Organisers were making important layout, staffing, and sponsor decisions without a live view of how people were actually moving through the venue.
Manual counts and survey feedback came too late. By the time the team understood which zones were congested or underperforming, the event day had already passed.
The solution had to preserve privacy. The useful output was aggregated audience intelligence, not identity tracking.
We started with edge inference on a small set of existing CCTV feeds. The first proof measured whether we could count attendees and estimate dwell zones without adding new on-site hardware.
We then visualised the output as heat and flow maps so the organiser could immediately see where crowds gathered, stalled, and dispersed.
That proof gave the team a concrete operational loop: adjust layout or staffing on day two based on what happened on day one.
EventSense runs computer vision against standard CCTV feeds and produces aggregated signals such as footfall, dwell, anonymised demographic mix, and spatial heatmaps.
The organiser dashboard ties those signals to the moments that matter: keynote, queue, booth activation, peak arrival, and exit.
Because the analysis is aggregated and anonymised, the platform gives organisers useful live intelligence without turning the event into a personal tracking system.
- Event teams gained live visibility into crowd behaviour.
- Sponsors and stakeholders could receive evidence-based post-event reporting.
- Operational decisions became faster because the signal arrived while the event was still running.