Retail automation in Hungary has traditionally trailed Western Europe and North America by about three to five years, partly because lower local labour costs reduced the immediate economic pressure to automate. Recently, however, adoption has picked up: developments range from self‑checkout tills through AI‑based image recognition to remotely monitored, 24/7 stores. We discussed practical experience and deployments with Attila Bessenyei, CEO of Laurel, a company that develops self‑checkout systems.
Where is automation happening?
- Self‑checkout remains the most visible automation element; although the technology appeared in the early 2010s, it became widespread in Hungary mainly in the post‑Covid period, later than in many other countries.
- Artificial intelligence is increasingly used for fruit and vegetable recognition and for automated analysis of video: image recognition identifies items, and behaviour analysis can detect suspicious events.
- Electronic shelf labels combined with dynamic pricing and inventory optimisation provide stronger returns than standalone electronic labels.
What can AI systems do today?
Current systems can:
- identify products by image (for example fruits and vegetables),
- monitor in‑store behaviour patterns and alert security within seconds if someone conceals an item,
- connect in‑store events with the payment process so that discrepancies between taken and paid items trigger warnings or halt transactions.
The central question is not only technical feasibility but investment viability: cameras, compute capacity, software and ongoing operations all carry costs.
Data protection and legal framework
From a data‑protection perspective, these systems operate on a similar logic to existing security camera solutions and can be implemented under GDPR. AI helps by spotlighting potentially suspicious events for human review rather than requiring continuous manual monitoring. The intent is loss prevention, not harassment of customers.
Autonomous and remotely monitored stores in practice
In practice, these are often not entirely staffless shops but hybrid models that extend opening hours via registration‑based automation. Purely vision‑based systems are cost‑effective only in limited scenarios; above roughly 20–30 square metres, purely camera‑based automation becomes harder to scale economically. Current deployments tend to target larger small shops (around 50–70 square metres) with a broader product range.
International and local figures:
- In the Czech Republic there are already more than 150 stores using registration‑based autonomous operation.
- In Germany the number of autonomous or remotely monitored stores rose from about 300 to more than 1,000 within a year.
A concrete Hungarian example is the Laurel–Coop Szeged project: the shop operates as a conventional staffed store by day and as a registration‑based automated outlet around the clock. Result: gross turnover rose by 24% in one year while daytime sales did not decline.
Theft, registration and operations
Requiring registration significantly reduces theft risk. In the Coop Szeged case there were over 30,000 purchases in a year and only a handful of theft incidents. The act of identifying oneself before entry appears to deter misconduct.
Remotely monitored shops are supported by live dispatcher services with detailed protocols and alert chains to manage incidents such as spills, broken goods or an intoxicated person. Many issues do not require on‑site intervention; remote operators can handle numerous software and permission adjustments.
Practical metrics and economics
- Intervention rate: interventions across the whole system are typically below 1% (this includes entry, door opening, registration and till issues).
- Investment scale: depends on existing infrastructure; a rough lower bound is about 10 million HUF for a retrofit.
- Payback: current experience indicates roughly a two–three year payback horizon, depending on store size and traffic.
Which locations benefit most?
The model is particularly promising for small rural convenience stores that face closure because running a store with full staffing is uneconomic. International examples include a 50‑square‑metre shop on a 400‑person village supported by the municipality, with daily one‑hour staff visits for restocking and cleaning, allowing the outlet to remain viable. The approach also suits campuses, hospitals, airports, pilgrimage sites and specialised small retailers.
A small convenience store will not typically compete with discounters on price, but automation reduces operating costs—especially labour—so that opening hours can be reduced to a few staffed hours per day and the rest handled remotely.
Payments and cash use
Loyalty schemes and coupons already work in current solutions: in Coop Szeged the same till system applies during staffed and automated periods, so discounts and loyalty are preserved. Most such stores today are cashless: although cash handling can be automated, practical problems arising from cash operations are harder to resolve without on‑site staff, so the regional norm is card‑only payment.
Future developments could see registration‑based automatic charging to linked payment methods, fully digital receipts and deeper integration of loyalty benefits, subject to technological and regulatory evolution.
Competition and outlook
Whether large chains’ expansion (for example if regulatory changes make new locations viable for big players) accelerates or slows small‑shop automation depends on the location and shop type. In some cases stronger competition may push small retailers toward automation to reduce costs; in others—very small communities—the big chains do not enter and automation offers a path to preserve local shops.
Online grocery delivery and instant‑commerce services address convenience too, but they are not identical solutions: deliveries often cannot provide continuous 24/7 immediate access. A remotely monitored, always‑open shop fills that immediate‑access niche.
Conclusion
AI and remote monitoring are enabling new retail operating models in Hungary. Pilot projects, such as the Laurel and Coop Szeged collaboration, demonstrate higher turnover, very low loss rates and plausible payback within two–three years. The model appears especially suited to sustaining small rural shops, though deployment depends on investment capacity, data‑protection compliance and market readiness.
(The article was prepared with support from the Hungarian‑language Alrite online dictation and video subtitling tool.)
