Bryce Parsons

Product
Designer

  • Kleiner Perkins Design Fellow
  • Prev. Design @ Glean
  • UC Berkeley Master of Design
Bryce Parsons
[SAN FRANCISCO, CA]

Work Experience

Product Design Intern, Assistant & Agents
  • Designed Agents in Assistant, a lightweight agent builder embedded in Glean's chat assistant, enabling users to create, test, and run custom agents; drove 59% adoption in internal trials ahead of GA release to Glean's 1M+ customers
  • Redesigned Glean's composer interaction model to support assistant capabilities, including a new affordance menu, keyboard shortcuts, and hierarchy principles for ordering referenced context and invoked actions
Product Designer
  • Led growth design initiatives redesigning lightweight onboarding, domain-hosting, and sharing flows, targeting conversion barriers from session analytics to drive retention and paid conversion
  • Designed composer shortcuts enabling users to inject project context into AI prompts, improving agent response quality and user flow
Senior Management Consultant, Products & Innovation
  • Led product design and strategy efforts for The Gates Foundation's grant and contract management platform processing $9B in transactions annually.
  • Redesigned the user experience of core product features, implemented core features in their content management system (CMS), and implemented a new brand design system for the leading brand of a Fortune 500 retailer; optimizing user experience and accelerating business results, reducing the time for front-end UI changes by 300%.
  • Designed and developed product performance data visualization dashboards in Tableau for the C-suite of a $100B+ multinational video game company; reducing time to insights by 50%.
  • Created a service design blueprint for a $60B cloud business; increasing operational sales efficiency and driving customer satisfaction.
Data Visualization Intern, Customer Success
  • Designed and developed Tableau BI dashboards that identified overlooked accounts at-risk to churn; enabling customer success managers to make data driven decisions 5x faster.
Professional Work

Glean

Enterprise AI platform

/ SUMMARY

I lead the redesign of Glean's composer — the front door to the assistant feature.

/ TEAM
Executive Stakeholders Christian Ervin, Emrecan Dogan (Head of Product)
Head of Design Christian Ervin
Designers Bryce Parsons (me), Ana Fucs, Kenji Kaneko
PM & Engineering Selene Kim (PM), Sonia Gu (Eng), Chinmay Goyal (Eng), Daniel Sun (Eng)
/ COMPANY SITE
Glean Visit

Problem + Context

Glean's composer left room for opportunity

Annotated breakdown of Glean's composer in its current state

An illustrative breakdown of Glean's composer in its current state, surfacing the inconsistencies and shortcomings of the existing system.

Problem

Glean's current composer (is):

  1. Inconsistent: Actions, create docs or images, and create skills, all follow different patterns.
  2. Doesn't showcase Glean's affordances: The composer only shows a small subset of the features and tools of the Assistant.
  3. Not scalable: As we build more functionality into assistant (e.g., creating agents) we need a principled approach to scale.

Context

As building with Assistant becomes more common, it is important that we intentionally design the composer to support creating artifacts and using system capabilities such as skills and agents directly within the composer. And as Glean's feature set grows, we need a composer-system that can scale with future capabilities rather than layering on one-off patterns.


This work proposes improvements to better surface the capabilities of Glean within the composer, nudging users to use more of Glean to tackle their work.

Exploration, Ideation, and Insights

A principled and craftful redesign

Design Principles

  1. Simplicity & consistency: Unifying the creation experience for all things.
  2. In line creation & referencing UI: Affordances for creating or referencing appear in line.
  3. Mutually exclusive, collectively exhaustive: All affordances of the composer can be enabled via / or @ shortcuts. Each shortcut menu has discrete functions.

Early Explorations

I began by designing the shortcut menus and their interaction model, as shortcuts encourage us to consider all of the potential affordances we might support the user in taking via the composer.

