From Blueprint to Business: AI Surgical Video Summarization
Designing and launching a hybrid physical-digital service for surgical video management
Results
Delivered an AI-driven service that shortens surgical videos for multiple contexts including clinical trials, ENT procedures, and conferences.
Reduced review times for clinicians and researchers by targeting only relevant video segments.
Created a scalable service blueprint that integrates with existing OR workflows and compliance systems.
Enabled potential storage cost reductions of more than 60 percent through summary-first archival.
Context and Vision
Surgical videos are critical for clinical research, training, and peer collaboration. They are also long, storage-heavy, and hard to navigate. In our HCDE master’s capstone, our team (Chronicle) investigated how to make these videos more useful without adding friction to already complex surgical workflows.
Our vision was to create a service that would deliver concise, context-relevant summaries without compromising compliance or clinical detail. The capstone ended with a validated service blueprint and MVP concept. Immediately after graduation, my teammate Jack and I spun the idea into a startup called Brevity AI, expanding the concept into a full-fledged physical-digital service.
Part 1: Chronicle – Discovery and Blueprinting
Understanding the Current Workflow
Our first step was to map the journey of surgical videos from capture to review. We documented both physical and digital touchpoints, including how videos were retrieved from the surgical tower, uploaded to secure storage, and later accessed for clinical or educational purposes.
Artifact: Workflow diagram
Caption: Early mapping of the surgical video lifecycle, from in-OR capture to secure storage and review.
Research Approach
We combined secondary research, contextual inquiries, and in-depth interviews with surgeons, an AI developer, and a clinical trial coordinator. This gave us insight into both the frontstage experience (surgeon interaction with tools and systems) and the backstage dependencies (hospital IT systems, storage processes).
Artifact: Research methods overview
Caption: Multi-method research plan integrating observational studies, interviews, and literature review.
Key Insights
We identified recurring pain points:
Inefficient review processes due to long, non-indexed footage.
Under-usage of stored videos because of retrieval complexity.
Risk of compliance breaches when sharing raw footage.
Artifact: Insights summary table
Caption: Pain points aligned to opportunities for intervention at specific touchpoints.
Mapping the Ecosystem
We mapped the ecosystem by creating a detailed user flow that captured every step in the surgical video process, from in-OR capture to post-surgery review and sharing. This helped us visualize where different users such as ENT residents, attending surgeons, and support staff would engage with the system.
From there, we developed early sketches and storyboards to illustrate how the service would fit into real surgical and post-operative workflows. These artifacts gave us a clear picture of how the physical and digital components needed to connect, and where the main opportunities for intervention were.
Artifact: Sketching and storyboards, user flow
Caption: Visual exploration of interactions and service touchpoints for the surgical video workflow.
Defining the Concept and Final Design
The synthesis of feedback from sketching, storyboarding, and user flow explorations led to alignment on a core set of 13 user interactions that became the foundation for our design work. These interactions, documented below, reflected the most critical actions users needed to take within the tool.
Artifact: User interactions list
Caption: Agreed set of 13 priority interactions guiding the design direction.
From there, we developed wireframes to establish the basic structure and layout of each key screen. These appear below and gave us a shared visual language for stakeholder discussions.
Artifact: Wireframes
Caption: Early structural layouts exploring arrangement of navigation, video playback, and annotation features.
We then moved through multiple rounds of design iterations, refining functionality and visual hierarchy. The designs below show this progression, from mid-fidelity layouts to more refined compositions, including feedback incorporated from peers, instructors, and our sponsor.
Artifact: Design iterations
Caption: Progressive refinement of the UI through mid- and high-fidelity prototypes.
To validate these changes, we conducted evaluative testing with both peers and our sponsor. The artifacts below capture this stage, highlighting observations, feedback, and adjustments made to improve usability and alignment with surgeon workflows.
Artifact: Evaluative testing sessions
Caption: Testing with sponsor to refine interaction flow and visual clarity.
This process culminated in a set of key screen flows that included the landing page, review bookmarks, search, and publish views, as well as a diagram for visually distinguishing AI-suggested, manually created, and bookmarked clips. Accessibility considerations were integrated before delivery of the final high-fidelity designs.
Artifact: Key screen flows and final high-fidelity designs
Caption: Final design package showing the product’s core interactions, visual patterns, and accessibility considerations.
The Chronicle phase ended with a concept for an AI-powered highlight reel tool, embedded into existing hospital workflows and compliant with privacy requirements.
