Velocity Engine 2025–Present
Velocity Engine Redesign
Overview
A powerful AI backend, ready for an interface to match.
Velocity Engine had built a powerful AI backend that helped marketing teams generate campaigns and content at scale. The interface had an opportunity to catch up: onboarding was steep, navigation was clunky, and users struggled to access the platform’s full capabilities.
We streamlined the core experience to make it easier to learn, easier to navigate, and easier to love.
| My Role | Fractional Head of Design |
| Team | 2 UI designers (Launch Design), CPO, Head of Growth, PM, SMEs (Velocity Engine) |
| Model | Embedded virtually via Slack · Weekly strategy & design sprints · Continuous SME alignment |
| Duration | March 2025 – Present |
The Problem
Users had access to powerful AI, but no guided path to put it to work.
New users hit a wall during onboarding. There was no opinionated setup flow to prompt them for the right information in the right sequence, so they were left guessing where to start. The sales team struggled to demo the product live, because the interface made simple tasks look complicated.
Business Goals
- Align the campaign creation workflow with how marketing professionals actually think, restructuring the UI to match the mental models uncovered through SME interviews.
- Give users a focused workspace that surfaces only what’s relevant to their current task, minimizing cognitive load.
- Elevate AI assistance from a background capability to the central interaction model, matching how users told us they want to work.
I kept asking myself, does this match what our marketers actually do? Does this flow feel like how they think, or how we think?
Strategy
Key Decisions
Figma → AI-Generated React Prototypes
By building prototypes in the same stack as production (React, TypeScript, Tailwind), we eliminated the fidelity gap between design and engineering. The higher-fidelity output also made SME feedback sessions dramatically more productive. Experts could react to something that felt real, not conceptual.
A functional prototype built in the production stack, used to validate interactions with Subject Matter Experts.
The step-by-step wizard encodes SME best practices directly into the campaign creation flow.
Opinionated Wizard-Based Workflow
SME interviews revealed that marketers wanted guided structure, not a blank canvas. We built a step-by-step wizard that encoded expert best practices into the flow, which increased beta customer satisfaction.
Alignment Artifacts
The research and frameworks that built shared understanding.
| Artifact | Purpose |
|---|---|
| Journey Maps | Mapped the end-to-end campaign creation workflow to identify friction points |
| SME Interview Synthesis | Distilled insights from marketing and writing experts using AI-assisted synthesis to align UI with user expectations |
| Design System | Established a unified component library, initially in Figma and later directly in React/Tailwind code |
| AI-Assisted React Prototype | Production-fidelity prototype built in React/TypeScript/Tailwind for realistic SME validation and faster engineering handoff |
Outcome
What changed.
- Customer churn decreased after the redesign launched
- Time-to-first-campaign improved significantly with the new wizard workflow
- Sales team successfully generated leads at the HubSpot conference. The streamlined UI made live demos far more effective
Key performance improvements following the redesign launch.
The Solution
An opinionated workflow that guides, not gates.
The campaign wizard replaced a sprawling, unstructured setup flow with an opinionated
step-by-step experience. Each step encodes best practices from SME interviews, so marketers get expert-level guidance baked into the interface, while still leveraging the intelligence of AI through a familiar chat interface.
The wizard-based workflow guides marketers through campaign creation step by step, encoding SME best practices directly into the flow.
A dashboard built around daily workflows.
The dashboard was rebuilt around the tasks that actually drive daily use, surfacing active campaigns, recent content, and quick-start actions above the fold so users reach the platform’s AI capabilities without navigating through layers of menus.
The redesigned dashboard prioritizes the most common workflows, reducing time-to-value for new users and removing friction for power users.
Learnings
What we took away.
Product Design
AI-powered products evolve at the pace of the underlying foundation models. Design teams must maintain continuous dialogue with SMEs and design partners to understand how each capability leap reshapes existing workflows, or creates entirely new ones.
Collaboration & Process
Moving from static Figma mockups to AI-generated React prototypes (using the client’s actual stack: TypeScript/Tailwind) eliminates the fidelity gap and enables realistic SME validation of complex logic, not just visual patterns.
Business
Aligning UI messaging with SME mental models is a growth lever, not just a design task. It directly impacts “time-to-aha” and long-term user retention.