On the occasion ofAI Pulse, Dam Mulhem, co-founder of XXII, was invited on the set of Tech & Co - BFM Business to explain how XXII is redefining video analysis.
At the heart of this exchange: CORE, the real-time video analytics platform designed to transform existing cameras into veritable sensors of operational intelligence.

In a context where organizations are seeking to better understand what is physically happening in their stores, warehouses or infrastructures, XXII proposes a new, more agile and comprehensive approach, driven by AI and vision-language models.

CORE: a video analysis platform designed for use and performance in the field.

Deployed in a wide range of environments (retail, logistics, transport, public spaces), CORE converts video streams from installed cameras in real time.
The aim is to provide a clear, actionable and immediate view of the field, without modifying the existing infrastructure.

Thanks to its AI engines, CORE enables :

• object detection and classification,

• analyze trajectories,

• understanding of flows,

• identify critical situations (intrusions, falls, long waits, congestion).

Integration is seamless: CORE adapts to complex environments and operational constraints.

CORE: integration of Vision-Language Models (VLM), a major step forward.

One of the key announcements made during the presentation concerned the integration of Vision-Language Models (VLMs), an emerging technology that is changing the way AI reads and understands a video scene.

VLMs enable CORE:

• describe and interpret what it observes in natural language,

• adapt to new scenarios without re-training,

• respond to complex requests ("show me the areas of congestion over the last hour").

This evolution brings video analysis closer to a true understanding of the physical world, paving the way for richer, more directly operational use cases.

Retail: analyze the customer journey and optimize operations.

In retail, the pressure on customer experience and store performance is high. CORE helps retailers make better decisions by providing reliable, real-time data.

The platform enables :

• identify hot and cold zones,

• measure flows, waiting times and customer paths,

• detect congestion or understaffing,

• optimize merchandising and restocking by observing shelf behavior.

Connected to supply chain and store management tools, CORE supports faster, more accurate decision-making and smoother operational execution.

Logistics: operational visibility and error reduction.

In warehouses and logistics centers, CORE provides a complete view of operations: pallet movements, zone occupancy, picking efficiency, anomalies, losses, errors.

Key benefits include :

• reduced errors thanks to automated tracking,

• optimized flows and activity zones,

• improved resource allocation,

• real-time alerts for immediate correction of deviations.

Thanks to its API and data warehouse connectors, CORE can be easily integrated into existing architectures, accelerating scale-up projects.

Transport and public infrastructure: better anticipation, better security.

CORE is also used in transport and infrastructure management, where understanding flows is a major challenge.

The platform makes it possible to :

• prevent incidents thanks to real-time analysis,

• detect atypical situations,

• improve traffic flow,

• reinforce security, while integrating with existing VMS systems.

VLMs offer an additional level of interpretation: CORE understands the intention, the situation or the abnormality, and triggers the appropriate alerts.

Data protection: a privacy-by-design and RGPD-compliant approach.

Regulatory compliance is a pillar at XXII.
CORE has been designed to analyze without identifying: anonymized silhouettes, processing without unnecessary storage, ethical governance overseen by an independent committee.

This approach enables organizations to use the power of video analytics while complying with RGPD requirements and societal expectations of privacy.

Why XXII is establishing itself as a key player in augmented video analytics.

Video analytics is entering a new phase. CORE doesn't just detect, it understands, explains and contextualizes.

The addition of Vision-Language Models marks a milestone for the industry:

• more agility,

• more intelligence,

• more operational value,

• rapid implementation using existing cameras,

• strong compatibility with confidentiality issues.