Twelve cars promised for today. One lift. Four technicians, and one of them on a split shift.
The Golf in for brakes came in yesterday and is still waiting on pads. The Mazda has been standing for three days because the customer never answered the estimate. And at 8:41 the phone rings: a service that wants a slot this week.
The question is not how much work there is. It is which of the twelve goes in first, because that is what decides whether all twelve leave today.
That moment repeats every morning in every workshop, and it is almost never settled by looking at a calendar: it is settled from memory, counting who is free and which lift is left. It works until the day fills up. After that, what decides the outcome is how well the day was laid out before it started.
And laying it out well is not a matter of tidiness, it is a matter of money: an hour of workshop time is the only thing in the business you cannot get back. The hour you do not fill today will not be filled tomorrow, because tomorrow already has its own hours.
How many appointments really fit into a workshop day
The usual answer — «as many as fit» — no longer holds, because today the limit is not demand: it is the people available to meet it.
In the United States, the market with the best public aftersales data, the TechForce Foundation calculates that the supply of technicians covers only 59% of annual demand in the automotive sector, and that most of that need comes from replacing people who retire or leave the trade, not from growth. In the same market, Cox Automotive’s 2025 service study reports that 57% of workshops and franchised service departments are not fully staffed, and that eight out of ten expect the situation to stay the same or get worse.
The practical reading is the same in any country: a workshop’s capacity no longer grows by hiring someone next week. It grows by making better use of the hours you already have. And for that, you need to see those hours.
A well laid-out day does not mean a fuller day. It means that what is in the workshop moves forward instead of waiting.
How to assign a technician and a bay to each appointment
An appointment is not an hour on a calendar: it is an hour, a person and a place. If any of the three is missing, the appointment exists on paper and does not exist in the workshop. That is where half the delays come from: two cars on the same lift, or a job assigned to someone who does not usually do it.
Laying it out properly means answering three questions while the customer is still on the phone:
- What slot is actually left? Not the day’s slot, but that technician’s and that lift’s, with the centre’s opening hours and holidays already taken out.
- Who does it best and fastest? A service has a real duration and a speciality behind it. The same service does not take the same time in different hands.
- Which bay can I count on? Some jobs only come out on the lift and some can be done in any bay. Knowing that beforehand avoids the eleven o’clock jam.
Once the system knows those three things, assignment stops being a mental calculation. In Cafler OS, when you book a new appointment the platform proposes the technician with the lightest load that day and a bay compatible with the service, with the alternatives beside it and the appointments each one already has. You say yes, or you pick someone else.
What to do when a job runs long and drags the rest with it
A day comes apart almost always for the same reason: something planned for an hour takes two. The problem is not the delay, it is finding out late. At 12:00 there is nothing left to decide; at 11:20 you can still move the next car to another bay, call a customer or change the order.
That is why it helps to separate two views that often get mixed up. The day’s calendar answers what is coming in; the view of work in progress answers what is being done right now. In the second one, a service planned from 10:00 to 11:00 that is still open at 11:30 shows up for what it is: a bar that has stretched and is pushing the next one along.
Alongside that, knowing what is stopped and why. A car waiting for a part and a car waiting for an answer to an estimate are not the same as a car under repair, and above all they should not be occupying a lift. Seeing them flagged as blocked is what lets you give them their place: call the supplier, pick the estimate back up, or take the car off the bay until the material arrives.
How to decide which car goes in first
Back to 8:40 and to the one question that really orders the day. It is a decision taken with four things in mind at once: the time promised to each customer, how long each job takes, the parts you have and the bays that are free. Getting it right twelve times, every morning, is a job in itself.
That is the part where Cafler AI, the AI for aftersales, goes first. When you open the diary it has already counted the day — how many handovers there are, which ones are tight, which lifts, technicians and bays are free — and, above all, it points at the bottleneck: if six cars depend on the only lift, it says so in those words and proposes what to do with the three most urgent ones.
Each car arrives with its reason in one line: the tightest handover, the bay that will free up soonest, the job that closes quickly and makes room for the rest. And with concrete options per car: bring it forward, move it to another bay, reschedule it with the customer or leave it as it is.
“Cafler AI suggests, explains where it got the data and lets you decide. It never interrupts the work.”
That sentence is the rule, and it is worth reading slowly because it is what separates help from a nuisance. Nothing moves on its own. Every proposal cites its source — the diary, today’s load, the customer record —, it can be recalculated when things change, and whatever you apply is logged and can be undone. The judgement is still yours; what changes is that at 8:40 you already have a proposal on the table instead of a blank board.
What to set up before the scheduler can run on its own
This is the part almost nobody tells you and the one that decides whether everything above works. A scheduler cannot suggest anything if it does not know what your workshop is like. Four things, done once, never touched again:
- The centre’s opening hours and holidays. Without them, the system offers slots that do not exist.
- Your services, with their real duration. Not the theoretical one: the time your team actually takes. It is the number everything else comes from.
- The bays and lifts, with the services each one takes. So the system does not put a job that needs a lift into a flat bay.
- The staff and their specialities. That is what lets it propose whoever really does that job.
With those four in place, automatic assignment and the day’s running order stop being a promise and start getting it right. Without them, any scheduling software is a pretty calendar.
How to confirm the appointment so it reaches the workshop
That leaves the last stretch, the one that happens outside the workshop: turning a booked appointment into a car at the door. Two notices solve it and nothing else. A reminder to the customer a few hours before, so the appointment is fresh and they can change it if something has come up, and an internal alert when the minutes you decide go by without them showing, so you can call them or give that slot to another car that is waiting.
A slot that frees up at 9:15 and is filled at 9:30 is an hour billed. The same slot spotted at midday is gone.
The following Tuesday
Same workshop, same time, twelve cars. Now the day’s order is proposed with its reason beside it. The Golf in for brakes shows up flagged as blocked on a part, so it is not occupying the lift the car that can move forward needs. The Mazda has a concrete reason to pick its estimate back up. The service that called yesterday went into Thursday’s slot with a technician and a bay, without hanging up. And the one that ran long yesterday was spotted at 11:20, not at 12:00.
There are no more hours than last week. They are better spread, and this time you can see them.
