In recent years, the travel industry has increasingly expressed interest in:<\/p>\n
Delegating customer inquiries to AI<\/p>\n<\/li>\n
Using chatbots to address staffing shortages<\/p>\n<\/li>\n<\/ul>\n
However, in practice, many organizations find that although the system is implemented, it is not fully utilized, and front-line staff ultimately have to cover the gaps<\/p>\n
This article examines the reasons why chatbot implementations are prone to failure, with consideration given to the specific characteristics of the travel industry<\/p>\n
Inquiries in the travel industry have the following characteristics:<\/p>\n
There are many combinations of dates, group sizes, cities, and flight numbers.<\/p>\n<\/li>\n
Relative expressions such as \u201clast time,\u201d \u201clast year,\u201d and \u201cthe other day\u201d are common.<\/p>\n<\/li>\n
Tour numbers and product names are often written ambiguously.<\/p>\n<\/li>\n<\/ul>\n
\nWhat happens if we take the same tour as last year, but with one more child this time?<\/pre>\n<\/blockquote>\nInquiries like this
cannot be framed as a standalone FAQ question.<\/strong><\/p>\nWith an FAQ-based chatbot:<\/p>\n
\n
- \n
The question may not match any FAQ.<\/p>\n<\/li>\n
- \n
It may return a similar answer instead.<\/p>\n<\/li>\n<\/ul>\n
These problems are likely to occur.<\/p>\n
\nReason for Failure (2)<\/h3>\n
Too Many Exceptions<\/h2>\n
In the travel industry, exceptions<\/strong> such as the following occur frequently:<\/p>\n
\n
- \n
Weather, natural disasters, and strikes<\/p>\n<\/li>\n
- \n
Sudden changes due to local circumstances<\/p>\n<\/li>\n
- \n
Cancellations that require individual judgment<\/p>\n<\/li>\n<\/ul>\n
These situations
are governed by rules, but the same rules do not necessarily apply every time<\/strong>
which means:<\/p>\nAs a result:<\/p>\n
\n
- \n
The more precisely the rules are written, the more they may diverge from reality.<\/p>\n<\/li>\n
- \n
Vague rules can lead to misunderstandings.<\/p>\n<\/li>\n<\/ul>\n
This creates a dilemma.<\/p>\n
\nReason for Failure (3)<\/h3>\n
Thinking \u201cAutomation Means AI Does Everything\u201d<\/h2>\n
One of the most common misconceptions about introducing chatbots is the idea that
**\u201cIf we leave it to AI, people will no longer be needed.\u201d**<\/p>\nIn reality:<\/p>\n
\n
- \n
Triage incoming inquiries<\/p>\n<\/li>\n
- \n
Provide initial information<\/p>\n<\/li>\n
- \n
Answer obvious, standard questions<\/p>\n<\/li>\n<\/ul>\n
These are areas suited to AI.
Responses that require judgment and accountability<\/strong> must be handled by people.<\/p>\nIf AI is introduced without drawing this line:<\/p>\n
\n
- \n
Staff may spend more time checking AI responses.<\/p>\n<\/li>\n
- \n
Responsibility in the event of a complaint may become unclear.<\/p>\n<\/li>\n<\/ul>\n
These are common outcomes.<\/p>\n
\nReason for Failure (4)<\/h3>\n
Knowledge \u201cGrows on Its Own\u201d<\/h2>\n
More systems now automatically read websites and PDFs
to build a knowledge base.<\/p>\nHowever, in the travel industry:<\/p>\n
\n
- \n
Outdated terms and itineraries<\/p>\n<\/li>\n
- \n
Multiple documents containing the same information<\/p>\n<\/li>\n
- \n
Limited-time notices<\/p>\n<\/li>\n
- \n
Options and wording that are no longer in use<\/p>\n<\/li>\n<\/ul>\n
These often coexist, so
deciding what the AI should learn is itself important.<\/strong><\/p>\nAutomating this process can create risks such as:<\/p>\n
\n
- \n
Not knowing when the information was current<\/p>\n<\/li>\n
- \n
The AI providing answers that staff are unaware of<\/p>\n<\/li>\n<\/ul>\n
These risks can arise.<\/p>\n
<\/h2>\n
How Should We Think About This?<\/h2>\n
When training AI for the travel industry,
**\u201cwhat it can do\u201d and \u201cwhat it should not do\u201d**
\nmust be clearly distinguished.<\/p>\nAreas Handled by AI<\/strong><\/h3>\n