A Dutch startup is bringing a new level of intelligence to the American back-of-house. Orbisk, Europe’s frontrunner in automated food-waste technology, is entering the U.S. with an AI platform already trusted by global hospitality brands. Its system captures and interprets kitchen waste in real time, turning ingredient recognition and pattern analysis into practical operational guidance. With sites typically cutting waste costs by around 70% and saving six-figure sums annually, Orbisk is positioning itself as a catalyst for smarter, leaner kitchen management. Its arrival comes as automation accelerates across the foodservice sector, setting the stage for a timely TechBullion interview with CEO Anastasia Dellis.
Please tell us a little more about yourself?
I’m Anastasia Dellis, CEO of Orbisk. I’ve spent 20+ years building and scaling purpose-led organisations and technology businesses across Europe and globally. I care deeply about high impact teams and building products that deliver value creation.
At Orbisk we combine AI, computer vision and deep operational design to help professional kitchens reduce waste, improve margins and scale their sustainability impact.
Orbisk is entering the U.S. market at a time when food waste is a major national challenge. What made now the right moment to launch, and what opportunities do you see uniquely in the American foodservice sector?
Now is the right moment for three reasons. First, the product is proven at scale: we have hundreds of monitors live, strong enterprise customers, and clear impact – millions of kilos of food saved and tens of millions of dollars in customer savings. Second, features like Orbisk’s AI-Powered Actions, stronger integrations and a repeatable rollout playbook significantly shorten time to value. We know how to move from pilot to multi-site deployment in a way that works for complex organisations. Third, market dynamics in the U.S. are compelling – food waste is a board-level issue for many large operators, labour pressure is acute, and operators are willing to invest in operational automation that delivers rapid ROI.
The U.S. is unique because it combines huge enterprise customers (hotel groups, cruise and corporate dining), large multi-outlet footprints and a high urgency on both cost and ESG reporting.
That creates an environment where a solution that ties waste reduction to operational decisions and real money saved is adopted quickly. Practically, we see strong opportunities in corporate dining, cruise operations and large hotel portfolios where multi-site rollouts and central procurement create leverage.
The Orbi device has been described as bringing predictive AI to back-of-house operations. Can you explain how the system recognises ingredients in real time and why that level of automation matters to kitchen teams?
The Orbi combines on-device computer vision with a cloud AI stack. The device captures photos of each waste event and uses a model trained on millions of labelled images and real kitchen scenarios to recognise ingredients, portions and preparation context in near real time. We then map that visual signal to menu items and Price Look Ups (PLUs), using menu and recipe metadata so the waste event is tied to a commercial line item.
Why this matters: kitchens cannot scale manual logging. Manual logs are slow, inconsistent and quickly abandoned. Real-time recognition eliminates the need for manual entry, gives immediate feedback to staff and creates a reliable dataset for AI-driven decisioning.
It also enables automations — for example, if we detect repeated waste of a specific ingredient, the system can suggest immediate adjustments to prep, par levels or ordering. That level of automation converts measurement into action, which is why teams adopt it: fewer tedious tasks, clear guidance, and faster proof of value.
Your new AI-Actions feature effectively acts as a recommendation engine. What types of operational insights does it generate, and how do those translate into measurable savings for chefs and managers?
AI-actions takes the data the Orbi collects and turns it into clear, prioritized recommendations for the team. It looks across sites, days, times and service types to identify patterns that humans would struggle to see at speed.
Typical insights include specific details and actions to tackle overproduction, menu and recipe optimisation, procurement alignment and operational routines.
For chefs and managers, this translates into a very tangible outcome: lower food costs, fewer last-minute stockouts and a more predictable margin. Because the recommendations are specific (“reduce prep of x by y% on these days) and tied into the operators own data, the team can act quickly and see the impact within weeks. In the U.S. where labour is tight and food costs are high, this kind of automated guidance is especially powerful. It allows culinary leaders to focus on the quality and guest experience while Orbisk’s AI-powered actions scan for waste and margin opportunities in the background.
Major brands such as Accor, Hyatt, and Carnival Cruise Line are already using Orbisk. What have you learnt from these global deployments, and how have those partnerships shaped the technology now entering the U.S.?
Enterprise deployment taught us three clear lessons.
