An enquiry that sits in an inbox until Thursday has usually booked someone else by Wednesday. Most small business owners know this and still lose jobs to it, because the person who should answer is up a ladder, in a consult room or halfway through a client meeting. AI lead qualification closes the gap between "someone asked" and "someone useful replied". Every enquiry, whether it arrives by web form, email, phone or chat, lands in your CRM, gets the right questions asked, is scored against rules you wrote and goes to the right person with a summary attached.
We build this into a custom CRM built around your business, not as another subscription bolted to the side. Your enquiries, customer records and rules sit in one place, and you can ask the system about any of them in plain English, by voice or by text.
How AI lead qualification works from the first message
It starts the instant something arrives. A web form, an email to your info address, a phone call, a message in the chat on your website: each becomes a lead record in the CRM straight away. Nobody writes a phone number on the back of an invoice and loses it under the seat of the ute.
The AI then reads what the person actually wrote. It pulls out the obvious details, such as name, suburb, contact number and what they need, and it notices what's missing. Someone who writes "need a quote for a hot water system" has told you the job type but not the address, whether it's gas or electric, or whether the old unit has already given up. Those gaps become the questions it asks next.
Once it has enough to go on, it scores the lead against your criteria and routes it. A high-fit urgent job might go straight to the on-call technician's phone. A routine quote request might land in the estimator's queue for the morning. Something well outside your service area gets a polite reply and a suggestion to look closer to home. The rules are yours. The AI just applies them consistently at 2am on a Sunday, which is more than most of us manage.
Phone enquiries follow the same path through our AI receptionist for after-hours and overflow calls, which asks the same questions out loud and writes the answers into the lead record.
The questions it asks and the criteria you set
Generic chatbots ask generic questions. Ours ask yours. On our first call we work out what separates a job you want from a job you'd rather pass on, and that becomes the question set. For most businesses it comes down to four things.
| Criterion | What the AI might ask | What you decide |
|---|---|---|
| Job type | "What do you need done, and is it urgent?" | Which services you want more of, and which you'd rather refer on |
| Area | "What suburb or postcode is the job in?" | Your service area, and whether jobs further out need a minimum |
| Budget | "Do you have a rough budget in mind?" | Whether a figure is required, optional or never asked at all |
| Timing | "When are you hoping to get this done?" | How soon counts as hot, and what happens to "sometime next year" |
You can add others. A physio clinic might want to know whether there's a GP referral; an accounting firm might care whether the books are up to date. The AI only asks what it still needs, so a person who explained everything in their first message isn't made to repeat it.
Questions are written in your voice and asked one or two at a time, because nobody wants to fill in the equivalent of a tax return to get a gutter cleaned. If someone would rather not answer the budget question, that's fine. The lead still comes through, marked as unknown on that point.
One opinion we'll stand by: elaborate points-based scoring models are mostly theatre for a small business. A handful of honest questions and clear rules beat dozens of weighted factors that nobody on the team can explain. We keep the scoring simple enough that you can read it, argue with it and change it.
An instant reply, then a person with the full story
The first response is the easiest part to automate well. The moment an enquiry lands, the person gets an acknowledgement on the channel they used: an email reply, an SMS or a message in the chat window. It says who you are, confirms the enquiry has arrived, explains what happens next and gives a rough idea of when a human will be in touch.
It doesn't pretend to be a person. The OAIC's guidance on privacy and commercially available AI products says public-facing AI tools such as chatbots should be clearly identified as AI to customers. We agree, and we build it that way from the start.
When a lead is ready for a human, the hand-over is the bit your team will notice most. Instead of a raw form dump, they get a short summary at the top of the lead record:
- who the person is and how they got in touch
- what they need, in a sentence or two
- their answers on area, budget and timing
- the score, and the reason for it in plain words
- anything unusual the AI noticed, such as a second property or a tight deadline
So your estimator picks up the phone already knowing it's a split system for a two-storey house in Brunswick that has to be in before Christmas. That first call goes a lot better when nobody has to open with "sorry, what was it you needed again?"
