"We use AI for customer support." This sentence could mean almost anything. It could mean a scripted chatbot with a few decision trees. It could mean an LLM-powered chat widget that deflects customers before they submit a ticket. Or it could mean an AI engine that autonomously resolves submitted tickets without any human involvement.

These are not the same thing. The confusion between them leads to poor buying decisions, misaligned expectations, and support operations that are partially automated at best.

This guide draws the clear distinctions — between scripted chatbots, LLM chatbots, and AI ticket resolution engines — and gives you a framework for choosing the right tool.

The Three Types of "AI" in Customer Support

Type 1: Scripted Chatbots (Rule-Based)

Scripted chatbots follow decision trees. They present the customer with a menu of options, navigate through predefined branches, and deliver a pre-written response at the end of the flow.

How they work: The customer selects "Billing issue" → selects "Invoice query" → selects "I need a copy of my invoice" → the chatbot sends a link to the invoice portal.

What they do well: Consistent, predictable responses for a small number of well-defined scenarios. Easy to build and maintain for simple use cases.

What they cannot do: Handle anything outside the predefined decision tree. If the customer's question does not fit a branch, the chatbot fails. They cannot understand natural language — they match keywords or button clicks.

Best for: Very simple, high-volume deflection scenarios with a small number of possible outcomes. Not suitable for teams with diverse ticket types.

Type 2: LLM Chatbots (Conversational AI)

LLM chatbots use large language models (GPT-4, Claude, Gemini) to engage customers in natural language conversation. They can understand questions phrased in any way, query a knowledge base, and generate contextual responses.

How they work: The customer types "I can't find my invoice from last month" → the chatbot understands the intent, queries the knowledge base, and responds with instructions for accessing the invoice portal.

What they do well: Natural language understanding, handling diverse question phrasing, knowledge base querying, conversational context across multiple turns.

What they cannot do: Process tickets that have already been submitted. Handle email-based support. Operate outside the chat interface. Most LLM chatbots are designed for pre-submission deflection — they intercept customers before a ticket is created.

Best for: Teams with high chat volume who want to reduce inbound ticket creation. Works well when customers are willing to engage with a chat interface.

Type 3: AI Ticket Resolution Engines

AI ticket resolution engines operate after ticket submission. They receive submitted tickets (via email, API, or webhook), analyse the full context, query knowledge bases and historical resolutions, generate a complete reply, and post it directly into the ticket — without any human involvement.

How they work: Customer emails support@company.com → ticket is created in the helpdesk → webhook fires to the AI engine → AI queries knowledge base and historical resolutions → AI posts reply to the ticket and closes it → customer receives a response within seconds.

What they do well: End-to-end resolution of L1 tickets across all channels, including email. Confidence scoring to determine when to send automatically vs escalate. Learning from approved resolutions. Operating at scale without headcount.

What they cannot do: Intercept customers before they submit a ticket. Handle complex, multi-system issues. Replace human agents for emotionally sensitive situations.

Best for: Teams processing high volumes of L1 tickets across email and other channels. Particularly effective when email is the dominant support channel.

The Core Distinction: Deflection vs Resolution

The most important distinction is not about technology — it is about when in the support lifecycle the tool operates.

Chatbots (both scripted and LLM) operate pre-submission. They intercept customers before a ticket is created. Success is measured by deflection rate — how many customers found an answer without submitting a ticket.

AI resolution engines operate post-submission. They process tickets after they have been created. Success is measured by automation rate — how many submitted tickets were resolved without human involvement.

These are different problems with different solutions. A chatbot cannot reduce the backlog of tickets already in your queue. An AI resolution engine cannot prevent customers from submitting tickets in the first place.

When to Use Each

Use a Chatbot When:

• Your primary support channel is chat (website widget or in-app messenger)
• Customers are willing to engage with a chat interface before submitting a ticket
• You have a well-maintained knowledge base that the chatbot can query
• Your goal is to reduce inbound ticket volume, not to resolve tickets faster
• Most of your support queries are simple enough to be handled in a short conversation

Use an AI Resolution Engine When:

• Email is a significant support channel (most businesses)
• You want to reduce the volume of tickets reaching human agents, not just prevent submission
• You have a high proportion of repetitive L1 tickets (password resets, billing queries, how-to questions)
• You want measurable, auditable automation with confidence scoring
• You need to handle tickets submitted outside business hours

Use Both When:

• You have high chat volume AND high email volume
• You want to maximise automation across all channels and all stages of the ticket lifecycle
• You use a platform like Intercom that supports both chat deflection (Fin) and email ticket resolution (via webhook)

The Decision Framework

Answer these questions to identify the right tool:

Question	Chatbot	AI Resolution Engine
What is your primary support channel?	Chat	Email or mixed
When do you want AI to act?	Before ticket submission	After ticket submission
What is your primary goal?	Reduce inbound volume	Reduce agent workload
Do you need to handle email tickets?	No	Yes
Do you need confidence scoring?	Rarely	Essential
Do you need to learn from past resolutions?	Rarely	Yes

Common Misconceptions

"A chatbot will solve my ticket volume problem"

A chatbot reduces the number of tickets submitted — but only for customers who engage with the chat interface and find the answer. Customers who prefer email, submit tickets outside business hours, or have already tried self-service and failed are not helped by a chatbot.

If your ticket volume is high, a chatbot addresses one part of the problem. An AI resolution engine addresses the other part.

"AI ticket resolution is just a smarter chatbot"

An AI resolution engine is not a chatbot. It does not engage customers in conversation. It processes submitted tickets — reading the full context, querying knowledge sources, generating a complete reply, and posting it to the ticket. The customer experience is receiving a fast, accurate email response, not a chat conversation.

"I need to choose one or the other"

Many teams use both. A chatbot handles pre-submission deflection for chat customers. An AI resolution engine handles post-submission resolution for email tickets and tickets that make it through the chat flow. The two tools are complementary, not competing.

Where TR-AI Fits

TR-AI is an AI ticket resolution engine — Type 3 above. It processes submitted tickets via webhook, resolves L1 tickets using your knowledge base and historical resolutions, and posts replies directly into the ticket thread.

It is not a chatbot. It does not intercept customers before they submit a ticket. It resolves tickets after they are submitted — across email, chat, and any other channel your helpdesk supports.

If you are evaluating support automation tools and want to see what a resolution engine looks like on your actual tickets, start a free trial — no credit card required.