LINE AI Business Assistant
Turns LINE messages into routed tasks, AI replies and trackable customer workflows.
AI automation studio
Turn complex workflows into AI tools people can actually use.
I am Alvis. Bunun Studio focuses on LINE AI assistants, n8n automation, document search and web presentation. I start with a demonstrable version, then improve it into a stable system step by step.
panda@bunun.co
Automation loop
RUNNING
Document Query Mode
01Enter a tender keywordLINE message / query intent
02Retrieve matched passagesRAG document search
03Summarize with AIanswer with source notes
04Reply and recordkeep searchable logs
Follow up / search again
Business Card Scan Mode
01Upload card imageimage input from LINE
02Read fields with OCRname / company / phone / email
03Save contact datastructured contact record
04Confirm resultprepare next upload
Upload next card
Selected Work
AI automation cases shaped with a product mindset
These are not only function demos. They turn messages, documents, queries and workflows into systems that clients can understand, test and gradually adopt.
Tender AI Search
Organizes tender documents into a searchable knowledge base to reduce manual review time.
n8n Automation Hub
Breaks large workflows into modules so testing, errors and future changes are easier to manage.
LINE Public Data Assistant
Wraps weather, air quality and public data APIs into simple LINE query experiences.
Business Card OCR Lab
Extracts contact details from images and prepares structured records for follow-up.
What I Build
AI automation services for small teams
I clarify the problem first, then build a small working version. The goal is not to pile on technology, but to make each step understandable and maintainable.
Workflow Review
Find repeated, time-consuming or error-prone work and turn it into a clear workflow map.
AI Assistant Prototype
Use LINE, n8n, RAG or APIs to validate the workflow and response quality.
Web and Demo Pages
Package services, cases and demos into pages that clients can quickly understand.
Documentation
Prepare operating notes, risks and recovery steps so the system is not maintained by memory alone.
How We Work
Start small, then make it stable.
AI automation can easily become too large too early. I split it into stages that can be reviewed, tested and improved.
Define the task
Confirm users, input data, expected output and data that must not be touched.
Build a prototype
Create a working demo without real tokens or production data.
Test and adjust
Check empty data, wrong formats, poor answers and exception cases.
Prepare launch
Add documentation, deployment notes and recovery steps before production use.
System View
Typical automation architecture
Besides the high-level architecture, I also prepare workflow diagrams so clients can see how the system moves from input to output.
LINE / Forms
n8n Routing
AI / RAG / OCR
Reply and Logs
Contact
Want to organize one workflow first?
Send a short description of where the work gets stuck, how much time it takes, and whether it uses LINE, Excel, PDFs, websites or other tools. Do not include API keys, tokens or passwords.