A single missing source can cause dozens of escalations
The system does not show unanswered questions as separate incidents. It groups them by topic, enabling the team to improve the knowledge base once for all channels.
Guests ask about car chargers
Over the course of a month, AI failed 17 times to find approved information about the number of chargers, the price and how to activate the service.
Topic ready for an update
The module groups questions, displays representative quotes and indicates what information is needed to close the gap.
- frequency and channels
- source conversations and reason for uncertainty
- owner and update status
Real conversations set knowledge priorities
A gap is identified from shortcomings visible in guest support, not just from a manually maintained list of topics. The module organises work on sources but does not create new facts itself.
Detects uncertainty
It collects questions that ended in an escalation, rejection of a suggestion or an answer based on an insufficient source.
Groups similar questions
It combines different phrasings of the same need and shows its scale across channels and over time.
Guides improvement
It allows an owner, status and the information needed to approve the knowledge to be assigned.
How does an unanswered question become a task?
Detecting the signal
The system records a missing answer, low confidence, manual takeover or rejected suggestion.
Group the topic
Similar questions from chat, messaging apps, email and telephony are assigned to a shared category.
Set the priority
Frequency, guest impact and the number of escalations help determine the order of work.
Complete and approve
The owner adds a verified source, and the hotel approves it before AI uses it.
Every gap leads back to the conversations that revealed it
The list records the channel, time, quote and reason for the flag. This helps the team understand the real context before adding new information.
What does AI use?
- conversation histories from supported channels
- rejected suggestions and escalations
- confidence rating and sources used
- knowledge-owner structure
What does the team receive?
- grouped list of missing information
- priority, quotes and source channels
- task with an owner and status
A missing answer must not automatically become new knowledge
The module highlights what is needed, but a subject-matter expert provides the fact and approves the source. Only then can the information be added to AI responses.
The team merges or splits groups, changes priorities, assigns owners and approves every new piece of information.
The module does not fill in missing facts, publish responses, change sources without approval or treat every escalation as an AI error.
Knowledge grows from real guest needs
The team can see which information is most often missing in day-to-day guest service.
Work starts with the topics generating the most questions and escalations.
One approved addition improves responses across all connected channels.
Signals come from conversations; improvements go into the knowledge base
The foundation is the history and decisions from communication channels. The task ends in the knowledge management system, which AI tools can use later.
Related process elements
Frequently asked questions
What counts as a knowledge gap?
A question without a reliable answer, a rejected answer, an escalation caused by a missing source or a recurring topic that needs clarification.
Are similar questions combined?
Yes. The module groups different wordings of the same need while keeping access to the source conversations.
Does AI fill in missing information by itself?
No. It can identify what is missing, but the fact and source must be provided and approved by the hotel.
How is priority determined?
It can take frequency, the number of channels, impact on guest service and the team’s manual assessment into account.