PROVIZ / Medical AI · NTNU
Taking AI from research to the radiology workflow.
Leading PROVIZ through product discovery, funding, regulatory preparation, and prospective clinical validation.
- My role
- AI Product & Project Lead; Innovation & Product Manager
- Period
- Product & innovation leadership, 2022–2025
- Status
- Prospectively validated · Patent pending
patients enrolled
From early-stage research to a product evaluated in a real clinical workflow.
The product challenge
PROVIZ is an AI decision-support system for prostate MRI. It processes scans and produces probability maps to help radiologists identify suspicious tissue. The product challenge was to bring that capability into a clinical workflow, with the evidence, documentation, and collaboration needed to use it responsibly.
My work at NTNU spanned product and innovation leadership: defining the concept and roadmap, securing funding, coordinating the team, and moving the product through prospective validation and commercialization preparation.
What I owned
Product direction
Translated feedback from 25+ users and stakeholders, and analysis of 20+ competing solutions, into requirements, priorities, and a product roadmap.
Funding & execution
Secured and managed 8M NOK in early-stage funding to support productization, validation, and commercialization activities.
Clinical & regulatory work
Designed and ran the prospective study; coordinated clinicians, IT, and engineers; and delivered technical documentation, risk-management work, and approvals needed for clinical evaluation.
Commercialization
Led IP work with NTNU Technology Transfer, developed the product proposition, and initiated collaboration with 10+ companies, generating 3+ qualified partnership opportunities.
Connecting product work to clinical reality
Define the value around the radiologist’s workflow.
Customer discovery, competitor analysis, and stakeholder workshops helped translate a research capability into a product concept. This work connected target users and clinical needs to requirements, milestones, and a commercialization pathway.
Build the evidence alongside the software.
The work extended from algorithm and pipeline development to regulatory documentation and prospective clinical evaluation. Coordinating those activities meant treating validation and clinical integration as part of the product roadmap.
Bring technical and commercial work together.
I worked with the technology transfer office on patenting and licensing considerations, while engaging potential industry partners. Participation in REACH and The Forge supported the project’s development toward commercialization.
What the work achieved
A progression from early-stage research to prospectively validated software, supported by funding, clinical collaboration, and regulatory preparation.
The prospective study was published in Insights into Imaging in January 2026. The published analysis included 89 patients for feasibility and safety, and 76 for diagnostic performance. These analysis populations are distinct from the overall enrollment figure above.
The study reported a patient-level AUC of 0.90 for the software and 0.86 for the radiologist; the difference was not statistically significant. The findings support further evaluation of the technology’s clinical potential. Read the published study ↗
NTNU Technology Transfer lists PROVIZ as patent pending and available for licensing. View the technology profile ↗
Explore the evidence
Next project