Exploration 1: / for creation of everything, @ for context, running skills and agents
Exploration 2: / for actions such as create and run, @ for adding context

Two different shortcut models: #1 "/" to create any artifact via Glean's assistant, "@" to invoke an agent or a skill or add any context, #2 "/" to create an artifact or invoke an agent or skill, "@" to add any context

Findings

  • Users preferred exploration #2, since they already invoke skills with "/" and see agents as related to skills, so they expected both to live in the same place.
  • Users struggled to understand the difference between agents and skills
  • While "Interactive" artifact adoption is high, users find the artifact name unclear

Results

Clear affordances, a delightful UX, and designed to scale

Addressing inconsistency

All actions and reference context now sit directly in line, with a redesigned set of UI chips to differentiate composer affordances from plain text.

Surfacing Glean's affordances

A considerable part of this work centered on identifying which of Glean's many capabilities should be accessible via the composer, and therefore executable through the assistant harness. Once identified, we determined how these affordances should surface across the +, /, and @ menus, creating heuristics for where each affordance type belongs and ordering items using empirical product data.

+ menu: includes and educates users on “Use” and “Create” affordances and how to tag context in the composer

/ shortcut menu: includes recommended agents and skills to invoke and create actions relevant to each user

@ shortcut menu: includes relevant documents, links, chats, people, and projects for users to add as context

Designing a scalable system

We designed a scalable system to organize and surface the many affordances of Glean's assistant within the composer, the feature's front door. As more affordances are added, we're confident they'll fit the model we've proposed: each one is either something to "Use" or "Create" (/), or something to "Tag" (@).

try it out!
Create, ask, or search anything. Type / for skills, agents, or to create with Glean. Type @ to tag people, docs, or projects.

Glean

Enterprise AI platform

/ SUMMARY

Contributed to the design & development of Glean's new Agents in Assistant feature.

/ TEAM
Executive Stakeholders Christian Ervin, Emrecan Dogan (Head of Product)
Head of Design Christian Ervin
Designers Bryce Parsons (me), Ana Fucs, Kenji Kaneko
PM & Engineering Selene Kim (PM), Sonia Gu (Eng), Chinmay Goyal (Eng), Daniel Sun (Eng)
/ COMPANY SITE
Glean Visit

Problem + Opportunity

Everyone Wants an Agent, but Building One Is too Complex

Overview of the agents problem and opportunity space

A glimpse into the complexity of Glean's current state agent builder.

Problem

Standard users increasingly want to automate basic tasks with agents. But Glean's agent builder, the only path to do this today, is too complex for their needs.

Opportunity

How might we let any user turn a natural-language prompt or successful Assistant interaction into a repeatable unit of work—without requiring them to learn agent concepts or navigate a complex builder?

Exploration, Ideation, and Insights

Removing the barrier to personal agents

A conceptual model

Diagram: Assistant can be prompted to build Agents (formerly automations)

A highlevel model of requirements to guide our initial solution discovery and design

Building agents via assistant

1 / 6

Prompting assistant

Managing your active agents

1 / 2

Viewing all active agents (automations)

Editing agents with the simplified builder

The simplified builder progressively discloses complexity to the user, rather than showing all of the complex configurations up front like the current agent builder.

1 / 5

Viewing the agent in the simplified builder

Promotion

Driving feature adoption

Promotion Framework

  1. Homepage Modal Animation (Primary Promo): All users who have access to create agents see the Homepage Modal on page load one time.
  2. Interactive Composer Card (Secondary Promo):
    • If a user exists out of Homepage Promo without building an agent, show them the Psychic Card Promo on next app load (persists for two session loads).
    • If a user clicks through Homepage Promo and completes building an agent via assistant, do not show them this Psychic Card.
  3. In-chat recommendation banner (Tertiary): Served to all users, dependent on chat context.

1 / 3

Homepage Modal Promo

Insights and results

Iterating through Dogfooding and Beta Testing

Internal Trial at Glean

Starting in July 2026, we ran a six-week internal study, releasing personal agent-building capabilities via Assistant to Glean's roughly 1,000 employees. Results were overwhelmingly positive: users adopted the feature in greater-than-expected numbers despite minimal promotion.

