The Predicates
Nine Questions That Should Always Have Answers
Each predicate evaluates a structural relationship between organizational domains. They return met, partial, or gap — and when something degrades, advisories fire automatically.
P_deal_promise Sales × Product Deals promise features that are actually being built
P_demand Sales × Product Feature demand from opportunities matches roadmap priority
P_claim Sales × Product Sales claims match actual product capabilities
P_revenue_attribution Finance × Product Revenue can be traced back to specific product features
P_skills_alignment HR × Product Team skills match the requirements of assigned features
P_cost_attribution Finance × HR True loaded cost per feature including time, expenses, and ROI
P_agent_pricing_coherence Agents × Sales × Finance Sales agent pricing aligns with finance budget targets
P_agent_capacity_coherence Agents × HR × Product Hiring pipeline matches product capacity needs
P_agent_delegation_alignment Agents × Governance Every business agent has appropriate human oversight
Connectors
Pull Data From Where It Lives
Coherence connects to your existing systems via webhooks and polling adapters. Events flow into the graph incrementally — no batch imports, no stale snapshots.
HubSpot
Deals, contacts, opportunities, quotes
QuickBooks Online
Invoices, expenses, revenue lines, budgets
BambooHR
Employees, teams, roles, time entries
Salesforce
Opportunities, accounts, contacts
Advanced
Guardrails and Remediation
Shadow Graph Evaluation
Before committing an action, evaluate it on an isolated copy of the graph. See which predicates would degrade — then decide whether to proceed, modify, or cancel.
Adaptive Remediation
When a check fails, 22 YAML-defined playbooks generate concrete alternative events. Each is verified on a fresh shadow graph before being suggested. Cross-predicate collateral damage is checked automatically.
Named Scenarios
Persistent what-if workspaces. Create a scenario, add events, evaluate, fork, and compare alternatives side by side before any real changes are made.
Effectiveness Tracking
The feedback loop: which playbooks are adopted, which actually restore predicates, and how long resolution takes. Calibrate thresholds with data, not guesswork.