Skip to main content
Home

Self-service kiosks · software

The kiosk is the easy part. The software is the moat.

On the trade-show floor the hardware looks solved: AI-ready kiosks, voice ordering, “autonomous” stores. But a kiosk is a box. What turns a box into a business is the software behind it: taking payment reliably, driving the equipment, and reconciling the money afterwards.

The teams that win the kiosk wave are the ones whose software runs the hardware and closes the books.

Shift reconciliationDIWAVE
Bay · taken
card + QR + cash, each event time-stamped (UTC)
Bank · settled
the acquirer statement is matched to the bay automatically
Cash box · matched ✓ · variance 0
Every morning · owner’s report: bay = bank = cash
Illustration: the three numbers that must match every shift

Three questions that decide whether a kiosk makes money

  1. 01

    Does the payment actually reach the machine?

    On an MDB vending bus or an RS-485 line the controller is the master and the reader answers only when polled. If they disagree, the reader can say “paid” while the machine does nothing. The software has to speak the device’s protocol, not assume it.

  2. 02

    What happens when the network drops?

    For unattended points this is the normal case, not the exception. The terminal may approve offline; the question is whether the bay still starts and whether the event survives the outage. The answer is a durable local log — what, when, which device, the response code, the outcome — stored in UTC and synced later.

  3. 03

    Who reconciles the shift?

    Three numbers must match: the post, the bank, and the cash box. That reconciliation should arrive by itself, every morning. A mismatch that repeats every shift is not a glitch — it is a design gap.

Each of these has its own technical depth. We cover them separately:

The layer underneath

The software runs the hardware

Three jobs the kiosk screen never shows, yet they decide whether the point makes money. Doing the first two and skipping the third is the common mistake: bays run, money comes in, and the owner still does not know what the point earned yesterday.

Taking payment
Card, cash, QR and app — through one ledger. Every event has a pair, “payment taken” and “service delivered”, or a shortfall can never be investigated.
Driving the equipment
The bay controller, bill validator, coin acceptor and contactless reader — each on its own protocol (MDB, ccTalk, RS-485). The software polls them rather than taking their word.
Reconciliation and telemetry
A money close every morning and a fault alert before the customer sees it. The owner sees each machine, not an average across the network.

The market is growing — but the software wins

The self-service kiosk market is on a steep curve: analysts put it at roughly $39.1B in 2025 and about $67.2B by 2031, a ~9.5% CAGR, with self-checkout the fastest-growing segment (Ken Research, 2026). That is a market tailwind, not our own figures: the bet is not the shiniest screen but the layer beneath it that closes the books.

Market figures source: Ken Research (2026). Diwave publishes no own kiosk results here: only figures reproducible by a query against the production database go on the site.

Frequently asked questions

Why does “AI-ready” on its own guarantee nothing?

“AI-ready” is a hardware label. It says the kiosk can host AI and richer payments; it does not say the books close by morning. It raises the stakes for the layer underneath: more payment methods, more devices, more edge cases — and the same unglamorous requirement to prove, two days later, exactly what happened at each machine.

What if the machine is older than the terminal?

This is the normal situation, not the exception. First establish three things for each model: whether the protocol (MDB/ccTalk) is physically present, what level it supports, and how much power is left for peripherals. That table saves more than any supplier comparison.

Can payment be added to an existing kiosk without replacing hardware?

Often yes. If the controller exposes data over a documented interface, the integration is quick. A closed device with no channel at all means workarounds or replacement — the most expensive scenario, best established at the first meeting.

Who owns the data after the project?

You do. It is worth fixing in the contract: the code and the data are yours, deployed on your infrastructure. If transactions live only in a supplier’s dashboard and cannot be exported, you can neither reconcile revenue nor move to another contractor.

Tell us about your kiosk

Describe the equipment you have and how payments are taken, and we will say plainly what can be connected to your existing hardware and where replacing it is the honest answer. No promises on timelines or price before a call.