Company: Engivio
Product: Care Jane – AI Recovery Companion for Post-Knee Surgery Patients
Industry: Healthcare AI
The Challenge
As a fast-moving healthcare startup with a top team, Engivio is building Care Jane, an AI companion designed to support patients recovering from knee surgery. While they have strong healthcare domain expertise, implementing robust safety, compliance, and security guardrails for generative AI posed a significant challenge risking to derail them from their core product and business.
In the high-stakes world of healthcare, large language models (LLMs) can become a liability if not properly constrained. Too few controls and the assistant risks generating unsafe or non-compliant advice. Too many, and it becomes too limited and useless in practice. Yet, manual review of every AI interaction is simply not scalable.
The Solution
To protect their product and ensure user experience, Engivio integrated Alinia Guard API, which offers business-specific and regulation-informed GenAI guardrails for compliance, accuracy, safety, and security. With minimal effort, they embedded real-time input and output controls directly into their AI workflow.
Thabet Burias, CTO at Engivio: “I really like the API, it’s super easy to plug in. It took me just two hours to set up. What surprised me is that adding Alinia Guard didn’t slow us down—it actually sped up our average response time. Risky queries are flagged and politely declined immediately, so we don’t waste processing power or generate non-compliant responses.”
Alinia’s guard models have been trained and tested with both internal and external datasets to best calibrate detector sensitivity. However, Alinia recognizes that customization is key and each business context and use case might require different thresholds.
“The good thing about the Alinia API is that it gives you the probabilities and you decide at which point you want to block medical advice or when to allow helpful content in the medical space (threshold)” shares Burias.
While Alinia specializes in compliance and regulatory guardrails, it also covers customer’s Security and Safety needs.
“It’s great at preventing adversarial attacks, that was one thing I was really afraid of before we integrated it: how am I going to protect the AI from being infiltrated with these type of questions? It would be expensive to hire a specialist security consultant for that specific purpose.
Engivio even benefited from Alinia’s recently released Investment Guard:
“We even tried asking our AI to give us nuanced investment advice, and Alinia detected it!” praises Burias.
Measurable Benefits
- Regulatory Peace of Mind: Ensures AI outputs stay within the safe boundaries of medical compliance.
- Faster Response Times: Risky queries are filtered right away, streamlining the user experience.
- Lower Operating Costs: Prevents unnecessary processing and reduces manual oversight needs.
- Customizable Risk Thresholds: Engivio can tune the system based on medical-specific risk sensitivity.
- Stronger Security: Proactively blocks adversarial inputs and prevents system manipulation.
Why It Matters
In regulated industries like healthcare, trust is everything. By embedding Alinia Guard into their product stack, Engivio didn’t just add guardrails—they unlocked the ability to scale safely, with confidence and control.