Enhancing Email Outreach with AI
AI Emailer
RESPOBSIBILITIES
Product design, research
ROLE
Lead designer
YEAR
Apr 2023- Feb 2024
TEAM
Principal product manager, product manager, developers, product designer
Overview
ZoomInfo is a B2B intelligence platform used by over 35,000 companies worldwide, providing data and tools to help sales, marketing, and recruiting teams find and connect with their ideal customers.
As a product designer at ZoomInfo, I worked on creating intuitive, data-driven experiences that empower users to discover relevant companies, streamline outreach, and drive business growth.I led the design of an AI-powered email generation experience that turns B2B data into personalized outreach. My work upgraded the emailer product and helped lay the foundation for ZoomInfo Copilot—our AI solution for sales teams.
The Problem
During the prospecting process, sales reps spend a significant amount of time crafting personalized emails. They often need to jump between different screens in ZoomInfo to gather the most up-to-date and relevant information about a potential lead—such as job changes, company news, or intent signals. We identified this as an opportunity to reduce manual effort and improve efficiency. Our solution focused on leveraging AI to automatically surface key signals about the prospect and present them in a clear, focused interface that streamlines the creation of effective emails directly within the workflow.
Main Challenges
First generative AI feature at ZoomInfo
As the company’s first generative AI initiative, we had to design an experience that felt approachable and intuitive for a diverse user base, including less tech-savvy users. Given how new and unfamiliar this type of functionality was, we treated the product as a learning tool to observe user behavior and gather insights on how to successfully introduce innovation into existing workflows.
Focused launch within ReachOut
The initial rollout was limited to ReachOut, a tool with a narrow use case and minimal screen real estate. This required careful UX planning to deliver both a streamlined editing interface and a clear, actionable output—without overwhelming the user.
Differentiating from generic GPT tools
As part of this project, I analyzed AI-powered writing tools to understand how they integrate into user workflows, focusing on email interfaces. These insights guided our approach to differentiating from generic GPT tools—designing an experience that highlights ZoomInfo’s unique value as a strategic, data-driven resource.
Disconnected systems and fragmented experience
At launch, the AI generation feature was not directly integrated into the existing email-sending service. This created complexity in how users transitioned between tools. One of the challenges was designing a flow that felt seamless and logical, despite the technical separation between drafting and sending emails.
Timeline
APRIL 2023
Internal Release
I started by creating an internal testing experience to refine the prompt and set the roots for developing the full experience
AUG 2023
ReachOut focused experienece
Next, I designed a single-contact solution focusing on sending a personalized email to an individual prospect within ReachOut
FEB 2024
Smart emailing at scale, built into the composer
We enabled selecting multiple contacts to email, and brought the custom AI experience directly into the email-sending flow.
Kickstarting AI Messaging
This was our first experiment integrating LLM technology into the product, aimed at testing a basic flow for AI-generated emails — before incorporating ZoomInfo data.We released it internally to assess the quality of the generated content and observe how users interact with AI writing tools in a sales context.
The experiment validated the need to embed such capabilities in our platform, and more importantly — confirmed the value of enriching them with ZoomInfo data, as planned for future phases.
It also helped us refine the underlying prompt logic to better align with user intent and business goals.
Evolving the Flow – ReachOut Integration
After laying the technical foundation in our initial pilot, we shifted to integrating AI directly into ReachOut — ZoomInfo’s Chrome extension. One of the key challenges was a technical constraint that prevented embedding the generator within the main composer, so I focused on designing a seamless side-panel experience that still felt connected to the user’s writing flow.The goal was to turn ZoomInfo’s signals — like job changes, funding rounds, and intent data — into personalized, context-aware emails. This marked a shift from generic GPT outputs to a smarter, B2B-specific workflow.The flow was divided into two parts: prompt editing and suggestion selection. While the editor went through numerous initial attempts due to evolving product manager requirements, after we reached alignment and validated it with users, the second phase—the suggestions screen—required additional testing and research with ~20 users to refine the final direction.
Option A - Chat interface
Option B - Tabs
Option C - Scroll
76% of the users preferred option B
”Simpler to use, less on the screen, less to read and think about”
These insights informed the design shown below
Last step - Seamless, scaled personalization
A few months after the initial launch, we decided to take the product a step further by enhancing the email experience with two key capabilities: allowing users to select a list of contacts to generate emails for, and green-lighting the development work needed to fully integrate AI into the email-sending system.To support this evolution, we made several refinements to optimize the user experience, including updates to the insights content and layout. We conducted multiple user interviews and design iterations to explore different directions, validate assumptions, and ensure the solution met real user needs. This iterative process helped us arrive at a final design that balanced usability, functionality, and technical feasibility
Validating the concept through user research
Loom videos shared with product leadership, showcasing user-tested design directions that shaped how the AI emailing experience scales across multiple contacts.
The Feedback I Got
Final Design
