Dressed in a lab coat and glasses, with a warm and friendly demeanor, Dr. Pixelpod is a digital tool designed to support healthcare professionals as part of the care team in a hospital. While it may resemble a nurse or doctor in appearance, Dr. Pixelpod serves a distinct role: it is a digital avatar that enhances patient engagement and hospital workflows.
Dr. Pixelpod is the first digital human avatar (DHA) created by QuantumBlack, AI by McKinsey, the firm’s AI arm. Digital avatars like Dr. Pixelpod are designed to complement human professionals by using AI, 3D modeling, and motion capture to facilitate interactions and improve access to information—ultimately enhancing the efficiency of healthcare teams.
They have enormous potential to provide a personalized level of interaction that resonates with end-users and works well across a variety of roles within industries. Avatars can also serve as travel advisors, customer service representatives, entertainment characters, and tutors.
They help organizations extend the reach of services and administrative tasks and can free up time for healthcare workers for higher-value activities like direct patient interaction.
How McKinsey uses NVIDIA’s Digital Human Technology
NVIDIA supplies the blueprint and core components—leading-edge gen AI software and the high intensity computing—needed to create avatars that can perform streamlined and continuous interactions.
McKinsey has developed the technology and expertise for training and deploying the avatars through a cloud-based platform that can be implemented at client sites. It serves as a one-stop shop: On day one, client teams can select from a broad range of avatars, created using NVIDIA AI, and then work with our QuantumBlack engineers and technologists to train them to fit specific use cases. In addition to practical deployment expertise, our proprietary agentic tooling brings new depth and realism to the avatar, extending use cases beyond the standard Q&A style interactions.
“We're looking at a new type of human interaction with a machine that will be an important component of rewiring businesses to become AI-centric,” explains Rickard Ström, a QuantumBlack data scientist. “With our industry expertise, we bring a deep understanding of individual use cases and are able to create the best avatar solution possible, including the change management and capability building needed so it’s deployed for maximum value.”
We help clients evaluate their avatar use case in terms of business, technical, and ethical considerations. How will it impact revenue? What can it produce in terms of efficiency, savings, service quality? Is the right technical infrastructure in place? And in the healthcare setting like the one where Dr. Pixelpod is deployed, what the impact could be on the patient experience and health outcomes.
In partnership with our clients, we also implement conversation guardrail mechanisms, providing extra safeguards to mitigate risks and ensure related topics are addressed when discussing a particular subject.
If a relevant topic doesn’t naturally arise in conversation, the avatar is trained to introduce it.
McKinsey’s digital human solution also emphasizes responsible AI use, factoring in compliance with data privacy standards and minimizing algorithmic bias, ensuring ethical deployment in all sectors.
Training an avatar typically involves three phases over approximately 12 weeks:
- developing the knowledge and memory
- refining appearance and behaviors including realistic facial expressions, gestures, and voice
- integrating industry-specific tasks, for example, a real-time booking capability for a travel assistant.
“We are excited about our work with NVIDIA because of the positive impact it can have for consumers across industries,” says Jess Lamb, a partner within QuantumBlack and McKinsey’s healthcare practice. “Used appropriately, this technology could make the healthcare system more understandable and approachable, improving the quality and experience for many.”