First, proof is essential. Large brands only scale once they see consistent operational results across sites. Proof of concept work remains a central part of our approach.
Secondly, reliability matters as much as accuracy. In complex environments the real product is the combination of device, model and service.
Third, integration removes friction. Connecting Orbisk to menus, PMS systems and procurement tools speeds up onboarding and rollouts.
These learnings shaped our roadmap. We built AI Actions to make insights prescriptive. We strengthened hardware and data flows for dynamic environments. We also invested in integration partnerships so that successful pilots translate into predictable multi site deployments. This is visible in our enterprise plans for brands where the motion follows a clear path from proof of value to contracted rollout.
This is exactly the playbook we’re bringing to U.S. hotel, cruise and corporate dining operators: start with proof of value, then scale in a structured, low-friction way.
Carnival Cruise Line collaborated with Orbisk to build a motion-resistant AI model for ships. How did you approach that engineering challenge, and what does the early 17% improvement in prep accuracy reveal about AI’s potential in dynamic environments?
Ships are demanding environments with constant motion, changing light and rotating crews. To solve for this we collected a dedicated dataset from onboard kitchens. We enriched the training set with motion and perspective variations. We also made hardware improvements related to placement, mounting and sampling to maximise signal quality.
The 17% improvement in prep accuracy is important because it proves that the approach works. With the right data and engineering response, AI models can performance reliably even in challenging conditions It also suggests strong potential for other dynamic environments including healthcare kitchens and large banqueting operations. Most importantly, higher accuracy increases chef trust and adoption – and that’s what ultimately drives real operational savings.
Kitchens using Orbisk typically cut food-waste costs by up to 70%. What are the most common behavioural or operational shifts that drive those results once teams begin using the platform?
High performing sites share four traits.
- They treat measurement as a daily discipline and work from trusted data, not assumptions.
- They build simple habits such as short prep checklists and daily reviews anchored in AI forecasts.
- They integrate waste insights into procurement and menu decisions which reduces avoidable variability.
- They remove decision friction by letting AI handle repetitive guidance so staff can act quickly and consistently.
The technology provides real visibility. The teams turn that visibility into routine practice. That combination is what delivers sustained reductions in both waste and cost. In practice this means faster pilots, quicker transitions to rollout and stronger credibility with large US operators.
Orbisk recently closed an €8 million Series A round led by sustainability-focused investors. How will this funding accelerate your commercial strategy, particularly across the U.S. hospitality and institutional foodservice market?
The funding accelerates three critical areas.
- Commercial scale. We can invest in experienced sales and customer leadership.
- Product and integrations. We can speed up key integrations and AI development, especially the API ecosystem that connects Orbisk into procurement and ERP systems.
- Strategic partnerships. We can strengthen regional presence and shorten our path from pilot to portfolio-wide scale.
Ultimately this allows us to support US operators more locally: faster implementations, stronger in-market support and a roadmap aligned with the realities of the American foodservice landscape.
With nearly 1,000 systems deployed across 42 countries, what differences do you see in how various regions approach food waste reduction, and where does the U.S. fall on that maturity curve?
Regions vary widely. Northern Europe and parts of APAC tend to move fastest due to strong governance, consolidated decision making and higher expectations on sustainability reporting. Other regions, including parts of the US, have scale but more complex procurement structures which lengthen decision cycles.
I’d place the US in the middle of the maturity curve. The opportunity is huge and awareness is high, but execution often requires a regional approach, owner engagement and deeper integration into procurement processes. Our go to market reflects this reality. We use regional hubs, targeted owner conversations and focused pilots that funnel into portfolio rollouts.
Looking ahead, what role do you believe AI-powered automation will play in shaping the next generation of kitchen operations, and how close are we to a truly zero-waste ecosystem?
AI will move kitchens from measurement into full automation.
Today, Orbisk provides visibility and prescriptive actions. The next stage is automatic adjustments to procurement triggers, menu guidance and staff scheduling based on real time demand.
Zero waste at a global scale requires policy, supply chain change and shared ownership. In the professional kitchen where we can measure and intervene the path to near zero for avoidable waste is achievable. Progress will come from integrated AI, simple interfaces for staff and clear incentives for operators.
