Your business runs on somebody remembering.
The enquiry nobody called back. The quote nobody chased. The review nobody asked for. None of that was a decision anyone made. It is just what happens when the operation depends on a person having capacity. We build the layer that does it anyway.
Most small businesses do not have a lead problem. They have an operations problem that shows up as a lead problem.
When enquiries are not turning into work, the instinct is to buy more enquiries. It is the wrong instinct most of the time, and it is an expensive one.
Here is what we find when we go and look. The enquiries are arriving. They are arriving at a business where the owner is on the tools until six, the office manager is across four different inboxes, and the CRM is a place where leads go to be entered rather than worked. So the enquiry gets a reply the next morning, or the next week, or not at all. By then the customer has called two other businesses and booked the one that answered.
Nobody in that business decided to let that happen. There is no meeting where somebody says we will ignore the quotes we send out. It happens because the operation has no layer that runs regardless of whether a human has capacity today, and every business without that layer eventually converges on the same failure: the leads keep coming in and the follow up quietly stops.
The reason this is worth caring about is arithmetic. Buying more leads costs money every month forever. Converting more of the leads you already get costs money once. A business converting one enquiry in ten does not need twice the enquiries, it needs to find out why the other nine went nowhere, and in almost every case the answer is not that they were bad leads.
That is what we mean by AI for operations. Not a chatbot bolted to your website. The operating layer underneath the business that answers, qualifies, follows up, books and reports, whether or not anybody remembered to.
Five things have to happen to every enquiry. Most businesses reliably do one.
This is the whole model. Everything we build sits in one of these five layers, and we build them in this order, because a layer is worthless if the one above it is leaking.
Answer
Every call, form, message and missed call gets a response inside a minute, at any hour, including the ones that arrive at 8pm on a Saturday. Nothing waits for someone to be free.
Qualify
Find out what the job actually is. What, where, when, how urgent, what budget. Written down in plain words and filed against the right record before a human ever picks it up.
Follow up
Not one attempt. A real sequence across SMS, email and phone that keeps going until there is an answer either way, and that stops immediately when there is.
Book
Into a real calendar, with confirmations and reminders, so the job that was agreed to actually gets attended and the no-show rate stops being something you tolerate.
Report
Which enquiries came from where, what they cost, and which of them turned into money. So the next decision about spend is arithmetic instead of a feeling.
The order matters more than the technology. There is no point automating your booking calendar if enquiries are dying two stages earlier at Answer. We start at the top of the stack, find the first layer that is actually broken, and fix that one first. Frequently that is where the engagement ends, because the first fix solved the problem.
What this looks like when we go and count.
837 enquiries in the CRM. Seven of them had been won.
This business was not short of leads. It was short of follow-up. We pulled every record out of their CRM and classified all 837 opportunities. 241 quotes had been sent and never chased, worth $776,688, sitting an average of 222 days untouched. Another 302 enquiries had never moved past the first stage at all, some for five months. Across the whole pipeline, only 42 records had been touched in the previous fortnight.
So we built the layer that does the chasing. Missed calls now get a text back inside a minute. An AI qualifier works the reply, finds out what the job actually is, writes it into the CRM and hands a written brief to a human. It never quotes and it never books. The quote backlog gets a real sequence carrying case studies and reviews instead of silence, capped daily so it goes out over a fortnight rather than in one blast. Alongside it runs a live report that attributes every inbound call to the tracking number that produced it, so the ad spend is finally answerable.
Figures from a full reconciliation of the client’s CRM, 837 opportunities across seven pipelines, pulled 23 July 2026. The call attribution report is live. Published without the client’s name.
Closing the loop on what the advertising actually produced.
The ad account said conversions. Nobody could say which ones became money.
Google Ads reported conversions every month and the business paid the invoice every month, but nothing connected the two. A conversion in an ad account is a form fill or a phone call. It is not a job and it is not a dollar. The owner was being asked to set next month’s budget on a number that could not tell him whether a single one of those calls had turned into work.
