The foundation everything
is built on.
Every Avennorth product runs on a shared fabric of data, integrations, and AI reasoning. The platform is not a product you configure — it is the infrastructure that makes everything else trustworthy.
ARCHITECTURAL COMPARISON
Not ServiceNow with a new name.
Four architectural decisions we made differently — and why they matter when your platform needs to survive an upgrade.
Every CI class, relationship type, and data entity has its own table with an explicit schema. No shared columns. No polymorphic extensions. Schema is always self-documenting and query-optimal.
The Task table underpins 40+ OOTB entity types via table inheritance. Sparse columns, implicit schemas, and extension tables that grow unbounded over years of customization.
Customer and partner extensions live in Ring 3–4. Platform core lives in Ring 0. An upgrade only touches Ring 0–1 — Ring 3–4 extensions survive unchanged, by architectural constraint, not by policy.
Platform customizations often modify core tables, ACLs, and business rules — effectively Ring 0 — so every upgrade becomes a manual merge exercise. Frozen instances are a predictable outcome.
Every AI output carries a cryptographic B1 receipt. input_hash → output_hash means every recommendation is deterministically replayable. C1–C10 validation checks include NLI falsifier-entailment (C10) — advice must survive semantic refutation before surfacing.
AI outputs carry confidence scores from the model. No cryptographic receipt, no deterministic replay, no structural requirement that an AI output state what would make it wrong.
Pathfinder eBPF agents observe kernel-level behavior continuously. The CMDB is a projection of live attestations — not a table someone maintains. Discovery runs on a ~60MB Go collector, agentless scan available.
400MB+ JVM-based MID Servers run scheduled scans using stored credentials. Discovery reflects what was reachable at scan time — the CMDB decays between scan cycles.
ARCHITECTURE
One platform. Eight products. Zero seams.
Products sit above the platform layer. They draw from the same data fabric, the same audit chain, and the same intelligence engine. There is no integration overhead between them.
CORE INFRASTRUCTURE
Six structural properties.
These are not features you configure — they are properties of the platform itself. Every product that runs on GPS inherits all of them automatically.
Bitemporal Data Fabric
Every CI, relationship, and attribute carries two time dimensions: when it was true in the real world, and when the system recorded it. No CMDB state is ever overwritten — only appended. Roll back to any point in history without losing current state.
Patent-Protected ServiceNow Integration
Pre-built connectors that write directly into ServiceNow's CSDM-aligned schema. No custom middleware, no ETL pipelines, no data quality degradation in transit. Integration logic is protected by pending patents.
Falsifier-Mandatory AI Framework
No AI assertion reaches the surface unless it carries a defined falsification condition. Every claim is paired with the test that could refute it. This is structural — not a guardrail bolted on top.
Immutable Audit Chain
Every write operation is hash-chained to the previous record. Tampering with historical data breaks the chain — audit integrity is cryptographically verifiable, not policy-asserted.
Zero-Credential Vault
Pathfinder's agentless discovery mode requires no stored credentials. Where agents are deployed, credentials are ephemeral and never persisted to disk.
Continuous Refresh Architecture
The platform does not run discovery once. Every change in the environment surfaces as a delta — propagating through service maps, health scores, and incident context in real time.
UPGRADE SAFETY
Ring 0–4 isolation.
Customization without the freeze.
The ServiceNow frozen-instance problem has one root cause: customers eventually modify core platform objects — tables, ACLs, business rules — that an upgrade needs to replace. When those objects are customized, the upgrade breaks them. The solution isn't better upgrade instructions. It's a platform architecture where Ring 0 is structurally unreachable by anything outside Avennorth.
GPS enforces five concentric isolation rings. Customer extensions live in Rings 3–4. Upgrades touch Rings 0–1 only. Ring 3–4 extensions survive every upgrade unchanged — not because of a policy, but because the architecture makes contact impossible.
No frozen instances. No upgrade negotiation windows. No "we can't upgrade yet because of customization X." Customer-built extensions are upgrade-invisible by construction.
Data fabric, audit chain, AI engine, identity layer, integration primitives. Never modified by customers or partners. Upgraded by Avennorth autonomously.
Pathfinder, Bearing, Contour, Assets, Vantage, Architect, Trust, Compass — all eight products. Avennorth-authored. Upgraded with Ring 0.
Partner-built modules distributed through the GPS marketplace. Upgrade-tested against Ring 0 as a certification gate — partners cannot publish without passing.
Customer-built using GPS Builders: Table, Workflow, Report, Data Loader, UI, Agent, Integration/Plexus, Micro-Model. Upgrade-compatible by design — Builders emit Ring 3 artifacts only.
Customer scripting and advanced extensions outside the Builder layer. Explicitly isolated and upgrade-gated. The GPS upgrade process flags Ring 4 artifacts and runs compatibility checks before proceeding.
AI PRIMITIVES
Provenance-first AI.
Not confidence scores — receipts.
Every AI output in GPS is cryptographically receipted, structurally refutable, and deterministically replayable. These are platform primitives, not product-level guardrails.
B1 Cryptographic Receipt
Every AI output is stamped with a B1 receipt — a cryptographic record of what input produced what output. The receipt is written to the immutable audit chain. The chain cannot be revised without detection.
C1–C10 Validation Checks
Ten graduated validation checks run on every AI output before surfacing. C10 is NLI falsifier-entailment: a separate model reads the output and checks whether it is semantically refuted by the evidence it cites.
