A risky comment does not receive a routine suggestion
The filter stops the standard process, explains the risk and passes the full context to the appropriate person.
Complaint with a compensation request
The guest describes an incident, publishes an employee's personal data and threatens legal action if the hotel does not pay compensation.
Issue blocked and escalated
The system shows the risk categories, highlights relevant passages and directs the issue to a member of staff without preparing a routine publication.
- visible risk level and type
- specific passages triggering the block
- recommended escalation route
A conservative block where human judgement is needed
The filter is designed to protect the brand from an uncontrolled response. A doubtful issue may be passed to a member of staff even when its category cannot be determined conclusively.
Recognises risk
Detects legal, safety, compensation, privacy and uncertain-fact issues.
Stops routine processing
Stops a routine suggestion or automation when an issue falls outside the approved scope.
Explains the escalation
Shows the reason, relevant passages and the suggested owner of the next decision.
How does the filter separate routine issues from exceptions?
Reading the comment
AI analyses the content, source, thread history and available context.
Checking the rules
It compares the issue with the risk catalogue and the hotel's relevant exceptions.
Classification
It assigns a risk level and highlights the passages supporting the result.
Blocking or escalating
A routine issue can proceed, while a risky one is passed to a member of staff.
The reason for the classification must be visible
The system should not hide its decision behind a generic label. The user sees the category, the relevant passage and the rule applied.
What does AI use?
- review or comment text
- public thread history
- risk and escalation rules catalogue
- optional hotel-specific exceptions
What does the team receive?
- risk level and category
- reason for the block with source passages
- suggested route for escalating the issue
Uncertainty favours human oversight
A falsely safe classification is riskier than an additional review. Crises and decision-critical issues always remain with people.
The team approves the risk and exception catalogue, can change the classification and always takes over legal, safety, compensation and uncertain-fact issues.
The filter does not settle disputes, provide legal advice, award compensation, publish responses or treat its classification as the final decision.
A safe boundary between routine issues and exceptions
Each issue receives a level, category and verifiable rationale.
High-risk content does not proceed to routine publication.
The issue goes to someone who can verify the facts and make a decision.
The filter works between the reputation panel and escalation
The Central Reputation Panel supplies the comment, the AI Platform applies the rules, and the context interview can add hotel-specific exceptions.
Related process elements
Frequently asked questions
Which issues should always go to a member of staff?
Legal, safety, personal-data, compensation and uncertain-fact issues, along with any other exceptions set by the hotel.
Can the filter flag spam or trolling?
Yes, but the decision on moderation or a response remains with the user.
Can the user see the reason for the classification?
Yes. The system shows the category, rule and relevant passage.
Is the risk flag final?
No. A member of staff can change it, and the team always makes decisions on high-risk issues.