- Who is this for?
- Operations leaders, transport operators, transformation managers and IT teams
- Question addressed
- Understand how Vigilo learns a workflow without silently turning every human correction into an automated rule.
Key takeaways
- An observed decision starts as a contextual record, not a new permission.
- Repetition is meaningful only when the records, constraints and expected outcomes are comparable.
- Every candidate rule must specify its sources, scope, exceptions and owner.
- Adaptation is managed as a configurable, testable and versioned product capability.
- A change in tools, screens or working practices requires validation of the affected workflow, not a claim of universal compatibility.
A real workflow never fits entirely inside a procedure
Written instructions, contracts, rate matrices and standard operating procedures provide a valuable foundation. They describe approved sources, expected checks and decisions that must not be automated. Yet part of the work still lives in the team’s experience: calling a contact before a sensitive amendment, requesting approval after a cut-off time or selecting a slot in response to a local constraint.
Vigilo is designed to represent this reality without reducing a company to a standard scenario. Within the agreed scope, it observes events and decisions, links each choice to the relevant record and makes recurring practices explicit. The goal is not to imitate every habit, but to distinguish stable rules from exceptions and decisions that must remain human.
What Vigilo retains when a decision is made
A correction in isolation does not explain why it was right. To be useful, the decision must remain connected to the situation that prompted it and to the outcome that was subsequently checked.
- the initial event, its channel, timestamp and source;
- the versions of the record and documents that were consulted;
- the constraints known when the decision was made;
- the selected option, its rationale and the person or mechanism that approved it;
- the action prepared or executed in the business system;
- the outcome actually observed and any subsequent corrections.
From repeated decisions to a candidate rule
- 01Observe
The human decision is retained with the record, its sources and the verified outcome.
- 02Compare
Vigilo looks for comparable situations rather than a superficial resemblance between words or gestures.
- 03Identify recurrence
The same decision appears in a sufficiently stable context to warrant review.
- 04Formulate a candidate rule
The proposal describes its conditions, supporting examples, excluded cases and intended authorization level.
- 05Obtain approval
The team confirms the business meaning, adjusts the scope and decides whether the rule remains advisory, supervised or active.
- 06Version and monitor
Every change remains attributable to a version, with an owner, date and suspension mechanism.
Example: a request received after the cut-off time
Several orders received after the cut-off time are sent for approval before creation. The action looks identical, but the reasons may differ: remaining capacity, type of goods, customer commitments or a slot that has already been confirmed. Vigilo retains these factors with each decision instead of immediately inferring a general rule.
If the same decision recurs across genuinely comparable records, Vigilo can propose a candidate rule: for this flow and customer scope, request approval after the defined cut-off. The team checks the exclusions, appoints an owner and selects the authorization level. An order outside that scope continues to be escalated to an operator.
Adaptation becomes a product asset, not a one-off project
Two companies can use the same TMS and still follow different operating methods. Conversely, two different systems may support a similar workflow. Vigilo’s adaptability therefore relies on stable building blocks — events, records, rules, actions, checks and escalations — configured around the company’s actual way of working.
Sources, identifiers, screens, action rights and rules remain specific to the deployed scope. Their configuration is tested and versioned. This approach allows the product foundation to be reused when a new flow or system is added, without claiming compatibility before access methods, screens, volumes and required controls have been assessed.
| Element | What can be reused | What must be adapted |
|---|---|---|
| Events | The observation and record-linking mechanism | Channels, formats and record identifiers |
| Rules | The proposal, approval, versioning and suspension lifecycle | Business conditions, responsibilities and exceptions |
| Actions | Controlled preparation, execution and verification | Functions available through an API or interface, and granted permissions |
| Supervision | Statuses, audit trails and stop reasons | Thresholds, recipients and authorization levels |
What the team validates before activation
This validation is not only a risk control. It turns tacit knowledge into transferable knowledge: a new team member can understand the rule, review the cases from which it emerged and know when not to apply it.
- the business purpose of the rule and the records that support it;
- the customers, flows, entities, systems and periods to which it applies;
- the explicitly excluded cases and situations considered sensitive;
- the source of truth when information conflicts;
- the permitted actions and the checks required after execution;
- the owner authorized to change, suspend or withdraw the rule.
When Vigilo must stop
Adaptability does not mean that Vigilo should process every case. Conflicting sources, a changed interface, missing mandatory data or a situation outside the validated scope must trigger an explicit stop. The record handed to the team then contains the event, reconstructed context, relevant rule and reason for stopping.
This behavior is part of the workflow. It allows autonomy to evolve on the basis of observable evidence without concealing areas where the organization, system or rule has changed.
Measured progression
- 01Observe without acting
Compare Vigilo’s reconstruction with the decisions made by the team.
- 02Propose
Present actions and candidate rules without changing business systems.
- 03Authorize a precise scope
Activate only the cases, systems and actions whose controls have been validated.
- 04Expand or reduce
Adjust coverage in response to outcomes, system changes and newly encountered cases.
Frequently asked questions
Direct answers
Does Vigilo learn every human correction automatically?
No. A correction remains a contextual record. A recurrence may produce a candidate rule, but only explicit approval can authorize its use within a defined scope.
Must every rule be documented before deployment begins?
No. Existing documentation speeds up scoping. Observation can also surface recurring practices, which must then be explained, bounded and validated.
Can Vigilo adapt to any TMS?
The foundation is designed to accommodate varied environments, through APIs or interface interaction where appropriate. Compatibility and reliability must still be validated for each workflow, access method and system version.
What happens when a working method changes?
The affected rule and configuration are reassessed, tested and versioned. They can be suspended until the new behavior has been confirmed.
Should a frequent decision always be automated?
No. It may remain advisory or supervised when its impact, sensitivity or dependence on context requires human judgment.
Sources and methodology
This resource describes Axonovia's product approach based on transport workflows. The examples are illustrative and do not constitute a performance commitment or a universally applicable contractual rule.
Product review: Axonovia team · next review recommended before 31 August 2027 or following a material change in scope.