AI Signup Flow
As B12’s AI website builder became more powerful, our signup experience lagged behind. It wasn’t capturing enough context to truly personalize the initial “reveal” websites. My goal was to redesign the AI Signup Flow so that DIY leads could easily share more about their business, upload assets, and understand what types of information would help our AI build a better first draft.
Organization:
B12
Role:
UI/UX Designer and Researcher
Project Type:
Onboarding and Form Design
What is B12?
B12 is a centralized, do-it-for-you platform that enables professional services firms to easily establish, run, and grow their business online. Within 30 days, B12 sets business' up with all of the tools needed to reach clients, offer their services, and get paid.
B12 supercharged AI instantly drafts a website, and their team of copywriters, designers, and SEO specialists brings businesses online in 30 days. They deliver a search-optimized online presence with the integrated tools their customers need to serve and charge their clients.
Overview
As part of improving the DIY user journey, I led a redesign of B12’s first signup screen to capture richer business context and assets from new leads. The goal was to guide users toward submitting more meaningful inputs resulting in better AI-generated websites and ultimately, higher conversion rates.
The Challenge
B12's AI websites had grown more powerful and personalized, but the signup flow wasn’t keeping up. Our first screen missed key user inputs like business goals, available assets (e.g., images, files), and stylistic preferences. Meanwhile, vague starter prompts left users confused about what to share.
Before
Key Problems
No Upload Option
Users wanted to share more (e.g., files, images) but had no way to upload them.
Unclear Prompts
Prompts didnt clarify what kind of info improves site results
Information Architecture
The business name field added friction without immediate value.
Result: Short, unclear business descriptions → less personalized reveals → lower DIY purchase rates.
Process
Design exploration
I explored multiple UI concepts:
Design iteration & Implementation
During implementation, we needed to adjust scope based on technical complexity and speed to release. While early designs included visual selectors and upload options for richer context capture, engineering prioritized a faster rollout so we could test and learn from real user behavior.
The team decided to start with “Enhance Prompt,” a lightweight improvement that automatically expands or refines a user’s input to generate a more detailed business description. This allowed us to ship quickly, gather feedback, and evaluate how users interact with the flow before layering in additional functionality.
We plan to revisit visual selectors and asset uploads in future iterations once we validate the impact of the current approach and have more user data to guide implementation.
Outcome
Success Metrics
Our design aimed to drive measurable improvements in early-stage funnel performance:
-
Increased length and depth of business description
-
Higher reveal quality and personalization
-
Lower drop-off after sign up
-
Improved DIY customer purchase rates
Business Impact
By capturing more relevant data upfront, we improved:
Reveal Quality → leading to stronger first impressions
Conversion Rates → Directly boosting DIY customer revenue
MRR Growth → Through higher new and upsell rates
Reflections
This project reinforced the value of designing onboarding as a conversation - not a form. By helping users share what they already know in ways that feel easy and intuitive, we empowered our AI to deliver more value upfront and earn trust early.