Fingermark.

The idea did not begin in a lab. It began behind a bar. Two decades ago, Luke Irving was running venues in Wellington and had a simple thought: turn the point-of-sale around and let customers order for themselves. That instinct - remove the friction between a person and what they want - became Fingermark, founded in 2005.

Today the company is a long way from a bar counter. From a base in Hastings, in New Zealand's Hawke's Bay region, roughly 90 people build the technology that keeps some of the largest quick-service restaurant chains on earth moving. Its customers read like a fast-food hall of fame: McDonald's, KFC, Yum! Brands, Chick-fil-A, Popeyes, Carl's Jr., Burger King and Hungry Jack's.

What Fingermark sells is deceptively plain. Restaurants already have cameras. They already have menu boards and order screens. Fingermark layers intelligence on top of that hardware and, in doing so, converts a noisy, chaotic drive-thru into a stream of numbers a manager can actually act on: how deep the queue is, how long each car waits, where the line stalls, which orders go wrong.

The company's positioning is refreshingly narrow. It does not want to do everything for restaurants. It wants to do one thing - performance - completely. "Obsessed with QSR performance" is not a slogan bolted on after the fact; it is the filter every product decision runs through. The drive-thru, that most ordinary corner of daily life, turns out to be one of the most demanding real-time systems in retail. Fingermark reads it, second by second.

Much of the noise in restaurant AI has centered on voice - systems that take a customer's order at the speaker box. Fingermark sits in a different lane. Its edge is computer vision: reading the physical reality of the lane itself, the cars, the queues, the timing, rather than the words spoken into a microphone.

That distinction shapes everything. Competitors in the space range from voice-ordering vendors such as SoundHound, Hi Auto, Valyant AI and PreciTaste, to Google Cloud's work behind Wendy's FreshAI, to digital-signage and analytics players like RevelDigital. Fingermark's vision-analytics niche overlaps with none of them cleanly, which is part of the point.

Its second differentiator is scope. Conventional venture wisdom says software companies should avoid hardware. Fingermark does both - vision software, physical kiosks, and menu-board management - because the QSR problem does not respect tidy categories. The messy, integrated work is the moat.

And it does this without asking a chain to rebuild its stores. Because Eyecue reads existing cameras and integrates with the POS, BI and kitchen systems already in place, the barrier to trying it is low - a practical advantage when your buyer runs thousands of locations and hates disruption.