Growth Flow

A mentorship workspace where AI handles the notes, the tasks, and the follow-through.

0 → 1

EdTech

SaaS

MixedMethods

Deployed

Project Overview

Most of what mentees gain in a mentorship session gets lost afterward — notes are messy, advice is forgotten, and there's no way to see progress. GrowthFlow is an AI workspace that helps with all of it: before, during, and after every session.

Role

UX Researcher · Product Designer · Developer · Sole end-to-end

Duration

4 months

Status

Deployed – [Live prototype ↗]

Tools

Figma · Figma Make · Dovetail · Cursor · GitHub

Problem

Mentees waste the start of every session explaining who they are. During the call, they're too busy writing notes to really listen. Afterward, the advice just sits there – they don't know where to start. And over time, they can't tell if they're actually improving, so they give up.

Opportunity

How might we help students and career switchers stay present during mentorship sessions and turn mentor advice into clear, trackable progress?

This question became the starting point for GrowthFlow.

Impact

Usability testing showed that the AI-assisted experience helped users complete mentorship tasks faster, with less mental effort and fewer mistakes.

52% faster task completion

Users completed key mentorship tasks in nearly half the time compared to the manual flow.

79% lower cognitive load

Participants reported much lower mental effort when AI helped capture notes, resources, and next steps.

97% fewer errors

Missed notes, lost resources, and incomplete tasks dropped from 65 errors to 2 across all participants.

SUS 90.25 – Excellent usability

 Participants rated the final experience as highly usable and easy to understand.

Research

Through 10 interviews and a competitive analysis, synthesized in Dovetail, three insights shaped the product:

Mentorship fails as a system, not at one point.

Context, notes, action, and motivation break in sequence – each failure feeding the next. This set the scope: a connected loop, not a single feature.

Advice without an order rarely becomes action.

Mentees had plenty of tips but no clear starting point. This shaped the roadmap's logic: tasks sorted by prerequisite, not by date.

Note taking pulled mentees out of their own sessions.

"I'm so busy typing the last sentence that I miss the next one." This shaped the AI Scribe as a background presence – not another tool to manage.

Design Process

In the project the goal was not to add AI features everywhere, but to use them only where users were losing time, focus, or momentum. Used AI tools to pressure test early concepts, with final decisions driven by user research and product judgment.

My process started with one question: where does mentorship break down, and what support does the user need at that exact moment? I mapped the experience across five moments – prepare, match, meet, act, and grow – and used that structure to guide the IA, user flows, wireframes, and prototype, keeping the product focused on the full journey instead of disconnected AI features.

Information Architecture

Mapping the IA showed that the five moments needed to share data one AI engine's output feeds the next which shaped a connected app, not five separate screens

User Flows – new user & task flows

Pinned down each moment where AI acts and where the user decides.

Wireframes – intake, profile, live session, summary

Locked layout and hierarchy before any visual polish.

Final Solution Walkthrough

GrowthFlow supports the full mentorship journey through five connected moments: prepare, match, meet, act, and grow.

1. Smart Intake + AI Profile Summary  

Users upload their resume and answer quick prompts about their goals. AI turns this into a clear profile summary, so mentors can quickly understand the user’s background, strengths, and gaps.

2. Mentor Matching 

GrowthFlow recommends mentors based on the user’s profile, career goal, and skill gaps. This helps users find mentors who are more relevant to their needs.

3. AI Live Session Workspace 

During the mentorship session, the AI scribe captures notes, resources, and action items in the background. This helps users stay present in the conversation instead of focusing on manual note-taking.

4. Post-Session AI Roadmap

After the session, AI turns mentor advice into a prioritized task list. Users can review, edit, and save these tasks to their dashboard.

5. Dashboard + AI Tutor + Goal Rings

Users track their tasks, ask the AI tutor for help, and see progress through visual goal rings. This makes career growth feel more visible and easier to maintain.

Usability Testing & Iteration

I tested the prototype with 10 participants using a within-subjects setup, where each participant completed tasks in both a manual Non-AI flow and an AI-assisted GrowthFlow flow.

The goal was to understand whether AI support helped users:

  • Share their background faster

  • Capture notes and resources with less effort

  • Turn mentor advice into clear tasks

  • Understand and use the dashboard easily

What teasting revealed

The core product logic was clear: users understood the value of AI support for capturing notes, saving resources, and turning advice into tasks. The biggest issue was roadmap clarity – users needed stronger guidance on which task to complete first and why.

Development & Deployment

After validating the designs, I built GrowthFlow as a functional web prototype using React and TypeScript, with Cursor as an AI-assisted IDE to speed up coding, debugging, and iteration. I translated the Figma flows into working interactions, reviewed and refined the code based on design intent, and deployed the final prototype through GitHub.

Build workflow:

  • Designed and refined the core experience in Figma

  • Used Figma Make to explore layout directions faster

  • Built the front-end prototype using React and TypeScript

  • Used Cursor to support coding, debugging, and faster iteration

  • Reviewed and refined the code to match design and functional requirements

  • Deployed the final prototype through GitHub

Key Learnings

1. I learned to design beyond handoff – from intended interaction to working product behavior. GrowthFlow was the first time I took a project past research, flows, and high-fidelity screens into a functional prototype. The design already had interaction logic, but building it forced me to validate that logic in use: what happens after a click, how the system responds, what states are needed, and where users need clearer guidance. It made me a more systems-minded designer.

2. I learned that AI-assisted building still needs design and technical judgment. Figma Make helped me explore UI directions faster, and Cursor supported React and TypeScript implementation. The real learning was turning design intent into working product behavior – reviewing code, debugging interactions, and refining states so the prototype stayed usable.

3. The biggest improvement was not another AI feature, it was clearer UX. In testing, users understood the value of AI support, but they didn't know where to start on the roadmap. Adding sequential task locking gave them a clear order. This taught me that sometimes the best fix is not more intelligence it's clearer guidance.

More Screens

Resume

Interested in Connecting?

Let’s talk opportunities, collaborations, or anything design!

© 2026 Ashlesha. All rights reserved.

CONTACT

EMAIL

LINKEDIN

RESUME