We employ risk identification and assessment processes for every phase and individual development initiative. We use a clinical feedback loop to continually develop and improve the clinical model to improve accuracy. Throughout the development cycle we work cross-functionally to ensure technical and architectural components work together in a way that continually delivers on clinical performance and safety. We implement in a phased and controlled way to ensure clinical safety. Our clinicians undertake continuous evaluation of patient cases to enable rapid identification and response to anomolies.

The model is designed to ensure appropriate transparency and so that the decisions it makes are traceable and easy to understand. All changes to the model are traceable and managed by humans. The system is intentionally designed not to be self-learning to minimise the risk of bias. The development of interfaces is data-driven and decisions about designs are traceable and explainable. There is clear visibility of data used in the machine learning model. The entire development process is evidenced in technical documentation.

Robustness and availability are ensured through the design of the architecture and infrastructure, along with well-defined processes for maintenance, testing, tracking and responding to issues. The architecture is designed to support scalability, resilience and fault tolerance, ensuring the system can handle varying loads and recover quickly from any failures. Protecting data and patient privacy are cornerstones of our work, with clear processes around data protection and handling. We adhere to industry standards that enable us to ensure balance between benefits and risks in a structured and accountable way. This compliance provides a framework for making informed decisions, maximising positive outcomes and mitigating risks.

We employ a holistic approach to understanding and solving the problems that exist for users across the healthcare system. Cross-functional teams are involved in the development of our human-centric designs, using concrete goals for fairness and inclusion and tested at all phases for usability. We implement in a phased manner with evidence collection that includes evaluations with users in the early phases before full roll-out. We work to consciously address potential sources of bias.

Responsible AI, always

When choosing AI for healthcare, trust matters. Visiba’s Responsible AI Principles explain how we develop secure, ethical, safe, and transparent solutions. Discover how we’re building AI you can trust.

Visibas four Responsible AI Principles

Clinically safe

Healthcare professionals, patients and commissioners should feel confident that our solution is clinically safe.
Secure

In addition to complying with relevant regulations, confidentiality, integrity and availability are cornerstones of our work.
Transparent

Our technology should be developed in a way that affords an appropriate level of transparency and explainability.
Ethical

The technology should work in a way that enhances and does not detract from the human element of healthcare.

Our commitment to high quality healthtech

We've been working with innovative healthcare providers to deliver high quality healthtech since 2014. When we launched our AI-enabled triage solution in 2018, we took the same quality-first approach. This approach was heavily guided by input from our medical team and clinicians using the system, alongside our fundamental belief in the importance of working responsibly.

Why Responsible AI Principles?

Marcus Olivecrona, Chief AI Officer at Visiba, introduces our Responsible AI Principles and why they are important to us and our customers.

Why Responsible AI Principles?

Marcus Olivecrona, Chief AI Officer at Visiba, introduces our Responsible AI Principles and why they are important to us and our customers.

Play video

Examples of our Principles in practice

These principles are embedded in every phase of our development process in the form of 'practices' and form an integral part of our daily work.

Clinically safe

We employ risk identification and assessment processes for every phase and individual development initiative.
We use a clinical feedback loop to continually develop and improve the clinical model to improve accuracy.
Throughout the development cycle we work cross-functionally to ensure technical and architectural components work together in a way that continually delivers on clinical performance and safety.
We implement in a phased and controlled way to ensure clinical safety.
Our clinicians undertake continuous evaluation of patient cases to enable rapid identification and response to anomolies.

Transparent

The model is designed to ensure appropriate transparency and so that the decisions it makes are traceable and easy to understand.
All changes to the model are traceable and managed by humans. The system is intentionally designed not to be self-learning to minimise the risk of bias.
The development of interfaces is data-driven and decisions about designs are traceable and explainable.
There is clear visibility of data used in the machine learning model.
The entire development process is evidenced in technical documentation.

Secure

Robustness and availability are ensured through the design of the architecture and infrastructure, along with well-defined processes for maintenance, testing, tracking and responding to issues.
The architecture is designed to support scalability, resilience and fault tolerance, ensuring the system can handle varying loads and recover quickly from any failures.
Protecting data and patient privacy are cornerstones of our work, with clear processes around data protection and handling.
We adhere to industry standards that enable us to ensure balance between benefits and risks in a structured and accountable way. This compliance provides a framework for making informed decisions, maximising positive outcomes and mitigating risks.

Ethical

We employ a holistic approach to understanding and solving the problems that exist for users across the healthcare system.
Cross-functional teams are involved in the development of our human-centric designs, using concrete goals for fairness and inclusion and tested at all phases for usability.
We implement in a phased manner with evidence collection that includes evaluations with users in the early phases before full roll-out.
We work to consciously address potential sources of bias.

Clinically led development

The medical team drive the development of Visiba Triage to ensure it is clinically safe. Leaning on extensive clinical experience they review hundreds of real-world patient cases weekly and use this information, along with feedback from practising clinicians, to develop a solution they want to use for their own patients.

Dr Annabelle Painter

Chief Medical Officer

Dr Christer Rosenberg

Senior Medical Specialist

Dr Katherine Leung

Clinical Product Owner

Dr Joel Ellbin

Medical Specialist

Dr Marcus Olausson

Medical Consultant

Dr Edwin Gidestrand

Medical Consultant

Dr Mats Halldin

Medical Consultant

Dr Ashiv Patel

Medical Specialist

Dr Martha Martin

Medical Specialist

Dr Kristina Bergman

Medical Specialist

Want to learn more?

Interested in understanding more about AI in healthcare in general and AI triage in particular? These informative videos from our knowledgeable medical and AI teams should help.

Responsible AI, always