Concurrent Beta Test

We beta-launched the Agents in Assistant feature with a select group of Glean customers. Feedback was positive, particularly around the ease of use and accuracy of agent generation.

Critical Feedback

Two issues surfaced during testing:

  • Placing active agents (automations) in a left-navigation field called "Active" was confusing to the user's mental model and cluttered the navigation panel. This was later addressed by adding the Active section back to the Agent Library.
  • Invoking agents via Assistant (the assistant harness) was difficult without a direct-manipulation way to select an agent by name; users had to recall the agent's name and instruct Assistant to run it. This was addressed by new affordances introduced as part of the Composer Redesign.

Onlook

AI-powered visual-first IDE for prototyping

/ SUMMARY

Designed and shipped a context-adding feature in the prompt window enabling users to inject contextual information into AI prompts.

/ TEAM
Executive Stakeholders Daniel Farrell (CEO), Kiet Ho (CTO)
Design Bryce Parsons (me), Daniel Farrell (CEO)
/ COMPANY SITE
Onlook (YC W25) Visit

Challenge and Background

Challenge

To enable users to effortlessly add contextual information to prompts for more accurate, relevant results in Onlook's Design-IDE AI Chat Interface.

Context

While working at Onlook, I relied heavily on Notion, Cursor, and Slack. I became accustomed to using the '@' feature to add context in those tools that I found myself instinctively trying to use it while prompting in Onlook.


This friction sparked a hypothesis: if I have this muscle memory, our users likely do too.

Concept Map: Initial Problem Space Understanding

Problem Discovery & Validation

Current state experience example

*another example of adding context in the "current state" includes copying the entire block of code for the header component which adds more specificity, but is a poor UX.

Validating the problem

  • Conducted extensive product testing to understand the current state friction of adding context to prompts
  • Interviewed and observed users to understand the most commonly referenced contextual items and their process for adding them
  • Surveyed Onlook's Discord community to understand current state pain points related to adding context to AI Chats

Key findings

  • Manual retrieval forces users to switch contexts, breaking the flow required for rapid, iterative prototyping.
  • Users most frequently need to reference code components (e.g., headers) and brand styles (e.g., colors, fonts).
  • To speed up multi-turn prompting, users need quick access to their most recently used context.

Solution Ideation

Design principles

      Leverage Familiar Patterns: Minimize the learning curve by adopting the interaction models users already rely on in tools like Cursor, Notion, and Slack.

      Balance Human-User Needs with Agent Requirements: Create a simple abstracted interface for users while ensuring the agent receives the deep, structured context it needs to perform accurately.

Design Exploration

Prototyping & Testing

Functional Prototype

I built a functional prototype of the AI Agent Context Feature in code using Onlook's own tool. I then used this prototype to conduct usability tests with users.

Final Conceptual Model

Results

Outcome

AI Agent Context Adding Issue Resolved: Users now can seamlessly integrate context into AI-agent-prompts without interrupting their flow, resulting in an improved user experience, productivity, and increased agent performance.

Final solution in production

Collaborating with Onlook's technical-lead, I implemented the final solution using a combination of Cursor and Codex.

Impact

  • Enhanced AI response accuracy through improved contextual prompting
  • Streamlined user workflow eliminating manual context retrieval
  • Increased user satisfaction and time on project through better AI-agent interaction experience

Onlook

AI-powered visual-first IDE for prototyping

/ SUMMARY

Led end-to-end research, design, and prototyping to increase new-user retention through a revamped onboarding flow, streamlined publishing, and collaborative sharing features.

/ TEAM
Executive Stakeholders Daniel Farrell (CEO), Kiet Ho (CTO)
Design Bryce Parsons (me), Daniel Farrell (CEO)
/ COMPANY SITE
Onlook (YC W25) Visit

Challenge & Background

Challenge

Increase new user retention, conversion to paid subscribers, and grow Onlook's active user base.

Background & Context

In-session analytics revealed critical drop-off points where users abandoned their sessions and workflows before completing key actions in Onlook's product. This indicated insufficient product onboarding and user experience around key features, resulting in poor business growth metrics.