Part 2: Brevity – Extending and Scaling the Service
Continuing the Journey from Chronicle
Following the Chronicle capstone, my co-founder and I moved the concept into an operational environment under our new company name, Brevity AI, Inc. Our focus shifted from validating the concept with ENT surgeons to designing and orchestrating a service that could operate within the constraints of real hospitals, regulatory bodies, and diverse surgical contexts.
We joined the UW CoMotion Labs incubator, which provided access to compliance experts, legal counsel, and health tech advisors. We also engaged Microsoft’s startup hub and Nvidia Inception to support the technical side of the service.
Expanding Research into Stakeholder Engagement
Where Chronicle centered on end users (surgeons and residents), Brevity required broadening the lens to the wider service ecosystem. We conducted targeted stakeholder interviews with:
An AI developer from the original sponsor team
A clinical trial coordinator
Healthcare lawyers specializing in compliance
Representatives from the UW Institutional Review Board
Advisors from the UW CoMotion incubator
These conversations clarified constraints that would directly shape the service blueprint, such as IRB approval timelines, encryption standards for PHI, and the operational realities of retrieving video from OR towers in busy surgical schedules.
Artifact: InitialIdeasonFlow.pdf
Caption: Updated service flow integrating physical OR video retrieval with secure digital upload and AI processing.
Mapping New Service Contexts
In Chronicle, the scope was limited to ENT surgical workflows. Brevity needed to adapt the service for three additional contexts:
Clinical Trials – aligning summaries with trial protocols for regulatory review
Conferences and Education – producing concise highlight reels for presentations and teaching
Other Surgical Specialties – exploring adaptation beyond skull-base procedures
To support these contexts, we mapped the expanded ecosystem, showing how frontstage surgeon activities connected to backstage processes like compliance review, storage optimization, and AI model configuration.
Artifact: Possiblestorageoptimization.pdf
Caption: Backstage process map showing how storage costs could be reduced while maintaining compliance and retrieval ease.
Prototyping the Multi-Context Service
We developed tailored service flows for each context:
Clinical Trials – Summary-first review to speed IRB and sponsor sign-off
Artifact: shortenedvideosforclinicaltrials.pdf
ENT Procedures – Highlight reels focused on relevant anatomy and procedural steps
Artifact: shortenedvideosforent.pdf
Conferences and Education – Condensed clips for live presentation and recorded lectures
Artifact: shortenedvideosforconferences.pdf
Each flow used the same AI processing pipeline but differed in onboarding steps, compliance reviews, and delivery methods. This reinforced the need for a flexible service backbone capable of supporting varied frontstage journeys without duplicating backstage systems.
Pilots and Service Validation
Our first live pilot tested the consultation improvement use case. Residents recorded endoscopic consults and generated AI summaries for attending review. These were evaluated against standard consult notes.
Result: AI-generated summaries received unanimous surgeon confidence scores of 5/5, with one attending noting, “Consult notes alone are essentially useless. Normally, I would have to scope the patient again. Your videos are significantly better.”
The pilot validated the value of the summaries in decision-making, while revealing integration challenges such as aligning retrieval timing with surgical scheduling and ensuring compatibility with existing storage systems.
Service Design Considerations
Brevity evolved into a hybrid service combining:
Frontstage: OR video capture, secure upload, AI preview, role-specific review interfaces
Backstage: AI segmentation, compliance checks, encryption, and optimized storage workflows
Support Layers: IRB engagement, legal compliance, onboarding for hospital staff
This layering allowed us to coordinate complex interactions between human roles, systems, and regulations without fragmenting the service experience.
Constraints and Outcomes
The AI model’s specialization in skull-base procedures limited immediate scalability. Expansion would require retraining on other surgical types, which was not feasible without funding.
Despite strong pilot results, the combination of legal, compliance, and operational costs—plus the lack of immediate funding—led to the decision to pause operations in mid-2025.
Reflection
From a service design perspective, Brevity demonstrated the shift from a contained prototype to a functioning service ecosystem. It reinforced the importance of:
Designing for compliance from the outset rather than bolting it on later
Maintaining a flexible service backbone to serve multiple use cases
Building relationships across the ecosystem, not just with end users
Recognizing that sustainability in healthcare services is as much about operational feasibility as it is about user value
If relaunched, I would expand the blueprint to include onboarding protocols for new hospitals, deepen EHR integration, and embed measurement points to track service impact on review speed, compliance events, and storage costs.