Restaurants have more data than ever. Decisions are still the hard part.

Plenty of data, still looking for an answer

A head of operations at a UK restaurant group said that to me before I first shared this story in June 2026. Ten sites, growing fast, plenty of data. A massive headache figuring out what to do with it.

I had heard versions of that sentence for years. Some operators were swimming in dashboards and still guessing at decisions. Others were spending hours cleaning data, copying tables and reading report after report. Neither felt like a good use of the time needed to run a restaurant. That gap is why I started Zvelto.

What the operator side taught me

I spent the better part of a decade on the operator side of hospitality. At Kout Food Group, I ran cross-functional programmes for Burger King, Pizza Hut and Taco Bell across 200 locations in the Middle East. At REEF, I led the launch of more than 250 delivery-first restaurant locations across 17 brands in the UK, including Pizza Express, Coco di Mama and Wendy's.

At Uber Eats, I managed enterprise partners including Caffè Nero, Fireaway Pizza and The Big Table Group, a portfolio bringing in £80 million a year. These are the experiences I described in my original account: different roles, different vantage points, but the same conversation with operators.

The tools were there: POS systems, delivery analytics, inventory software and staff scheduling apps. What was missing was a clear way to bring the information together. Should I run a promotion tonight? Is this dish hurting my margin? Why are sales at this site down this week? Getting from the numbers to a useful answer took time most operators did not have.

From another dashboard to a useful next step

I had been thinking about that problem for years. I went full-time on Zvelto in January 2026. Manos joined as co-founder and CTO in early May 2026. As I described in June, he brought more than a decade of technology experience in the sector, had solo-founded a platform that grew to more than 100 restaurant customers, and had owned and operated a cafe-bar.

A dashboard tells you what happened. What I want a hospitality decision engine to do is help you decide what to do next. Our ambition for Zvelto is to connect information from the tools a restaurant already uses and surface useful actions in plain language, rather than give an operator another report to interpret.

That might mean assessing an offer, reviewing a dish's margin or reconsidering a staff shift. Recommendations that an operator could act on in one click, or choose to automate, are part of that ambition. This is a direction for the product, not a claim that every integration or automated action is available today.

Early lessons from two Athens locations

In May 2026, we ran two pilots, one at each Mandraki restaurant location in Athens, during Saturday peak service. In my June account, I reported 14 user feedback responses in the first few days, a Net Promoter Score (NPS) of +57, 100% order accuracy, and 93% of diners saying the experience was better or faster than usual. More than half said it was both.

These are our own reported findings from a small early pilot, not independently audited results or a promise of what another restaurant will achieve. They do not establish that the wider decision-engine ambition has been delivered.

The detail that stayed with me was a staff member on the top floor saying they no longer had to go downstairs to pass on and clarify each order. That practical change meant more to me than a polished case study. I would rather share early numbers with their limits than make them sound conclusive.

Who we are building for

The direction I set out in June 2026 was to build for operators running several sites, or one busy restaurant, with the UK and Greece first. The problem is not a shortage of software. It is the work still left between reading a report and making a decision. That is the work I want Zvelto to help reduce.