Catch a drowning before a lifeguard can.

Questions buyers ask us

FAQ

How does AI drowning detection actually work?

Detection runs on classification models trained on hundreds of thousands of swimming and drowning sequences. The model watches each swimmer's trajectory, posture, and vertical position. When motion patterns match a documented distress profile (face-down stillness, vertical-bob without breath, sudden disappearance under the surface) the system fires an alert. The lifeguard or on-site responder gets a phone or radio page with the camera feed pre-cued.

Underwater cameras or above-water? Which do we need?

Above-water systems like Lynxight read posture and trajectory from the deck. They miss what happens once a swimmer goes under. Underwater systems like SwimEye read body position below the surface, where the high-confidence drowning signal lives. Most aquatic centers do both. HOA and condo pools often start with above-water and add underwater later as budget allows. The consultation names the configuration that fits each pool.

What state laws or federal regulations apply to pool monitoring?

The Virginia Graeme Baker Pool and Spa Safety Act of 2007 is the federal floor: drain covers, vacuum-release safety, applies to public pools and spas across all 50 states. State-level requirements vary. Florida, California, Arizona, and Texas all have residential pool safety codes that govern fences, gates, and alarms but not AI monitoring. AI detection is currently optional; insurers are starting to discount premiums where it is deployed.

What is the realistic false-positive rate, and why does it matter?

Vendor-published rates land in the 1 to 3 alerts per pool per day range during normal use, dropping once the system is tuned. False positives matter more here than in most analytics use cases. If the system pages the lifeguard ten times a day, the lifeguard learns to ignore the page. Tec-Tel's deployment plan tunes confidence thresholds aggressively before pagers go live to keep alert volume low.

Does this replace a human lifeguard?

No, and any vendor that says yes is selling you something the legal team will not sign. Aquatic-center programs use AI as a force multiplier for guards on duty. Unguarded HOA and hotel pools use it as a backstop for hours when no human is watching. The CDC's Healthy Swimming guidelines treat lifeguards as the primary control. AI is a layer, not a replacement.

How does it integrate with our PA system or lifeguard pagers?

Most platforms output webhook or SIP-based alerts that land in standard pager systems, PA control panels, two-way radios over IP, or staff smartphones. SwimEye, Lynxight, and AngelEye Vision all expose APIs for the integration. Tec-Tel handles the alert routing during deployment so the alert lands on the right channel for each shift, not a queue nobody reads after hours.

What does it cost to deploy?

Single-pool installs typically run $25,000 to $80,000 depending on whether the configuration is above-water only, underwater only, or both, plus a software subscription in the $400 to $1,500 per pool per month range. Aquatic centers with multiple bodies of water and competition lanes land higher. Insurance discounts for AI-monitored pools currently average 5 to 15 percent in the carriers we have worked with, which shortens payback.

Can we add this to a pool that is already in service?

Yes. Above-water systems mount on existing poles, building eaves, or dedicated deck poles. Underwater systems install during a routine pool drain and refill, which most facilities do annually anyway. The consultation walks the pool with the operations lead and flags whether the existing electrical, networking, and structural anchor points support the install or whether minor work is needed first.