Case 01
Building an AI Coaching Platform for a Global Psychology Institution
Client: VIA Institute on Character
Engagement: AI Advisory & Product Development
The VIA Institute on Character is one of the world's leading positive psychology organizations — a Cincinnati-based non-profit behind the VIA Character Strengths framework, used by researchers, therapists, coaches, and educators across 190 countries. With nearly 35 million surveys taken globally, VIA's science is among the most validated and widely adopted in its field.
They wanted to explore whether AI could deliver meaningful, personalized coaching around character strengths development. Not a chatbot that answered questions. Something clinically grounded, safe enough for real users, and useful enough to actually change behavior.
The challenge wasn't technical. It was epistemic.
What does responsible AI coaching look like when the subject is something as personal as character? How do you build something that satisfies clinical psychologists, holds up under stress testing, and genuinely serves the person using it? These aren't problems a language model solves on its own.
Working directly with the CEO and a multidisciplinary team — clinical and positive psychologists, coaches, technologists, and a physician — we took a safety-first approach from the first prototype. Working across successive generations of OpenAI's models, we developed the prompting architecture and coaching mechanisms, built guardrails for user safety, and designed a moderation process that could operate at scale without compromising the clinical integrity of the experience.
The work was deliberately iterative. We ran quality control tests, stress tests, and proof-of-efficacy evaluations before anything reached a real user. We built a contextual data platform to give the AI meaningful grounding in VIA's framework. We presented the project at the 11th European Conference on Positive Psychology in 2024.
The result was Viana — VIA Institute's first commercial AI pilot, launched as a 12-month program. The first AI coaching product from a globally respected psychology institution, built with the rigor that reputation demands.
"The VIA Institute on Character embarked on a challenging project to create an AI character strengths-based coach to elevate the best within ourselves to inspire human flourishing globally. Our partnership with MADNESS has been transformative, unveiling the nuanced blend of creativity and analytical precision that defines Prompt Engineering. The profound importance and potential of this field became evident through our collaboration. The MADNESS team dedicated themselves to thoroughly grasping our business mission and objectives, which they masterfully leveraged to surpass our expectations."
Jeffrey Levy, MD
CEO, VIA Institute on Character
What This Engagement Reflects
Proof matters more than pace.
Some AI hypotheses are worth testing carefully, with the right people in the room, over the right amount of time. Speed is not always the objective.
When the product will interact directly with people around something as personal as who they are, proof matters more than pace.
Have an AI hypothesis worth testing?
An Explore engagement gives you real evidence before you commit budget to a full build. Working prototype plus go/no-go recommendation in one to four weeks.
Not sure where to start?
Tell us what you're dealing with. We'll tell you which engagement makes sense — or whether we're the right fit at all.
Let's Talk →