ✓✓✓ Visual flow builder, human-in-loop gates, conditional branches, error handling, retries
Governance Workflow ✓✓✓ Draft → review → approve → publish cascade with version history, dependency graph, impact analysis
Human-in-Loop Approval Gates ✓✓✓ Any step in any flow can pause for human review and sign-off before continuing
Knowledge Graph + Ontology ✓✓✓ Neo4j-backed, D3 force graph visualization, ontology-driven RBAC on node types
The Point
A fourth, fifth, or sixth appliance doesn't rebuild any of this. It inherits it. Every new domain pack is configuration against an existing platform — not a new product build.
Same engine. Same delivery playbook. Same security model.
Three different buyers. Three different revenue streams. One platform investment.
Land with MAiGRATE → Expand with ATR → Grow with APQR