The words chatbot and AI agent are often used as if they mean the same thing. For a small business, the difference matters because they solve different problems.
A chatbot is useful when customers ask the same questions again and again. An AI agent is useful when the answer is only the first part of the job, and the business also needs details collected, a rule checked or a task prepared.
Use a chatbot when the job is mostly answering. Use an AI agent when the job includes a workflow, a handoff, or a clear next step.
What is an AI chatbot?
An AI chatbot is a chat interface that answers customer questions. It may use your website, FAQ page, product notes or written instructions to reply in plain language.
On a small-business website, a chatbot might answer questions about opening hours, service areas, pricing, appointment types, delivery, refunds or what happens after someone sends an enquiry.
A chatbot is usually best when the conversation can end with a useful answer or a link. It should not need to make a decision, update a system or judge whether a job is a good fit.
What is an AI agent?
An AI agent is software that uses business knowledge and rules to help complete a defined task. It can ask follow-up questions, organise details and prepare a next step for a person or system.
For example, an AI agent for a trades business could ask for the address, urgency, photos and access details, then create a clean job summary. An agent for a service business could qualify an enquiry and draft a reply for the owner to approve.
Green Kauri builds custom AI agents around these narrow workflows. The goal is not to replace judgement. The goal is to remove repeated admin before a person makes the important decision.
1. A chatbot answers; an agent acts within limits
The simplest difference is action. A chatbot replies to a question. An AI agent can take a limited action or prepare an action, as long as the rules are clear.
That action might be small. It could tag an enquiry, draft an email, create a summary, check whether a request is inside the service area or ask for one missing detail before passing the conversation to a person.
The limits matter. A good agent should know when to stop and hand the task to a human, especially when money, risk, complaints or sensitive information are involved.
2. A chatbot uses answers; an agent uses rules
A chatbot needs good answers. An AI agent needs good answers and business rules. Rules tell the agent what it may do, what it must not do and when it should ask for help.
A chatbot might answer, "Yes, we service heat pumps in Lower Hutt." An agent might check the suburb, ask whether the unit is wall-mounted or ducted, collect a photo and mark the enquiry as a maintenance request.
If your rules are unclear, start there. Write down your service area, price signals, booking steps, escalation rules and the cases that always need human review.
3. A chatbot is often public; an agent may work behind the scenes
A chatbot usually appears on a website as a visible chat window. An AI agent may appear there too, but it can also work behind the scenes after a form, email or booking request arrives.
That hidden work can be valuable. The customer may only see a simple form, but the agent can organise the message, identify missing details and prepare a cleaner summary for your team.
This is why an agent is not always a bigger chat bubble. Sometimes the best agent is an internal helper that makes every enquiry easier to handle.
4. A chatbot can live on weak content; an agent needs clearer knowledge
A basic chatbot can answer from a short FAQ. An AI agent usually needs a clearer knowledge base because it is helping with a process, not just giving a reply.
Your website is often the first knowledge base. Clear service pages, FAQs, contact details and process notes help an agent answer consistently. They also help Google and AI assistants understand the business.
If the website is vague, an xEO optimisation pass can improve the content structure before the agent is built. For new sites, website development with xEO included gives the agent cleaner foundations from the start.
5. A chatbot handles common questions; an agent handles repeated workflows
A chatbot is a good fit when customers ask the same questions. An AI agent is a better fit when your team repeats the same workflow after those questions arrive.
Look at the work after the first message. Do you ask the same five follow-up questions? Do you copy details into a spreadsheet? Do you write the same first reply? Do you sort enquiries by urgency or service type?
If the answer is yes, an agent may help. If the task is rare or changes every time, a simple chatbot, better FAQ or clearer service page may be enough.
6. A chatbot improves response speed; an agent improves handoff quality
A chatbot can give a customer a fast answer. An AI agent can make the handoff to your team better by collecting the right context before a person replies.
That handoff quality matters for small businesses. A busy owner does not only need "new enquiry received". They need to know who the customer is, what they need, where the job is, how urgent it is and what information is missing.
The agent does not need to close the sale. It can prepare the conversation so the human reply is faster, clearer and less likely to miss important details.
7. A chatbot is easier to start; an agent is better when the task is clear
A chatbot is usually faster to start because the scope is smaller. You can begin with common questions, approved answers and a simple path to contact the business.
An AI agent takes more planning because it needs workflow rules. That extra setup is worthwhile when the task happens often and the steps are predictable.
Start small. One focused agent that handles enquiry intake well is usually more useful than a broad assistant that tries to do everything.
Which one should a small business choose?
Choose a chatbot if your main problem is repeated questions. Choose an AI agent if your main problem is repeated work after the question: collecting details, checking fit, drafting replies, preparing job notes or routing enquiries.
Many small businesses need both over time. The chatbot helps customers get clear answers. The agent helps the business handle the next step without losing context.
Before building either one, make sure your core business information is clear. A chatbot or agent will only be as useful as the services, policies, prices, areas and examples it can rely on.
What should you prepare before building an AI agent?
Prepare the information your team already uses to answer customers. This includes your services, service areas, common questions, price signals, booking steps, handoff rules and examples of good replies.
You should also decide where a person must approve the next step. For example, an agent may draft a quote note, but a person should approve the price and final promise before it is sent.
If you already have a website, this preparation may show which pages need clearer content. If you are starting fresh, it can shape the website and the agent at the same time.
Chatbot vs AI agent FAQs
Is an AI agent just a smarter chatbot?
No. A chatbot mainly replies to questions. An AI agent is designed around a task or workflow. It may still chat with the customer, but it can also collect details, follow rules, prepare summaries and hand work to a person.
Can a small business start with a chatbot first?
Yes. A chatbot can be a sensible first step if customers ask repeated questions. It helps you learn what people need. Later, those questions and answers can become part of an AI agent that also helps with intake or follow-up.
When is an AI agent better than a chatbot?
An AI agent is better when the business needs more than an answer. It suits repeated workflows such as lead qualification, booking intake, quote drafts, job summaries and routing enquiries. The task should be clear, common and safe to prepare before human review.
Does an AI agent need human review?
Often, yes. Human review is important for prices, promises, refunds, complaints and sensitive topics. A safe AI agent collects facts and prepares the next step, then asks a person to approve anything that affects risk, money or trust.
Green Kauri helps small businesses choose the right level of AI support. We can improve the website knowledge first, then build a focused agent around one useful workflow when the task is clear.