Back to Articles

ai

LINE Chatbot for Business: Use Cases, Costs and PDPA

A practical guide to building a LINE chatbot for business: use cases, how the Messaging API works, rule-based vs AI chatbots, rough costs and PDPA essentials.

September 25, 20267 min read

A LINE chatbot lets your LINE Official Account answer customers automatically, 24 hours a day, inside the app many Thai customers already use to contact businesses. It can handle FAQs, bookings, order status checks and lead capture. Simple rule-based bots are quick to build, while AI chatbots that combine a large language model (LLM) with RAG can answer open-ended questions from your own documents. Cost depends mostly on how many systems the bot connects to and how much it has to understand.

Why a LINE chatbot makes sense for Thai businesses

For many customers in Thailand, LINE is the default way to message a shop, clinic or service provider. That creates a familiar problem: the same questions arrive all day, staff answer them manually, and messages sent after hours wait until the next morning.

A chatbot does not replace your team. It takes the repetitive questions and structured tasks, and hands complex or sensitive conversations to a person.

Common use cases

Use caseWhat the bot doesTypical integration
FAQAnswers questions about prices, opening hours, locations, policiesKnowledge base or documents
Booking and appointmentsShows available slots, confirms bookings, sends remindersBooking system or calendar
Order statusCustomer enters an order number and gets the current status and trackingE-commerce platform, ERP or shipping provider
Lead captureAsks qualifying questions and sends the lead to your sales teamCRM, Google Sheets or email
NotificationsSends order confirmations, payment receipts and service updatesYour backend system

Most successful bots start with one or two of these, measured by how many conversations they resolve without staff.

How the LINE Messaging API works

You don't need to understand every detail, but knowing the basics helps you brief a developer and evaluate quotes.

  1. LINE Official Account and channel. Your business needs a LINE Official Account with the Messaging API enabled. This gives you a channel secret and a channel access token.
  2. Webhook. When a customer sends a message, taps a button or adds your account as a friend, LINE sends an event to a webhook URL on your server.
  3. Signature check. Your server verifies each request using the signature LINE includes in the request header, calculated with your channel secret. This stops anyone else from sending fake events to your bot.
  4. Reply or push. The bot answers using a reply token, which can be used once and only shortly after the event. Messages you send on your own initiative, such as reminders or broadcasts, are push messages and count toward the monthly message quota of your LINE Official Account plan. Check LINE's current plan pricing when estimating running costs.
  5. Interface elements. Rich menus (the button panel at the bottom of the chat), quick reply buttons and Flex Messages (card layouts) make bots easier to use than typing. For forms and longer flows, LIFF opens a web page inside LINE, and LINE Login can identify the customer securely.

Rule-based vs AI chatbots (LLM + RAG)

Rule-based chatbotAI chatbot (LLM + RAG)
How it worksButtons, keywords and predefined flowsUnderstands free text and generates answers from your content
Best forBooking, order status, menus, structured tasksFAQs with many variations, product advice, policy questions
AccuracyPredictable, only says what you wroteVery good with RAG and guardrails, but must be tested
Build effortLowerHigher: knowledge preparation, prompts, testing
Running costHosting onlyHosting plus LLM usage fees per conversation

Rule-based chatbots

Rule-based bots follow flows you design: tap "Book appointment", choose a branch, pick a time. They are reliable and cheap to run, but they struggle when customers type questions in their own words, especially in Thai, where the same question can be phrased many ways.

AI chatbots with LLM and RAG

An LLM can understand natural Thai and English and write a fluent reply. On its own, though, it does not know your prices or policies, and it may guess. RAG (retrieval-augmented generation) fixes this by first searching your documents, such as price lists, FAQs, manuals and policies, and then asking the model to answer only from what it found. We explain the approach in more depth in what RAG AI is and how businesses use it.

The hybrid approach we recommend

In practice, the best LINE chatbots combine both: buttons and fixed flows for structured tasks like booking, an AI layer for free-text questions, and a clear handoff to a human when the bot is unsure or the customer asks for staff. LINE Official Account supports switching a chat to manual replies, so your team can take over in the same conversation.

How much does a LINE chatbot cost?

Costs depend on scope. As a rough guide:

  • Simple rule-based bot with a rich menu and a few flows: the lowest cost, often a few weeks of work.
  • Bot with integrations to booking, order or CRM systems: more, because each connection needs development and testing.
  • AI chatbot with RAG over your documents: more again, because of knowledge preparation, evaluation and guardrails.

Vectorkub's AI solutions start from ฿44,999, and a typical AI chatbot project takes around 3–6 weeks. Running costs to plan for include server hosting, LLM usage fees (charged by the amount of text processed), your LINE Official Account plan if you send push messages, and ongoing updates to the knowledge base.

Privacy and PDPA

Thailand's Personal Data Protection Act (PDPA) has been fully in force since 1 June 2022, and it applies to chat data too. Names, phone numbers, addresses and order history collected by a bot are personal data.

  • Collect only the data the bot actually needs for the task
  • Show a privacy notice, for example in the welcome message or rich menu, explaining what you collect and why
  • Get consent where required, especially before using data for marketing
  • Check where an AI provider processes and stores data, whether it uses your data for training, and what that means for cross-border transfers
  • Avoid sending sensitive data such as ID card numbers or health information to an LLM unless necessary; mask it where possible
  • Set retention periods and delete chat logs you no longer need
  • Restrict who on your team can access conversation history

FAQ

Do I need a LINE Official Account to build a chatbot?

Yes. The Messaging API works through a LINE Official Account. You can start on the free plan and upgrade if you need to send more push messages.

Can an AI LINE chatbot answer in Thai?

Yes. Modern LLMs handle Thai well. Quality depends on how well your knowledge base is written and on testing with real customer questions, including informal Thai and typos.

Will an AI chatbot give customers wrong answers?

It can if it is not designed carefully. RAG, instructions to answer only from your content, and a fallback to human staff reduce the risk significantly. Test thoroughly before launch and review conversations regularly.

How long does it take to build a LINE chatbot?

A simple rule-based bot can be ready in a few weeks. A typical AI chatbot project takes around 3–6 weeks, depending on integrations and how much content needs preparing.

Start with one conversation that matters

The best first step is to pick the one question or task your team handles most on LINE and automate that well. Learn more about our AI solutions for business or contact our team to discuss your LINE chatbot.

From Our Team

Related Articles

More thinking, field notes, and practical guidance from the Vectorkub team.

Ready to build something great?

Tell us about your project — we'll help you get there.