Security teams are about to get a major upgrade in how they hunt through endless camera footage. Conntour, a Y Combinator-backed startup, just closed a $7 million funding round led by General Catalyst to build what it's calling an AI-powered search engine for security video systems. Instead of scrubbing through hours of footage or relying on basic motion detection, security teams can now type queries like "show me anyone wearing a red jacket near the loading dock yesterday" and get instant results. It's the kind of natural language interface that's been transforming other enterprise software, now coming to an industry drowning in video data.
Conntour is tackling one of the most tedious problems in corporate security - finding specific moments in massive archives of surveillance footage. The company's $7 million seed round, announced today and led by General Catalyst with participation from Y Combinator, comes as enterprises are scrambling to apply AI to operational workflows that have remained stubbornly manual for decades.
The pitch is straightforward but powerful. Instead of security personnel spending hours rewinding through camera feeds or setting up complex alert rules, they can simply ask questions in plain English. "Show me every delivery truck that stopped for more than 10 minutes," or "Find instances of someone entering the server room without a badge." The system processes the query, analyzes the video data using computer vision models, and returns the relevant clips.
It's a workflow transformation that mirrors what's happened across enterprise software over the past two years. Natural language interfaces powered by large language models have moved from novelty to expectation, and Conntour is betting that security operations are ripe for the same revolution. The difference is that instead of querying documents or databases, the AI has to understand and search through visual information in real-time.
The market opportunity is substantial. According to industry estimates, the global video surveillance market exceeded $50 billion in 2025, with enterprises managing hundreds or thousands of cameras across facilities. But the tooling for actually using that footage has barely evolved beyond digital versions of VCR controls. Security teams regularly cite video review as one of their most time-consuming tasks, particularly during investigations or compliance audits.
General Catalyst's involvement signals conviction in the enterprise AI infrastructure play. The firm has been aggressively backing companies that apply AI to unsexy but critical business processes - the kind of workflows that don't make headlines but consume massive amounts of human hours. This follows a pattern of enterprise-focused AI investments across inventory management, quality control, and operational monitoring.
The technical challenge is non-trivial. Unlike text-based AI search, video analysis requires processing enormous amounts of visual data, often in real-time across multiple camera feeds. The system needs to understand context, recognize objects and behaviors, and do it all with the kind of accuracy that security teams demand. A false negative in a security context isn't just inconvenient - it could mean missing a critical incident.
Conntour is entering a market where incumbents like traditional security camera manufacturers have been slow to innovate beyond basic analytics. Companies like Verkada and Rhombus have modernized the hardware and cloud infrastructure side, but the search and query capabilities have remained relatively primitive. That's created an opening for software-first startups to build on top of existing camera infrastructure.
The Y Combinator pedigree matters here too. The accelerator has been pushing companies to apply AI to specific vertical use cases rather than building horizontal platforms. It's a recognition that the real value in enterprise AI comes from deep domain expertise and solving concrete workflow problems, not just wrapping an API around a foundation model.
For security teams, the promise is shifting from reactive to proactive. Instead of only reviewing footage after an incident is reported, they can run regular queries to spot patterns, anomalies, or compliance issues. "Show me every instance this month where the emergency exit was blocked" becomes a routine audit task rather than a manual inspection project.
The funding will presumably go toward scaling the engineering team and expanding beyond early customers. Enterprise sales cycles in security are notoriously long - trust and reliability aren't negotiable - so the capital gives Conntour runway to prove the technology at scale before needing to show massive revenue growth.
What's particularly interesting is the timing. As enterprises deploy more AI tools, they're getting comfortable with the idea of natural language interfaces replacing traditional software UIs. That mental model shift makes it easier for a company like Conntour to sell into organizations that might have been skeptical of "AI magic" just two years ago. The technology has moved from experimental to expected.
The competitive landscape will likely heat up quickly. Once the approach proves viable, expect both startups and incumbents to rush in with similar offerings. The question is whether Conntour can build enough of a technical moat and customer base to stay ahead. In enterprise software, being first matters less than being best and most reliable.
The $7 million round is a clear signal that investors see enterprise AI moving beyond productivity software into operational infrastructure. Security video represents exactly the kind of data-rich, labor-intensive workflow where AI can deliver immediate ROI. If Conntour executes, it could reshape how thousands of security teams interact with their camera systems daily. The bigger test will be proving the technology is reliable enough for high-stakes security environments where missed detections carry real consequences. But with backing from General Catalyst and Y Combinator, the startup has the resources to find out whether natural language search can finally make those endless hours of security footage actually useful.
