Inquiry Translation Response Process, ChatGPT Integration

/ Workflow Apps / Inquiry Response, OpenAI API

Inquiry Translation Response Process, ChatGPT Integration 20231228
This process manages the inquiry form from reception to response. The enquiry text in English is automatically translated into Japanese and sent to the enquiry desk. The person in charge prepares a draft response in Japanese and uses the AI’s English translation as a reference.
Worker (#of Swimlane: 2)
  • Inquiry Desk Leader
    • Responsible for the inquiry response team.
  • Responder
    • Person in charge of responding to inquiries.
Business Flow (Complexity of Workflow: 18)
  • 61. Designate a responder
    • The person in charge of the inquiry response team will designate a responder.
  • 132. Response
    • The person responding to the inquiry enters the answer.
  • 193. Evaluation of response
    • The person responsible for the inquiry response team evaluates the response results.
  • 8x1. Subject
    • The workflow platform sets the subject.
  • 40x1’. Subject (no correspondence)
    • The workflow platform sets the subject line.
  • 43x2′. AI_Inquiry Translation（gpt-3.5）
    • The workflow platform translates the query into Japanese.
  • 33x2. AI_Inquiry Translation（gpt-4）
    • The workflow platform translates the query text into Japanese. (If the process times out, refer to gpt-3.5)
  • 9x3. Destination, response email text
    • The workflow platform sets the response email destination.
  • 44x4′. AI_Response Translation（gpt-3.5）
    • The workflow platform translates the draft response text into English.
  • 35x4. AI_Response Translation（gpt-4）
    • The workflow platform will translate the draft answer into English. (If processing times out, refer to gpt-3.5)
Business Process Variables (#of Data Item: 25)
  • ▼ Inquiry Source ▼ 0
  • Name q_name 1 *
  • Company Name q_company 2 *
  • Email Address q_email 3*
    • The email address of the inquirer is stored.
  • Phone Number q_tel 4
    • Stores the telephone number of the inquirer.
  • Previous Inquiries q_note 5
    • Stores notable past queries regarding the inquirer.
  • ▼ Inquiry Form ▼ 6
  • Received Date/Time q_receiveDatetime 7
    • The date and time the inquiry was received is stored.
    • processInstanceStartDatetime
  • Inquiry ID q_id 8
    • #{#sformat(‘%d’, processInstanceId)}#{#sformat(‘%04d’, processInstanceSequenceNumber)}
  • Message q_inquiryDetail 12*
    • Inquiry details (English) are stored.
  • Inquiry Details (Japanese translation) q_inquiryDetail_Jp 19*
    • Inquiry details (Japanese) are stored.
  • File q_inquiryFile 13
    • Contains files related to queries.
  • ▼ Response Content ▼ 14
  • Responder q_operator 16
    • Stores the person in charge of creating the query answer.
  • Response Date/Time q_responseDatetime 17
    • The date and time of the reply to the inquiry source is stored.
  • Response Email Destination q_mailTo 20*
    • Stores the destination address for inquiry response emails.
  • Response Email Sender q_mailFrom 21
    • The sender address of the inquiry response email is stored.
    • customer-service@example.com
  • Response Email Body q_mailBody 22*
    • The inquiry response email body is stored.
  • Draft Response (English) (A) q_response 25
    • The draft text of the inquiry response email (English) is stored.
  • Draft Response (Japanese) (B) q_response_Jp 24*
    • The draft text of the inquiry response email (Japanese) is stored.
  • Response-Related Files q_responseFile 33
    • Stores files related to query answers.
  • ▼ Internal Communication ▼ 29
  • Correspondence Score q_point 30
    • Stores the rating for handling inquiries.
  • Knowledge should be shared internally q_knowledgeFlag 31
    • Stores whether or not the query contents and results should be shared within the company.
  • Comments q_comment 32
    • Memos and comments are stored.

Field Name, num, Initial Value

Download

This archive contains the BPMN icons that are not available in the Basic edition.

Notes

  • For the AI part of the process, the model gpt-4 (or gpt-4-32k) may cause a processing time-out. In case of a time-out, the model gpt-3.5-turbo has been used to cover this. With gpt-3.5-turbo, the accuracy of the response changes slightly, but the likelihood of a time-out is significantly reduced.

Capture

See also

OpenAI ChatGPT: Chat
Inquiry Management Process, ChatGPT Integration 
Inquiry Response Process, ChatGPT Connection Response Review
Scroll to Top

Discover more from Questetra Support

Subscribe now to keep reading and get access to the full archive.

Continue reading