A driving school app has to know where the last pupil was dropped off.

A driving instructor's diary is a route, not a calendar

A calendar app books 10:00 to 11:30, then 11:30 to 13:00, and is pleased with itself. An instructor cannot do that, because the second pupil lives three miles from where the first one was dropped off. Every booking in a driving instructor app carries two addresses rather than one, and the gap between them is real time that nobody pays for.

That makes the booking record wider than a calendar event. Pickup address, drop off address, lesson length, which vehicle, which instructor, and whether this pupil is being collected from home, from school or from work, because it is often a different address on a Tuesday. Recurring lessons should be a rule rather than copied rows, since a pupil who takes 15:00 on Wednesdays will move that slot four or five times before the test, and each move is a different pickup.

Reminders belong in this section rather than in a marketing one, for a plain financial reason. A pupil who forgets costs the instructor the hour and the fuel, and most forgetting is fixed by a message the night before rather than by a stricter policy. It is the same machinery as any appointment reminder app, it is well understood, and it is worth building first rather than treating it as a nice to have.

What the lesson record has to hold

The other half of the app is the record of what was actually taught. DVSA publishes a driver's record for exactly this job. It tracks the level a pupil has reached against 27 driving skills, split across 8 groups, using five named levels: introduced, helped, prompted, independent and reflection. It is a good model to copy even outside Great Britain, because it describes progress the way an instructor already thinks about it.

Two design points fall straight out of that. A level is not a percentage and it is not a tick, and a pupil can drop back a level on a wet Friday, so a record that only moves upwards will quietly lie. And you want a row per skill per lesson rather than one current level per skill, because the question that matters in month four is not where the pupil is, it is how long they have been stuck there.

That record is also the whole case for a learner driver app on the pupil's own phone. Once the pupil can see the same groups the instructor is marking, the argument about whether they are ready for the test stops being an opinion against an opinion. Private practice hours logged with a parent or a friend fit in the same place, though the required amount differs by country and by state, so store it as a rule per jurisdiction rather than as a constant in the code.

A school with more than one instructor has a second reason to keep this tidy. Pupils get moved when somebody is ill or leaves, and the record has to move with them rather than living in a notebook in a footwell. If your instructors are employed rather than self employed, the hours behind the wheel are also payroll, which is a separate job with the same shape as an employee time clock app.

Cancellations, prepaid blocks and how the payment is allowed to work

Most schools sell lessons in blocks, which means the app holds a balance rather than a payment history. That balance has to survive the things people actually do: pause for four months, switch instructor, pass the test with three hours left, or ask for the money back after two. Decide what happens in each case before you build, because a credit with no expiry rule and no owner turns into an argument later.

The cancellation window is the other rule worth putting in software. If the policy is 48 hours, store the moment the notice arrived against the booking and let the app decide which side of the line it falls on. That is not about being strict. It is about the instructor not being the person who adjudicates it, pupil by pupil, in a message thread at ten at night.

Then there is where the money moves, and here the platform has an opinion. A driving lesson is a real world service consumed in a car, not inside the app, and Apple's App Review Guidelines cover that in 3.1.3(e), Goods and Services Outside of the App: if your app enables people to purchase physical goods or services that will be consumed outside of the app, you must use purchase methods other than in-app purchase to collect those payments, such as Apple Pay or traditional credit card entry. The neighbouring 3.1.3(d) points the same way for real time person to person services between two individuals. So lesson fees must not go through in-app purchase, and Apple takes no commission on them. The reverse holds for anything consumed inside the app, a theory test practice subscription for instance, which falls under 3.1.1 and does need in-app purchase.

Which card provider you use, how the blocks are priced and whether any of it becomes a subscription is a wider question than this page. The wider question of charging for lessons has its own trade offs, and none of them are settled by the software.

Apple App Review Guidelines, 3.1.3(e) Goods and Services Outside of the App

What you can realistically build this with

Driving school software exists, and for plenty of schools buying it is the right answer. People go looking anyway because of a rule the product does not have: a block of lessons shared between two siblings, a pupil who can only be collected from work on alternate weeks, a test day that has to block three hours including the drive to the test centre and back.

If you are weighing up the AI app builders, check what each one actually produces before you commit, because the output differs and the front page does not always say. Firebase Studio is the example worth knowing about. Its App Prototyping agent builds web apps with Next.js, not a React Native or Expo project, so what comes out is a site rather than something a pupil installs. It is also on the way out: Google's own documentation says new workspace creation and user signup were disabled on 22 June 2026, and that Firebase Studio sunsets on 22 March 2027, with existing projects to be migrated before that date.

The distinction that matters most in this category is whether the result ends up on a home screen. An instructor works from a phone propped on the dashboard in the two minutes between lessons, often one handed and sometimes with no signal worth the name. A page that has to be loaded each time is a worse tool for that than an icon, whatever is running behind it.

The smallest useful version is smaller than people expect. One instructor, a booking with two addresses, a list of skills with a level against each, and a balance per pupil. That is a fortnight of evenings, and it removes most of the paperwork before anything clever gets added.

Firebase Studio documentation, sunset and project migration