Daniel, Onlook's CEO and Product Design Lead, challenged me to address the retention and growth issues through strategic design decisions.

Concept Map: Creating a Shared Mental Model of the Problem Space

Problem Discovery & Validation

Research Objectives

  • Why do users struggle to discover and adopt Onlook's current features?
  • What are the "core" features that all users should discover?
  • Which features are most important for driving growth? Do they exist in the current state?

Current State UX Friction Points

Findings

  • Onlook onboards some new users manually but does not have a self-service onboarding flow—something users requested.
  • New users perceive the platform as complex with numerous feature offerings.
  • Engagement with "Site Publishing" and "Custom Domain Hosting" features leads to higher user conversion and retention.
  • Onlook's core features: AI Chat, Design Panel (Brand Styles), Code Window, Design-Preview Toggle, Site Publishing, and Custom Domain Hosting.

Solution Ideation: Onboarding Flow

Approach

Addressing initial findings that there is demand for a self-service onboarding flow and that new users perceive the platform as complex with numerous feature offerings—I ideated on the prompt, "How might we create an intuitive and informative core-feature onboarding experience to drive adoption and retention?"

Solution Exploration and Ideation

Solution Ideation: Publish & Share Features

Approach

Addressing initial finding that engagement with "Site Publishing" and "Custom Domain Hosting" features leads to higher user conversion and retention—I ideated on the prompt, "How might we encourage discovery and completion of site publishing and custom domain flows?"


Additionally, "How might we spur new user growth through strategic design and product decisions?"

Solution Exploration and Ideation

Solution Development: Onboarding Flow

Onboarding Feature Overview

  • Flow initializes as the product loads in the background
  • Focuses on core-features
  • Interactive to encourage user-engagement

Onboarding Prototype Demo

Solution Development: Publish & Share Features

Publish & Custom Domain Features

  • Added an attention-grabbing vertical carousel to the Publish/Share button to drive feature discovery and engagement.
  • Enhanced the Publish/Share modal with improved loading states and preview functionality to increase publish flow completion by boosting user confidence and reducing decision fatigue.
  • Sequenced the Custom Domain call-to-action to appear after site publication, capitalizing on completion momentum to drive higher adoption of premium features.

Sharing Feature

Designed a sharing feature integrated into the existing publish modal, drawing on proven design patterns from Figma, Notion, and Framer to create an intuitive experience that encourages viral growth through network effects.

Share Feature Prototype Demo

Results

Outcomes

  • Increased new user retention through improved feature awareness and enhanced publish flow
  • Increased revenue due to higher conversion to paid subscriptions via optimized add custom domain flow
  • Improved collaboration and user growth through frictionless sharing capabilities

Key Success Metrics Defined

  • User retention rate improvements
  • Conversion rate from free to paid plans
  • Publish completion rates
  • Custom domain addition rates
  • User session duration and feature adoption
  • Monthly Active Users (MAU)
Work I did in design school

Archetype

Simulated user testing with synthetic users

/ SUMMARY

Building a tool for product teams to test software with their target customers using synthetic users.

/ TEAM
Role Design & Product Lead
Technical Lead Hanyang Gu
/ COMPANY SITE
Synthetic Archetype Visit

About

Archetype is a product and design research tool that models synthetic users based on your product's exact customers. This work addresses the opportunity space for synthetic users to effectively model usability studies on coded prototypes. Synthetic personas are initially derived from user behavior frameworks, then continually refined based on session recordings that expose real user behaviors through the PostHog API. The Archetype platform consists of a web app for managing synthetic personas and a plug-in for the CLI tool, Claude Code, enabling software builders to quickly run experiments on prototypes as they build.


As of September 2026, I am no longer working on Synthetic Archetype.

Achievements

  • Accepted into Berkeley Skydeck's Pad-13 Accelerator Program (Summer 2026)
  • Top 10% of applicants for YC W25

Additional Information & Resources

#7???

#8???