Confidence Banding
Output confidence is mapped to three action bands: auto-trust (surfaces immediately), confirm-required (held for practitioner review), or block (rejected and logged). The band thresholds are configurable per product and per context.
Deterministic Replay
input_hash → output_hash. Given the same input state, GPS AI produces the same output — verifiably. Recommendations are replayable weeks or months later for audit, incident review, or regulatory inspection.
GPS
Intelligence
Most platforms give you answers. GPS Intelligence gives you answers you can defend.
Intelligence is the reasoning layer underneath every product. When Bearing scores a CI, when Vantage diagnoses an incident, when Compass generates a brief — Intelligence is what separates a confident assertion from a falsifiable claim.
It ships with 40+ day-one indicators grounded in live discovery data across four architectural domains: Business Applications, Capabilities, Services, and Offerings. No surveys. No manual configuration. Indicators are active the moment Pathfinder runs.
Every question produces multiple competing explanations simultaneously. No single hypothesis is favoured from the start.
Each hypothesis is paired with conditions that would eliminate it. Those conditions are evaluated against live platform data.
Only hypotheses that pass every falsification check reach the surface — with confidence band, assumed facts, and data lineage.
Organizational
Intelligence
Org charts lie. The official structure is not how decisions get made or how work actually flows.
Organizational Intelligence builds three simultaneous structural models of how your organization actually operates — grounded in directory services, behavioral discovery, service maps, and engagement signals. It finds what the org chart cannot show: ghost influence, capability drift, single points of informal power, and maturity gaps against industry benchmarks.
The official org chart — roles, reporting lines, cost centers, and business units ingested from directory services. The baseline. It is never treated as the complete picture.
How work actually moves. Which teams interact with which systems. Which services span which business units. Derived from behavioral discovery data — not from what anyone declared in a spreadsheet.
Who actually influences decisions regardless of their title. Where authority is informal. Where capability sits outside the people formally responsible for it. The map most leaders don't have but every leader needs.
Ghost Influence
Individuals who carry significant informal influence over decisions or teams but hold no formal accountability. High retention risk. Invisible to succession planning. Org Intelligence surfaces them with evidence.
Capability Drift
Where the organisation's formal capability allocation has drifted from where work actually happens. Teams holding titles for capabilities they no longer exercise — or exercising capabilities they are not resourced for.
Single Points of Influence
Key person dependencies that do not appear on any org chart. Found by mapping who decision flows converge on — not who the org chart says they should converge on.
Maturity Gap Detection
Industry ontologies and capability frameworks (TOGAF, BIAN, eTOM) used to score each capability area against maturity benchmarks. Where the org is ahead. Where the gaps are. What the risk is if the gap persists.
Capability maps scored against industry benchmark frameworks. Identifies where the organisation is ahead, where gaps exist, and what the risk is if the gap persists without intervention.
Stakeholder
Hub
Relationships are managed by gut feel and memory — until something breaks.
Stakeholder Hub turns relationship and trust risk into a measured, evidence-backed signal. It surfaces declining engagement and overdue commitments before they become failures — so leaders can act on relationship risk instead of discovering it too late.
7-Dimension Trust Scoring
Each stakeholder relationship is scored across seven observable dimensions — not rated by gut feel. Scores are derived from evidence, not from surveys. The trust model is explicit and auditable.
Engagement Health Tracking
Going-cold signals surface automatically. A stakeholder who was previously active but is becoming less engaged triggers an early-warning flag — before the relationship deteriorates into a risk event.
Commitment Ledger
Every commitment is recorded with a defined evidence condition: what would prove it was kept or broken. Falsifier-mandatory — not a to-do list. Overdue commitments are surfaced, not forgotten.
Feedback Loops
Structured feedback from stakeholders is recorded, processed, and acted on within the evidence chain. Feedback that changes a trust score is traceable to its source. Nothing is informal.
Immutable Evidence Chain
Every trust score change, commitment update, and engagement event is written to an immutable chain (INV-stakeholder-engagement-001/002). Trust history cannot be revised — only appended with corrections that reference the prior record.
Vantage + Compass Integration
Stakeholder Hub feeds directly into Vantage stakeholder assembly (who gets paged for a P1, in what order) and Compass advisory routing (which executive receives which type of brief).
Measures interpersonal and professional relationship trust — engagement health, commitment fulfillment, and influence dynamics. Stakeholder-to-stakeholder. Distinct from system access or identity governance.
Measures system-level trust: identity governance, access rights, compliance posture, and audit fabric. Person-to-system and system-to-policy. Distinct from relationship scoring.
INTEGRATIONS
Works with your full stack.
The GPS Platform connects to your existing tooling — it does not require you to replace it. Discovery signals, monitoring data, identity records, and collaboration channels all flow into the platform fabric. Select a category to explore.
BUILT ON THE PLATFORM
Eight products. One fabric.
Closed-loop discovery that populates the fabric
Learn more →Health scoring across every CI in the fabric
Learn more →Service mapping built from discovered relationships
Learn more →Hardware and software asset lifecycle from discovered CI data
Learn more →Incident intelligence drawn from live fabric state
Learn more →Solution design and CSDM compliance validated against the fabric
Learn more →Compliance, governance, and audit fabric for regulated environments
Learn more →Executive intelligence filtered through the fabric
Learn more →The AI reasoning layer that powers all products
Learn more →Ready to see the platform in action?
We deliver a fully operational, trusted CMDB in 90 days. Book a demo and we will show you what your environment looks like in the platform.