The contact form is dying
Most website visitors leave without making contact. They had a question, couldn't be bothered filling out a form, and went to your competitor instead.
This isn't a design problem. It's a fundamental UX problem. Contact forms are passive. They sit there waiting. They don't engage, they don't qualify, they don't help.
AI agents do all of those things.
What an AI agent actually does
An AI agent is not a chatbot from 2018 with scripted responses and decision trees. Modern AI agents, powered by large language models, have real conversations. They understand context, ask follow-up questions, and respond naturally.
Here's what that looks like in practice for a small business:
- A plumber's website visitor asks "do you do emergency callouts after 6pm?" at 11pm on a Saturday. The AI agent answers yes, captures their details, and books the callout.
- A dental practice visitor asks about insurance coverage. The AI checks the practice's accepted providers and answers instantly, then offers to book an appointment.
- A restaurant visitor wants to know if there's a gluten-free menu. The AI shares the options and asks if they'd like to make a reservation.
None of these people would have filled out a contact form. All of them became leads or bookings because an AI agent engaged them at the moment they were interested.
What changes when an agent answers
The shift is structural, not magical. Enquiries that used to land in an inbox overnight get answered in the moment, while the customer is still on your site. Routine questions (hours, service areas, pricing ballparks, availability) stop consuming staff time. And every conversation is captured, so you see what customers actually ask, which is its own kind of market research.
We do not publish blanket conversion percentages, because the honest answer is that results depend on your traffic, your industry, and what your agent is allowed to do. What we can show you is the conversation log and the captured leads, which is better evidence than any average.
Why "just get a chatbot" isn't the answer
Not all AI agents are created equal. The difference between a good AI agent and a frustrating one comes down to three things:
Knowledge depth: Does it actually know your business, or does it give generic responses? The best AI agents crawl your entire website and build a knowledge base specific to your services, pricing, and processes.
Learning ability: Does it improve over time? Our improvement review reads conversations, finds gaps and suggests fixes that a person approves before anything changes. Most chatbots stay exactly as dumb as the day you installed them.
Escalation intelligence: Does it know when to hand over to a human? Frustration detection and intelligent escalation rules are the difference between capturing a lead and losing a customer.
Setup without the homework
The biggest objection we hear is "I don't have time to set up and train a chatbot." Fair enough. That's why the setup does the homework for you.
You paste your website URL. The agent crawls your site, extracts your services, FAQs, brand colours, and key information. It builds its knowledge base automatically. You customise the personality and goals, copy one line of code to your website, and you're live. Self-serve setup in Studio is quick to start; a fully managed build, trained and tuned by our team, takes one to two weeks.
No prompt engineering. No training data. No technical skills required.
What to give the agent before it goes live
An agent is only as good as what it knows. The automatic crawl gets the basics from your website, but the most useful details are often not on it. Before you switch it on, give it:
- The services you offer and the ones you do not, so it can say no politely instead of guessing
- The areas you cover, written out by suburb or region
- Your hours, including what happens after hours and on public holidays
- Price ranges or starting prices where you are comfortable sharing them, and what makes a job cost more
- Your policies on bookings, cancellations, deposits and warranties
- The questions it should never answer itself, such as medical, legal or financial advice, or a firm quote that needs a site visit
The last list matters as much as the rest. A good agent knows when to say "that needs a person" and how to collect the details so the person can follow up quickly.
How to tell whether it is working
Read the conversation log every week for the first month. You are looking for three things: questions the agent answered well, questions it could not answer, and conversations where a customer gave up. The second and third lists tell you exactly what to add to its knowledge.
Then look at the captured leads. Are they real enquiries with usable contact details? Are they arriving at times your office is closed? Are they for the work you actually want? If the answers are mostly yes, the agent is doing its job. If not, the log usually shows why within a few conversations. Keep a short running note of what you added to its knowledge and when, so you can see which changes cut down the questions it could not answer.
When an AI agent is the wrong tool
An agent is not the answer to every problem. If your sales depend on a long conversation with an expert, the agent's job is to book that conversation, not to replace it. If you work in a regulated field, keep it to general information and booking, and have a person handle anything that could be read as advice. And if your team cannot follow up leads quickly, capturing more of them will not help until that is fixed: a lead that waits two days for a reply is often a lead lost. In those cases a clear contact page and a fast reply from a person will do more than any agent.
Being upfront with your visitors
Tell people they are talking to an AI assistant. Most visitors do not mind, and they mind a great deal if they find out later. Collect only the details you need to follow up, and make sure your privacy policy says how chat conversations are stored and who can see them. Businesses covered by the Privacy Act need to handle personal information collected through a chat the same way as information collected through a form, so treat the transcript with the same care as any other customer record.
What inaction actually costs you
Every hour your website runs without an AI agent is an hour you're losing leads to competitors who have one. The cost of inaction isn't zero. It's every visitor who had a question, didn't get an answer, and went somewhere else.
