Chatbot
Also: Conversational AI · Chat Assistant · AI Chat
Quick definition
A chatbot is software that simulates conversation with a user, typically through a website, app or messaging platform. Rule-based chatbots follow scripted decision trees. AI-powered chatbots use large language models to generate responses from context, making them far more flexible but also less predictable.
How it varies across Australia
Chatbot adoption varies sharply across Australian industries. Financial services and telecommunications have invested heavily in AI-powered chat for support deflection. Most small and mid-market businesses still run rule-based chatbots with limited depth, which means the gap between what visitors expect and what they get is widening as consumer familiarity with tools like ChatGPT raises the bar.
See digital maturity scores across Australian industries →The two types of chatbot
Follows a pre-written decision tree. Good for narrow, high-volume questions where the answer is always the same.
Predictable but brittleUses a large language model to generate responses from context. Handles open-ended questions but can hallucinate or drift off-brand.
Flexible but needs guardrailsWhat it actually means
The word chatbot covers a lot of ground. At one end you have a widget on a pricing page that asks 'Looking for something?' and routes you to three links. At the other end you have an AI-powered assistant trained on a company's documentation that can answer nuanced product questions, qualify a lead against an ideal customer profile (ICP), and hand off to a human with a full conversation summary.
Most marketing chatbots sit closer to the first end than the second, which is fine when the use case is narrow. The problems start when rule-based chatbots are deployed to handle questions that require judgement. Visitors leave frustrated. Conversion drops. The chatbot gets blamed when the real failure was a mismatched deployment decision.
For marketers, the two most useful applications of chatbots are lead capture and support deflection. Lead-capture chatbots intercept high-intent visitors, ask qualifying questions, and feed warm leads into a CRM or nurture sequence. Done well, this lifts conversion rate on paid traffic without increasing spend. Support deflection chatbots handle repetitive queries so the customer experience (CE) team handles complex ones. When either of these is working, it shows up clearly in cost per acquisition and customer satisfaction scores.
The generative AI shift is changing expectations fast. Visitors who use ChatGPT daily expect chatbots to handle nuance. A button-menu bot on an otherwise sophisticated brand website now reads as a mismatch. Closing that gap is a digital maturity decision, not just a customer service one.
A chatbot is not a strategy. It is a channel. What it says, and when it says it, is the strategy.
How it shows up
Chatbot performance shows up across several metrics at once. Conversion rate on pages where the chatbot is deployed. Lead volume and lead quality in the CRM. Support ticket volume (down is good for deflection). Customer satisfaction scores on post-chat surveys. Time-to-first-response for sales enquiries.
The diagnostic signal most teams miss is drop-off rate mid-conversation. If a large share of chatbot conversations end without a resolution, the bot is intercepting intent it cannot serve. That is a configuration problem, not a volume problem.
The Australian context
Australian consumer privacy expectations are relevant for AI-powered chatbots specifically. The Privacy Act 1988 and its ongoing reforms affect how chatbot conversation data can be stored, used for training, or shared with third-party AI providers. Australian businesses deploying chatbots powered by offshore large language models need to be clear about where conversation data goes and whether that meets their privacy obligations.
ACMA's rules on consent also apply if a chatbot initiates outbound contact or feeds into an email nurture sequence. The chatbot lead is only as valuable as the consent it captures.
Where people get this wrong
Related terms
Common questions
What is the difference between a chatbot and live chat?
A chatbot responds automatically using software. Live chat connects the visitor to a real person. Many platforms offer both: the chatbot handles the first contact and qualifies the query, then hands off to a human for complex conversations. The two work better together than either does alone.
Do chatbots help with lead generation?
Yes, when configured for it. A chatbot that intercepts high-intent visitors, asks two or three qualifying questions, and captures contact details can improve lead volume on paid traffic without increasing spend. The key is deploying it on pages where purchase or enquiry intent is already present.
Are AI-powered chatbots worth the investment for small Australian businesses?
For most small businesses, a well-configured rule-based chatbot on the pricing or contact page is enough. AI-powered chatbots add value when query volume is high, queries are varied, or the cost of a human handling each conversation is significant. The configuration and oversight cost is real and should be factored in.
How do I know if my chatbot is working?
Track resolution rate (did the visitor get what they needed), conversion rate on pages where the bot is active, and drop-off rate mid-conversation. If a large share of conversations end mid-flow, the bot is intercepting intent it cannot serve. Deflection rate alone is not a useful health metric.
Keep exploring
About New Rebellion
New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.
How we think →