The human context: why safety in human services is different
In human services, safety isn’t just about security systems or compliance checklists. It’s about protecting real people in deeply human moments. The individuals supported in this field often have diverse needs that sometimes require additional support during moments of stress or change. Staff members are trained to respond with care and professionalism, but even the most skilled teams can only act on what they can see in time.
Unlike hospitals or corporate campuses, most human services organizations operate in distributed settings such as group homes, day programs, classrooms, and clinics. Cameras help, but no one watches them around the clock. When an incident occurs, it can take hours or even days to find the right clip. That delay matters. In those unseen moments, stress compounds, trust erodes, and risk increases.
This is the context in which Amego Inc., a Massachusetts-based nonprofit serving individuals with disabilities, decided to act.
The challenge at Amego: 30 hours to find one incident
Amego supports hundreds of individuals across group homes, day services, and clinical programs. Their challenge was clear: identifying incidents quickly enough to respond in real time.
In one instance, it took more than 30 hours for three staff members to review camera recordings just to find a single event. Without anyone monitoring footage around the clock, the process was reactive and exhausting.
It didn’t align with Amego’s mission to provide compassionate, responsive support. They needed a way to surface incidents as they happen, not hours or days later.
The organization began looking for AI safety technology for care facilities that could help detect aggression, self-injurious behavior (SIB), or abuse in real time without compromising privacy.
The decision: choosing Oddity.ai
When Amego began evaluating solutions, they found that many vendors simply weren’t designed for the realities of behavioral health or group home environments. “Most organizations do not care for our industry,” the team recalls. “Once we reached out to each and every organization, we came to the realization that they were not willing to work with us. But Oddity was the only company that was willing to work with us and try to accommodate our needs.”
That willingness to adapt made the difference. Oddity.ai’s aggression detection software is built specifically for high-sensitivity environments, from AI monitoring for group homes to video analytics in healthcare. The solution integrates securely with existing camera systems and flags physical aggression within half a second. That speed gives teams a crucial head start.
For Amego, the goal wasn’t technology for technology’s sake. It was to make sure that every staff member could respond faster, with greater confidence, and with better outcomes for the people in their care.
The collaboration: plug & play integration and rapid feedback
The rollout was seamless. “Integration was smooth, pretty much plug and play; via a secured network to the cloud and you’re good to go,” the team explains. Oddity.ai’s platform connects over encrypted networks to ensure compliance and security. There is no third-party data access and no external storage, keeping everything private.
Beyond deployment, what stood out to Amego was the collaboration. “Communication with Oddity is constant. Quick replies; recommendations are implemented,” they share. Feedback loops are tight, meaning Amego’s observations translate directly into model improvements and workflow optimizations.
That responsiveness matters, especially in a pilot phase. The system quickly began identifying real incidents such as abuse, restraints that turned violent, and self-injurious behaviors. At the same time, Oddity’s engineering team worked with Amego to fine-tune sensitivity thresholds and reduce false positives, ensuring alerts were both timely and actionable.
The result: incidents visible in real time, calm in the team
Within the first month, the system brought to light incidents of aggression and self-injurious behavior that might previously have remained unseen or been discovered long after the fact.
Now, instead of spending hours reviewing footage, Amego’s staff can act in real time. The result is not just time saved, but a stronger sense of safety and calm across the team.
As the team summarizes it plainly: “The solution meets its objectives. It has the potential to become a stand-out in the industry.”
By combining real-time alerts with human oversight, Amego maintains both accountability and compassion, the essence of human services work. AI doesn’t replace the caregiver; it empowers them.
Reflection from Amego: honest, grounded, human
Amego views Oddity.ai not just as a vendor but as a partner in innovation. The AI is still evolving, but the direction is clear. “It’s meeting its priorities and its use case,” they explain. “It’s still maturing with feedback, but it’s very promising.”
The team is particularly focused on refining detection balance, reducing false positives while maintaining high sensitivity. “Once those are addressed, we’ll get to a balanced point where we’re all in agreement on the alerts that are coming through.” That iterative approach reflects Amego’s culture: thoughtful, data-driven, and centered on the wellbeing of both clients and staff.
Why this model works for the whole sector
Across the human services landscape, organizations face similar challenges. Distributed sites, high staff-to-client ratios, and unpredictable behavior patterns make constant supervision impossible. AI for behavioral health is emerging as a game changer, not to monitor people but to detect moments that matter.
Oddity.ai’s model works because it’s built for these environments:
- Real-time detection (<0.5 seconds): alerts staff the instant aggression occurs.
- Privacy-first architecture: data remains within a secure cloud, fully encrypted.
- Plug-and-play integration: works with existing camera systems.
- Continuous learning: feedback from partners like Amego helps refine detection models.
This balance of speed, privacy, and collaboration ensures that technology enhances care without disrupting it. As Amego’s pilot shows, AI in special needs care can bring visibility, reassurance, and faster responses where they matter most.
Case snapshot
- Organization: Amego Inc. (Massachusetts, USA)
- Sector: Human Services & Behavioral Health
- Use case: Real-time detection of aggression, self-injurious behaviors, and abuse
- Integration: Existing camera network, secured VPN connection
- Impact: 30+ hours saved per week with instant physical aggression alerts, faster responses, fewer blind spots
FAQ: AI safety technology in human services
-
How does Oddity.ai protect privacy when using video analytics in healthcare?
Oddity.ai operates on a fully private, encrypted cloud environment. Footage is therefore always safe. The system detects aggression patterns, not identities, so privacy remains intact. -
How fast can Oddity.ai detect incidents?
Detection occurs in under 0.5 seconds, allowing staff to react almost instantly. This is up to ten times faster than traditional manual monitoring or post-incident review. -
Can it integrate with existing camera systems?
Yes. Amego’s experience shows the integration is plug-and-play. Oddity.ai connects with existing cameras directly or through an NVR or a VPN. -
What types of incidents can be detected?
The system identifies physical aggression, self-injurious behaviors (SIBs), and certain restraint incidents that escalate into violence. Each deployment is tailored to the organization’s risk profile. Oddity.ai is now working on improving the detection of these and other behaviors. -
How does this support caregivers and staff wellbeing?
AI for caregiver support means fewer hours spent searching footage and more time providing care. Real-time alerts reduce stress, increase confidence, and give teams peace of mind.
Faster to the moment that matters
Amego’s journey shows what happens when innovation meets empathy. By combining AI precision with human judgment, Oddity.ai helps organizations see the unseen, respond sooner, and build safer environments for everyone involved.
Interested in how Oddity.ai strengthens safety in care organizations? Plan a demo.
