AI Assisted System Flow Diagram
Context
This project explores integrating an AI powered virtual assistant into the TreeRing Yearbook platform to simplify user support, improve efficiency and reduce support costs. I worked as Senior Product Designer over a four month sprint leading the strategy, design and technical framing to create a scalable solution that aligned with business goals.
Problem
-
Users submitted many repetitive support questions because platform information was difficult to find.
-
Customer support was overloaded with routine inquiries which increased staffing needs and operating costs.
-
There was no unified self service experience which created user frustration and slowed down task completion.
User Scenario Diagrams
User & Competitive Research Mural Hub
Why It Matters
The growing number of support requests affected both the platform experience and the business. Without a more efficient model users would remain frustrated and support costs would continue to rise. An AI assistant offered a sustainable way to reduce support load, improve the user journey and add a modern scalable support layer before busy yearbook seasons.
My Role & What I Did
-
Led discovery by mapping current user flows and identifying patterns from user and support data.
-
Assessed technical approaches by comparing a custom AI solution with an API based accelerated option to balance speed, cost and scalability.
-
Created a clear systems framework showing how AI context layers and response logic would work across the platform.
-
Designed conversational flows, wireframes and prototypes for an intuitive natural language chat experience.
-
Built a phased roadmap that outlined immediate value, long term scalability and future expansion opportunities.
First Figma Iteration
NLP & Response Logic Engine Thinking
Key Decisions & Strategic Moves
-
Selected a flexible model that used existing AI APIs for immediate results while leaving room for custom intelligence later.
-
Aligned engineering, design, support and business teams by creating clear artifacts that explained how the assistant would function within the platform.
-
Designed the assistant to respond to natural language intent rather than using rigid button based flows.
Visuals & Technical Tasks
-
Recreated the current experience to expose usability constraints and connect them directly to research insights which uncovered a new opportunity and direction for building the intelligent virtual assistant.
-
Evaluated multiple AI assistant models including in house, hybrid and fully outsourced solutions and assessed their strengths and tradeoffs which allowed the team to choose an approach that balanced speed, cost and long term scalability.
-
Researched and selected the right tools for each stage of the assistant, documenting costs, limitations and decision criteria to guide engineering and leadership.
-
Defined the capabilities and limitations of the assistant by creating the conceptual framework, response logic structure, contextual layer and sentiment handling flow which ensured the system aligned with real user frustrations and product goals.
-
Designed the visual direction and the full set of user interactions that shape how people engage with the assistant across the platform.
-
Built the foundation for a scalable AI assistant and created a complete version based roadmap that supports ongoing growth, iteration and future product innovation.
Scalable Solutions and Designs Specs
API Guardrail Specs and Edge Case States
API Curation Diagram
Impact & Expected Benefits
-
Reduced support load by shifting routine questions to the assistant which allowed human agents to focus on complex issues.
-
Improved usability by giving users a fast accessible way to get help without searching through pages of information.
-
Positioned TreeRing for sustainable growth with a scalable support layer that lowers long term operational costs.
-
Established a path for the assistant to evolve into a core platform feature that supports smarter workflows and guided creation.
-
Phase two expands the assistant into workflow support and platform actions. Phase three introduces proactive guidance that anticipates user needs which increases efficiency and improves the overall creation experience.
Want to get in touch? Chat? See pictures of my dog?