If you want the chasing handled too, qualified leads can flow straight into automated follow-ups for quotes and enquiries, so a quote sent on Tuesday doesn't quietly die by Friday.
Filtering spam and tyre-kickers without losing good jobs
Any business with a contact form gets junk: SEO offers, fake invoices, bots pasting links. Spam is the easy part. The AI recognises it, files it out of the way and never sends it to your phone.
Tyre-kickers need more care, because a vague enquiry today can turn into a solid job later in the year. So we don't delete them; we sort them. Someone with no timeline and no budget, or who wants a full kitchen renovation for the price of a tap washer, might get a helpful reply and a spot on a low-priority list rather than a call from your busiest person. Someone who sounds unsure but fits everything else goes to a human anyway.
The rule we work by is that the filter should fail towards a person. If the AI isn't confident whether a message is spam or a strange but genuine enquiry, it goes to your team with a flag on it. Losing a real job to an over-keen filter costs more than reading a few extra junk messages. Nothing it filters disappears, either; you can always check what was set aside and why.
You stay in control of every lead
AI lead qualification should make decisions faster, not take them away from you. Every score can be overridden and every routing rule can be changed. If you think a lead was marked too low, you bump it up and add a note saying why, and that note feeds into how we tune the rules. You can also pause the automation for a promotion, a week of leave or a particular service whenever you like.
And you can simply ask. Type or say "show me last week's leads from the northern suburbs that nobody called back" and you get a list you can act on. For the bigger picture, the same lead data feeds CRM reporting dashboards that show where enquiries come from and which ones turn into paid work.
Your data stays under your control as well. We tell you which AI providers process your enquiries and where the data is hosted, and it is never used to train those providers' models. Enquiries are full of personal information, so where the Privacy Act 1988 applies to your business, the Australian Privacy Principles govern how it's collected and used. Even outside the Act they're a sensible standard, and we design the questions and privacy wording with them in mind.
How we set it up and tune it after launch
We start with a 30-minute call to learn how enquiries move through your business today and where they get stuck. Maybe calls after hours go to a voicemail nobody checks until mid-morning. Maybe the website form still emails someone who left last year.
Then we scope the build and give you a fixed quote before any work starts. The scope covers the channels, questions, routing rules and who gets notified. If enquiries need to reach your Microsoft 365 or Google Workspace calendars, or your accounts in Xero, MYOB or QuickBooks, our CRM integrations that bring your systems into one place cover it. Existing leads come across from spreadsheets or your current CRM, so you don't start from a blank screen.
After launch we train your team and stay on for support. The first few weeks are where the real tuning happens. We look at the leads the AI scored and compare them with what your team actually thought of them, then adjust. A question that confuses people gets reworded. A suburb that should have been in your area gets added. A score that kept running hot gets cooled down. Changes and fixes are done within days, not parked for a quarterly release.
That ongoing review isn't optional polish. The same OAIC guidance makes the point that using an AI product shouldn't be "set and forget", and that's how we run it.
If enquiries are piling up somewhere in your business, book a 30-minute call with our team and we'll map out where they're getting stuck.
Questions we get asked
Will customers know they're dealing with an AI?
Yes. The acknowledgement and any follow-up questions by chat or SMS say clearly that they come from an automated assistant working for your business, and the person can ask for a human at any point. It's the honest approach, and it's the one Australian privacy guidance points to.
What happens if the AI scores a good lead badly?
You override it in the CRM and the lead is routed the way you choose. We review overrides as part of tuning after launch, so the same mistake doesn't keep happening.
Can it qualify enquiries that come in by phone?
Yes, when calls come through our AI receptionist. Callers are asked the same qualifying questions, and the transcript, answers and score land in the same lead record a web form would create.
Do we have to start from scratch with our customer data?
No. Lead qualification runs inside the CRM we build for you, and we move your existing data across from spreadsheets or your current CRM as part of the build, so your history comes with you.
How long does it take to set up?
It depends on how many channels and rules are involved. You get a fixed quote and a clear plan before any work starts, and once it's live, changes to questions or rules are done within days.
