Enter password


Designing Apollo’s First AI Assistant

Leading the vision, interaction systems, and shipped experiences that integrated conversational AI into Apollo’s core workflows.
Before Assistant, AI inside Apollo existed as a collection of disconnected features. AI fields, enrichments, and generation tools existed across the product, but there was no cohesive interaction model tying them together. AI was present, but fragmented.
The goal of Assistant was to introduce a unified conversational layer across Apollo. Rather than requiring users to navigate isolated AI features, the Assistant would centralize reasoning, generation, and execution into a single interface integrated across prospecting, sequencing, workflows, and research.
This required defining Apollo's first conversational AI interaction model. The work extended beyond chat design into interaction systems, workflow integration, approval patterns, generated actions, and AI-native UI components that could evolve alongside rapidly changing model capabilities.
Assistant became the foundation for later explorations into autonomy, delegation, and continuous AI systems. It established the first cohesive AI-native workflow layer inside Apollo and fundamentally shifted how the product approached AI interactions moving forward.

AI existed everywhere, but nowhere cohesively

Before Assistant, AI capabilities existed across Apollo as isolated tools and workflows. Users interacted with generation, enrichment, and automation separately, without a unified system connecting them.
Campaign
Research with AI

Creating Apollo’s first conversational AI system

Assistant introduced a centralized conversational layer that unified reasoning, generation, and execution across Apollo’s workflows.
Engineering demo
Early explorations focused on different treatments for how conversational AI could exist in the UX, and how it could evolve beyond isolated chat interactions.
Explorations

Designing the interaction system

A reusable interaction system was developed to support AI-native experiences across multiple workflows, enabling consistency as Assistant expanded throughout the product.
Design system

The Assistant experience, Alpha to GA

Assistant launched publicly as Apollo’s first conversational AI system, introducing AI-native workflows to users across the platform.
0%
Paid Core WAU
+0%
QoQ growth
$0M
AI-linked ARR
+0%
NRR Lift
Assistant launched publicly as Apollo’s first conversational AI system on March 4, 2026.

Designing for changing model capabilities

As models rapidly evolved, the interaction system required continuous adaptation, balancing usability, capability, and trust while expanding the role of AI throughout the product.
Finding delightful solutions for latency
Constant iteration and testing on the Assistant home experience.
Evolution
Introducing mini-mode for a compact interactive experience