From draft to sent in the user's own words.
HeyFollowUp
Most networking tools solve the handshake. Nobody solved what happens three days later.
People leave events with new contacts and good intentions. The follow-up window closes fast and once it does, it rarely reopens.
I researched why users weren't sticking around and redesigned the full experience:
Capturing a connection before context fades
Organizing contacts by priority
Sending a follow-up with AI that drafts in your own voice
Role
UX Designer
Duration
9 Weeks
Team
Collaborated on stakeholder interviews; research synthesis, design, and prototyping led independently
Tools
Figma, Miro, Google Docs
Problem and Opportunity
The moment nobody designs for You leave an event with fifteen new contacts saved in your phone. Three days later you haven't messaged any of them. Not because you forgot. Not because you don't care. Because the blank message box is impossible to fill when you can't remember why the conversation mattered.
Every tool solved the wrong problem Popl, Blinq, Clay and Covve, all of them focused on the handshake. Getting information from one person to another as fast as possible. None of them supported what happens after that moment.
The gap Not better contact capture. A system that meets users a few days after the event, when momentum is fading and the connection is about to go cold, and makes the next step feel possible.
The design challenge Reduce the distance between meeting someone and sending a message that feels worth receiving.
Research and Discovery
What already existed Some stopped at contact exchange. Clay and Covve went further but both skewed toward power users and CRM workflows, not someone trying to send one good message to someone they met last Tuesday.
What that revealed The gap wasn't a missing feature. It was a missing philosophy. Collecting contacts at an event is easy. Every tool stops there or sorts them for sales pipelines and marketing goals. Nobody was helping you figure out who actually mattered and what to say to them.
Key User Insights
Three personas, two breakdown moments Milton, Hannah and Mindy had different experience levels and different networking styles. But the research pointed to the same two failure points regardless.
Post-event chaos No system to capture context before it fades.
Follow-up fatigue The blank message box winning every time.
Those two moments became the design targets and shaped everything that followed.
“I don't think I have a pattern... it's a very disorganized way of reaching out to people.”
- Milton
Milton isn’t new to networking. The follow-up is where his confidence disappears.
Design Process
The obvious answer The first wireframes were built around a contact list: organized, searchable, filterable by relationship type and date. Working through the flows made clear it was wrong.
The real problem was not the list but the distance. To follow up with someone a user had to open the app, navigate to contacts, find the right person, remember why they'd met, and then start writing.
Five steps before they'd typed a word. That's where people were giving up.
Dashboard Iterations
The pivot Dashboard-first instead of contact-first. Rather than asking users to manage a list, the redesign surfaces the right person at the right time based on relationship status and how long ago they'd been captured. The user opens the app and their next action is already visible.
The filter for every decision Does this reduce the steps between a user and their next follow-up?
That question drove three specific changes:
Quick actions above the relationship feed
AI draft accessible directly from the dashboard reminder, no navigation required
AI suggestion surfaced during capture, not just at follow-up
Design Decisions
The redesign targets three moments where existing tools had failed:
Capture — selfie, QR, or business card scan while the face and conversation are still memorable
Organize — tags and context feed the AI so reminders are actually useful
Follow-up — AI drafted, but in your own voice
Each screen exists to reduce friction at exactly the point where people were giving up.
From “who did you meet?” to sent in four screens.
Three ways to capture, one place everything lands.
Most people lose a contact not because they forget but because stopping to record them mid-conversation is awkward. HeyFollowUp removes that moment entirely, offering three capture modes depending on what the situation allows:
Selfie together — a face to match the name
QR scan — instant contact exchange
Business card scan — for when that's all you get
Each takes under thirty seconds. Everything lands in the same place with notes, context tags and a follow-up reminder already set.
The right person, at the right moment, with a message that sounds like you.
The hardest part of following up isn't wanting to, it's starting. HeyFollowUp surfaces the right contact at the right time, then drafts a message from the user's own notes. The AI fills in structure. The voice stays theirs.
Validation and Reflection
Testing 7 users across the two highest-risk moments, post-event organization and AI-assisted follow-up.
What worked
AI drafts feel natural when built from the user's own notes
Email and SMS integration reduced app-switching friction
Reminders, notes and tags kept connections from going cold
The finding that surprised me Users were more comfortable with AI than expected as long as the output sounded like them. The hesitation wasn't about using AI. It was about losing their voice.
If I were to continue The original product positioned itself as a networking buddy not just a tool. I'd want to build that out through adaptive guidance that meets users where they are. Milton needs confidence scaffolding and habit building. Mindy needs efficiency. The product should infer which mode fits rather than asking users to self-identify before they've seen enough to know.