Perspective · 22 Aug 2026
Beyond the wolf: what two adjacent use cases would need
When an independent committee reviewed HERDER, our wolf-recognition project, they didn't just approve it. They pointed out it's modular enough to scale to problems we hadn't built for: bird-strike deterrence at airports, and wildlife management in solar parks. We took that seriously enough to think it through. The short version: the core carries further than you'd expect, one of the two fits us better than the other, and neither is a product.
AI on cameras isn't new here
First, the honest part. Recognition on camera isn't new at airports or solar parks. Security and CCTV systems already run it, and they run it well. But that's a different job: live video, wired and powered cameras, a control room, real-time response, built for security. What we'd bring is a different paradigm entirely, cheap off-the-shelf cameras, no live feed, running on battery or solar, producing a record you own rather than a stream someone watches. Not a replacement for a security system. A different layer, often needed on the same site, for a different reason: monitoring what moves through a place over time, not watching it in real time.
What already carries over
Underneath the wolf work is a generic core that any recognition-at-a-site problem needs, whatever the species or the domain: camera-agnostic intake (FTPS, plain FTP, or manual upload), region-scoped species recognition, a review queue with provenance stamped on every detection (which model, which version, confirmed by whom), effort-corrected metrics so counts stay comparable over seasons, and open export (Camtrap DP, CSV). None of that is wolf-specific. It's the operational layer any camera-based monitoring problem needs, and it's the part that's genuinely reusable.
Solar parks: the closer fit
Of the two, solar parks fit us better. They're expanding fast across the Netherlands and the wider EU, and nature-inclusive design and impact monitoring are increasingly part of permitting. Fencing and panel layout affect ground-nesting birds and ground-dwelling species, and someone has to show what actually uses the site, before and after. That's a monitoring-and-reporting job, and it's close to what we already do: the project and location model, effort tracking, and structured, ownable export map neatly onto an ecological-impact or compliance report. Cheap distributed cameras across a large, fenced, off-grid area is exactly the shape of problem our approach suits.
The honest gap: there's no dedicated solar-park product, and no customer in this space yet. This is an the architecture would fit observation, not a claim of use.
Airports: monitoring, not the trigger
Airports are the more speculative of the two, and it matters to be precise about why. Bird strikes are a real safety and cost problem, and the response to them, real-time deterrence, is safety-critical and low-latency. That is precisely the paradigm we don't work in. We don't do live video, and "no live feed" is an advantage for monitoring but a non-starter for a real-time deterrence trigger. So the honest fit isn't the alarm; it's the monitoring side. Different species use the margins of an airfield differently and at different times, and knowing which bird, when, and where, over seasons, is the kind of pattern that informs habitat management and a deterrence strategy, rather than firing it.
The honest gap is larger here: no live low-latency trigger, no safety-critical or aviation-certification path considered, and no airport operator spoken to. The recognition pattern carries over to the monitoring question; it does not make us a safety system.
Why we're not building either yet
Plainly: we'd rather finish what's in front of us than chase two verticals nobody has asked us for. The wolf and predator-conflict work, the Smart Wildlife Box feasibility study, and the maturity of the platform itself all come first. We start small and scale on evidence. If you work in either field and this resonates, we'd like to hear from you, because that's what would turn a speculative post into a roadmap item.
The bigger point
The committee's own phrase was a more data-driven approach to nature policy and management. The value in that isn't one clever point solution per problem. It's a single recognition-and-monitoring architecture that generalizes, the same operational layer under a wolf, a ground-nesting bird, and a species crossing a solar park. That's the thing worth building well, and building once.