We built the report that closes the loop. Every inbound call is attributed to the tracking number that produced it, then reconciled month by month against the ad account and split commercial against residential. It runs from the CRM and the ads export, so it is the same arithmetic every month rather than somebody’s recollection. It is live, updated monthly, and the written commentary under the campaign table is generated from the figures rather than typed, so it cannot quietly go stale when a new month is added.
Describes a delivered and deployed reporting system in monthly use. Published without the client’s name.
The system we deliberately made less capable.
We built an inbox that could reply on its own. Then we switched that part off.
The inbox held 958 conversations with more than two hundred waiting on a reply, most of them unread. The obvious build is a system that reads each message, works out what it is, and answers it. So we built exactly that: it classifies every incoming message every two hours, drafts a reply, and files it as a task with the draft ready to go.
Then we let the automatic sending run, and it messaged four real prospects before we caught it. So we retired auto-send permanently. It drafts and files, and a human sends every reply. Two findings made that call easy. The platform’s own sentiment tags were wrong in every case we checked by hand: three marked interested had already signed with somebody else, two were vendors selling to us, and one marked not interested was handing over a phone number. And about a quarter of the inbox never triggers a notification at all, so anything built on those alone silently misses it. We would rather hand you a system that does less and never embarrasses you.
Describes a live internal system and a real incident. The sentiment tag failures come from a hand-checked sample of the live inbox.
The operations we automate most often.
None of these are exotic. They are the jobs that every business intends to do consistently and no business actually does consistently.
Missed call text-back
A call that rings out leaves no record and no follow up. Within about a minute the caller gets a text that opens a conversation, and the enquiry stops being invisible.
After-hours cover
A third of enquiries arrive when nobody is working. They get a real response at 8pm on a Saturday, get qualified, and are sitting on your desk qualified on Monday morning.
Quote follow-up
The quotes you sent that went quiet get a structured sequence carrying proof and reviews, rather than one hopeful just checking in and then silence.
Review generation
The request goes out automatically at the point the customer is happiest, which is the day the job finished, not three weeks later when you remember.
Job and enquiry records
Calls, texts and form fills get written into the CRM as structured records, so the pipeline reflects the business instead of being a data entry chore nobody keeps up.
Reactivation
Old enquiries, dead quotes and past customers are already paid for. They get worked properly, with the guards that stop you burning the list.
Four steps, and you can stop after any of them.
Audit
We go into your systems and count. How many enquiries arrived, how many were answered, how fast, how many were followed up more than once, and where they stopped. This is data, not opinion, and it usually surprises people.
Map
We show you the five layers, mark which ones are working, and put a number on the gap. You get this written down whether or not you continue with us.
Build
We build the broken layer first, on top of the systems you already run, with the guards agreed in advance. Live in two to three weeks.
Hand over
You get documentation, the numbers to watch, and the ability to leave. We do not hold your systems hostage and there is no lock-in contract.
What this is not.
- It is not a magic box. If your enquiries genuinely are bad, faster follow up gets you to the same no more quickly.
- It does not replace judgement. The system finds out what the job is and writes it down. A person still decides what to quote.
- It is not instant. Two to three weeks to the first working layer, six to ten for a full stack, and longer if the CRM needs cleaning first.
- It will not fix a business that cannot deliver. If you are already at capacity, more converted enquiries makes your problem worse, not better.
What it does do.
- Every enquiry gets a response, at any hour, without anyone having to remember.
- The quotes you have already sent get chased properly instead of going quiet.
- Your CRM reflects reality, because records are written by the system rather than by a person who has better things to do.
- You find out which advertising is producing money and which is producing noise.
- The work keeps happening on the weeks you are flat out, which are exactly the weeks it used to stop.
The things people ask before they start.
Where to go next.
Is your business sales ready?
The five questions that decide whether your operation can convert what it already gets.
Read more →The Lead Conversion System
The full stack applied to one problem: turning enquiries you already pay for into booked work.
Read more →Score your follow-up in two minutes
Ten questions, a score out of 100, and your three biggest leaks. No email required.
Read more →Start with the audit. It is free and it is useful either way.
Twenty minutes on the phone, then we go and look at your numbers. You get a written picture of where enquiries enter your business, where they stop, and what the gap is worth each month. If the answer is that your operation is fine, we will tell you that and you will have lost twenty minutes.