Milestone Systems, the Copenhagen‑based provider of video management platforms, announced a trio of generative‑AI services aimed at reducing manual effort in security operations while staying within tightening data‑privacy regulations. The offerings—AI Search, XProtect Video Summarization, and Video Anonymization—are slated for rollout through the company’s VMS, VSaaS and analytics stacks, with on‑premises options expected by the end of 2026.
Why generative AI matters for security teams
Security operators have long depended on labor‑intensive processes: scrubbing hours of footage, manually drafting incident reports, and painstakingly redacting personal data before sharing video with law‑enforcement agencies. Missed alerts or false positives can have severe repercussions, and the emerging EU AI Act, alongside GDPR, forces organizations to prove that any AI‑driven decision‑making is both accurate and compliant.
Milestone’s new tools aim to address those pain points. By leveraging vision‑language models (VLMs) that are fine‑tuned on security‑specific video data, the company claims the AI outputs are “consistent, reliable and purpose‑built for video intelligence”—a direct contrast to generic large language models repurposed for visual tasks.
The three new services
AI Search – A natural‑language interface that lets investigators describe a scene (“red sedan turning left at the intersection”) and instantly retrieve relevant clips across multiple cameras. The feature eliminates the need for complex filter configurations and manual timeline navigation.
XProtect Video Summarization – Generates structured, text‑based summaries of video segments on demand or automatically as part of an alert workflow. The output standardizes incident documentation, cutting review time dramatically.
Video Anonymization – Applies automated face and license‑plate redaction to footage before export, helping organizations meet GDPR, the EU AI Act and other privacy mandates without manual pixel‑bashing.
In a use‑case scenario, an operator dealing with a traffic collision could locate the incident with AI Search, obtain a concise written summary via Video Summarization, and then share a redacted version of the clip through Video Anonymization—all within a single workflow.
Architecture and data compliance
Milestone’s generative AI stack rests on a proprietary, “compliant data library” that the company says is the world’s largest collection of responsibly sourced video for AI training. The library is built from fair‑exchange agreements with existing Milestone customers, anonymized using brighter AI’s technology, and annotated with over 300,000 categories via NVIDIA tools.
The curated data is split into domain‑specific subsets—initially focusing on EU and US traffic scenarios—and used to fine‑tune NVIDIA’s Cosmos Reason foundation models. Milestone reports a 19.4 % gain in flow‑and‑direction correctness and a 4.4 % boost in alert verification accuracy compared with the base models, according to internal testing.
All training material is fully traceable and stripped of personally identifiable information, aligning the resulting VLMs with EU AI Act and GDPR requirements. The city of Genoa, Italy, has joined as the latest data contributor, expanding geographic diversity.
Deployment options and timeline
Milestone plans to make AI Search available as a cloud service by the close of 2026, with an on‑premises variant delivered through BriefCam for air‑gapped environments. XProtect Video Summarization and Video Anonymization will be integrated into the existing XProtect and Arcules suites, leveraging the same VLMs that power AI Search. The company emphasizes that the on‑premises option “delivers the full intelligence of AI with no internet connection required,” a point likely to resonate with highly regulated sectors.
Executive perspective
Andrew Burnett, Chief Technology Officer at Milestone Systems, framed the launch as a trust‑first approach:
“How can we build on our decades of experience in video management to prepare our customers for the next generation of artificial intelligence? The answer is trust. Trust that the data is compliant, trust that the AI is accurate, and trust that these tools will genuinely make operators’ jobs easier. That’s what we’re delivering today.”
Market implications
Milestone’s move reflects a broader shift in the enterprise AI market: vendors are transitioning from generic LLMs toward vertical‑specific models that can be audited for compliance. By anchoring its VLMs in a traceable, anonymized video corpus, Milestone sidesteps many of the “black‑box” concerns that have hampered AI adoption in security‑critical environments.
Competitors such as Axis Communications, Genetec and Avigilon have also begun to embed AI‑driven analytics into their platforms, but few have publicly tied their models to a compliant data pipeline of this scale. If Milestone’s performance claims hold up in real‑world deployments, the company could set a new benchmark for responsible AI in video surveillance.
What to watch at ISC West
Milestone will showcase the full suite—VMS, VSaaS, and analytics—at ISC West, booth #18053. Attendees can expect live demos of the AI Search interface, a walkthrough of Video Summarization’s output format, and a preview of the upcoming Video Anonymization workflow. The company also plans to highlight the new XProtect App Platform and the BriefCam analytics engine, underscoring its strategy of building an open, extensible ecosystem for security AI.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
