Trigger automation
When this happens, do that
- Starts from a technical event
- Automates a local task or integration
- Often assumes context and data quality
- Explains execution, not always business authority
Executable enterprise policy
Model how your business may operate. Enforce it on every message.
A DataInbox rule is not a trigger-action shortcut. It is part of an executable operating model that connects business meaning, context, identity, purpose, risk, permissions, decisions, actions, and evidence. AI can help author the model. The deterministic runtime remains authoritative.
- Business and domain models
- Security, data and AI policy
- Governed state and decisions
- Validated configuration changes
A different class of rule engine
Not IFTTT. Executable enterprise modelling.
Simple automation asks what should happen after a trigger. DataInbox first asks what the message means, which operating context applies, who has authority, which policy permits the next step, and what evidence must remain.
DataInbox policy model
Given this business context, what is permitted now?
- Starts from a validated business message
- Evaluates domain state, identity, purpose, and risk
- Constrains people, systems, agents, and Companions
- Produces an accountable decision and traceable result
The operating model
Rules express how the enterprise works
Routing, validation, and approval are outcomes of a richer model. The model can describe the business object, its current state, applicable policy, permitted actors, available capabilities, and the evidence required after execution.
Message and domain model
Define what a customer, order, payment, case, decision, or document means, which fields are required, and which business invariants must hold.
Identity and authority
Determine which person, system, agent, or Companion may see context, propose a decision, approve it, or perform an action.
Purpose and permitted use
Connect data use to purpose, consent, classification, minimisation, retention, deletion, and the conditions that apply to every operation.
State and decision logic
Model valid states, transitions, exceptions, approvals, escalations, deadlines, and the next action that is allowed in the current context.
Risk and regulation
Translate applicable security, data, AI, and sector requirements into controls that can be evaluated when a business message is processed.
Outcome and evidence
Record the input, applicable policy, decision, approval, action, result, and exception as traceable business-message evidence.
Policy becomes operation
Turn complex requirements into controls the runtime can evaluate
Regulations such as GDPR, NIS2, and the EU AI Act do not become one universal checkbox. Their applicable requirements, together with internal security and data policies, must be translated into explicit controls for each operating context.
DataInbox can support enforcement and evidence. Compliance still depends on your organisation, role, use case, risk classification, configuration, contracts, and operating procedures.
Security policy
Classify a message, verify identity and scope, restrict sensitive fields, require stronger approval, and reject a capability that is not permitted.
Data policy
Evaluate source, purpose, consent, minimisation, residency, retention, and deletion before data is exposed or used for another operation.
AI policy
Select eligible AI Companions, constrain the context they receive, define the output contract, require human oversight, and validate the result before use.
Operational policy
Validate business state, route the message, request approval, invoke an allowed capability, handle an exception, and capture the resulting state change.
AI-assisted authoring
Let AI do the modelling work without giving it the keys
An AI agent can analyse policy documents, ask for missing decisions, inspect the supported configuration model, and draft complex Inbox rules. The DataInbox CLI then provides a controlled path from proposal to approved configuration.
01
Interpret the requirement
Business experts, legal teams, security teams, and engineers define the desired operating model. AI can help turn policies and regulation into explicit requirements and testable decisions.
02
Draft against the specification
An AI agent can inspect the DataInbox configuration specification and the current Inbox configuration, then propose structured changes without inventing unsupported fields.
03
Validate and compare
The CLI validates the candidate configuration and produces a diff. A failed validation or an unexpected change stops the process before anything is applied.
04
Dry-run and approve
The proposed configuration is dry-run first. The responsible reviewer can inspect scope and impact before an approved change reaches the Inbox.
05
Enforce and evidence
The deterministic runtime evaluates the approved model for each applicable message and records decisions, actions, outcomes, and exceptions for review.
Deterministic at runtime
One governed decision path for people, systems, and agents
Once approved, the operating model is evaluated when a business message arrives. AI may interpret or propose inside its permitted scope, while rules decide which context is available, which result is acceptable, and which next action may occur.
Business message
Contract and context
Applicable policy
Permitted decision
Result and evidence