How to select operations that can utilize Robo Claw and candidates for introduction in Travel/TourismSTARTUP
In terms of the impact on customer safety and rights, whether the AI makes a final decision, whether there is personal, reservation, itinerary and location information, whether there is a reservation, price and inventory update, whether there is external transmission to customers, whether there is money and refund processing, whether human approval is required, whether there is a need to escalate to a safety officer, ease of suspension, correction and rollback, and whether it can continue to operate with a small number of people, we recommend that priority be given to routine tasks that can continue to operate with a small number of people centered on reading and drafting.
When Travel/TourismSTARTUP selects operations for Robo Claw implementation, the following criteria are considered: (1) The occurrence frequency is high, (2) the procedures are standardized, and (3) Impact on customer safety and Prioritize operations such as: ④ AI does not make final decisions such as reservation confirmation, fare confirmation, refunds, and departure decisions; ⑤ handling of personal, reservation, itinerary, and location information is organized; ⑥ approval can be assumed for external transmission to customers; ⑦ operations can be continued even with a small number of people. The more tasks that meet these requirements, the easier it will be to verify the effectiveness of Pilot.
Who This Is For
Who This Is For
For managers, bookers, CS leaders, and others who are considering where to start with Robo Claw.
What You'll Decide
What You'll Decide in This Step
In this Discover step, you will identify and prioritize the 1 tasks that Robo Claw will cover first. Detailed design of requirements, permissions, and approvals will be done in the following Refine step.
Fit Check
Suitable work／Unsuitable work
Suitable work
- Frequently occurring tasks with standardized procedures, such as primary classification of reservation inquiries and organization of word-of-mouth trends
- A task that makes it easy for humans to check and modify the first draft of judgments, such as FAQ search and drafting of multilingual guides
- Work that cross-checks multiple SaaS (reservation management, PMS, experience booking platform, etc.) information
- Even if the reservation, price, and inventory information are updated, work that requires approval before updating
- Work that does not involve customer safety or financial impact, but only organizes and drafts information
Unsuitable work
- Operations that occur extremely infrequently and are unlikely to have the effect of automation
- Tasks that require immediate booking confirmation, fare confirmation, and refund without approval
- Work that attempts to delegate final judgment to AI, such as decision making in the event of a disaster or bad weather, and confirmation of passport or visa requirements
- Work that is unorganized and contains a lot of sensitive personal information, itinerary, and location information
Industry Challenges
Challenges specific to travel/TourismSTARTUP
Many travel/TourismSTARTUPs face the following challenges when choosing a target business:
Too much candidate work to narrow it down
There are many options for booking, itinerary creation, word-of-mouth analysis, and multilingual guidance, and you can't decide where to start.
No hypothesis of effect can be made
The actual workload and man-hours of part-time members are not visualized, and it is difficult to estimate the introduction effect in advance.
I don't know if I can connect to the booking management/PMS/experience booking platform
It is not possible to determine if the reservation system you are using supports API integration or if there is no dedicated person in charge.
Lost in handling customer safety-related work
There are no criteria for deciding whether it's safe to use itineraries or work related to disaster response.
Method
Implementation Steps
1. Identify candidate work
Inventory tasks related to booking correspondence, CS, and product planning, such as reservation inquiry classification, itinerary suggestion creation, word-of-mouth aggregation, and multilingual guidance.
2. Check the workload and frequency
Check the frequency of each operation (regular, daily, weekly, etc.) and the man-hours required to respond.
3. Check the impact on customer safety and rights
Check whether the covered business involves decisions that affect the safety and rights of customers, such as booking confirmation, pricing confirmation, refunds, and implementation decisions.
4. Confirm the handling of personal, reservation, itinerary and location information
Check whether the target business handles the customer's personal information, booking information, itinerary information, and location information.
5. Check the availability of the booking management/PMS/Experience booking platform
Confirm the location of the data required by the target business and whether the target system can be linked.
6. Check whether there is external transmission for customers
Check whether the target operation is accompanied by a notification/email to the customer, and if so, whether it can be premised on approval.
7. Confirm whether it is possible to continue operation with a small number of people
Check whether the approval and review system can continue even for part-time members.
8. Prioritize
List the evaluation results and decide which region/facility or experience product/business to start with first.
Evaluation Table
Candidate Job Evaluation Form
Below is an example of an evaluation sheet. Please replace it with your own Candidate Jobs. The closer the job is to ◎, the higher the Discover priority.
| Candidate Jobs | Frequency | Stereotypical | Impact on customer safety and rights | System Connection | Write |
|---|---|---|---|---|---|
| Primary classification of booking inquiries | By Day | High | None | Booking Package | None |
| Review trend analysis | Weekly | High | None | Reservation site/social media | None |
| Draft multilingual guidance | at any time | High | Low | CMS/FAQ | None |
| Help creating itinerary suggestions | at any time | Medium | Medium (human verification required) | Booking Package/Map Services | None |
Data & Systems
Data and Systems Used
Discover step will review the status of the data system to assess candidates for the job, including:
Human-in-the-loop
Where Human Approval Is Required
Although it will not be implemented at the time of Discover STEP, when evaluating Candidate Jobs, the following judgments will always be made by humans.
- Final decision on whether to select the target task as a candidate for automation
- Determine whether to include work that involves booking, pricing, and inventory updates
- Determine whether to include tasks dealing with personal, itinerary, and location information as candidates
Measurement
KPI Candidates
In Discover STEP, the following is recorded as a hypothesis for each Candidate Job and will be used as verification material in the subsequent STEP.
Current Response Effort
Estimated human effort currently required for Candidate Jobs
Effect hypothesis
Hypothesis of man-hours and hours that can be expected to be reduced when automated
Deployment Priority Score
Priority calculated from Frequency, Stereotypical, Impact on customer safety and rights, etc.
Pitfalls
Common Pitfalls
Select only tasks where the effects are easy to see
If you only select tasks that stand out without considering frequency and exception handling load, it will be difficult to verify the effectiveness of Pilot.
Postponing Booking Package/PMS connection confirmation
If you decide on Candidate Jobs and then check whether the connection is possible, you will have to rework in the later steps.
Downplaying the impact on customer safety
If you select tasks without evaluating the magnitude of the impact, you will need to redesign the approval flow in subsequent processes.
FAQ
Frequently Asked Questions
What is the right number of tasks to start with?
In most cases, we recommend starting with one region, one facility, one experience product, or one business. When multiple tasks are carried out at the same time, authority design and pilot evaluation become complicated.
Should work related to customer safety be excluded from candidates?
You don't necessarily have to remove it. By automating the process of organizing information and presenting confirmation items, and creating a design in which the final decision is made by the safety officer, it is possible to treat them as candidates.
What should I do if I don't know whether I can connect to Booking Package/PMS?
Please check whether the API/data linkage of the Booking Package/PMS/Experience booking platform you are using is possible. If you do not have a dedicated person in charge, we can also provide individual consultation.
Who should evaluate Candidate Jobs?
We recommend that the management, reservation manager, and CS manager do this together. It is necessary to evaluate based on both the actual business situation and system constraints.
Continue
Next step
Once the target business has been determined, the next step is to specify the requirements, authority, and approval design for that business.
Would you like to help us organize the selection of target operations?
Based on your current business volume, frequency, and Booking Package/PMS configuration, you can discuss Robo Claw application candidates and priorities with the official LP.