Andrea Salad
Loom (feat. Bonobo) — Ólafur Arnalds · some kind of peace
Process

How I Work: AI-Native Product Strategy & Design

AI has been a force multiplier for my role, and I was an enthusiastic early adopter. It lets me spend more time crafting inspiring visions, validating directions, influencing roadmaps and solving the problems that create the most customer and business value.

I love leveling up alongside my colleagues, and crafting reusable templates, systems and AI skills that make life on the job a little easier.

Favorite tools: Claude Code, Co-Work and Design · ChatGPT for specific tasks · Outset AI · Google’s NotebookLM · Lovable · Firefly · Whispr

Here’s my process in a nutshell:

01

Discover

AI agents surface past research, customer feedback, analytics, support conversations, experiment results and prior design decisions, so the team builds on what we already know.

I pair that with external research in Claude, ChatGPT, Perplexity and custom MCP workflows, plus a custom Claude Landscape Analysis skill connected to Mobbin that turns competitive reports from days into minutes.

I synthesize signals to find where customer needs, business goals and market opportunity meet, and I bring partners along from day one.

02

Define

Prototyping in Claude, Replit, Lovable and Figma Make lets me explore many divergent concepts and test everything meaningful before converging, so I can recommend strategic directions with less risk.

Outset AI is my favorite way to scale research. Alongside interviews, analytics, JTBD mapping, opportunity/solution trees and hypothesis-driven experiments, AI synthesis surfaces patterns across hundreds of conversations without losing nuance.

03

Design

I use AI as a design partner to explore interaction models, simulate edge cases, check accessibility, generate realistic content and test information architecture before investing in build.

With Claude Design, Replit, Lovable, Cursor and other AI coding tools I build prototypes on real APIs and data, using our design system so they already match engineering architecture, tokens and accessibility standards.

Because the work is grounded in production reality, the conversation shifts from “Can we build this?” to “Should we build this?”

04

Deliver

By delivery, most of the risk is gone: the work is grounded in customer evidence, aligned with strategy, validated through testing and built against production systems.

Depending on the team’s needs I deliver executive-ready strategy, vision work with sizzle reels and narratives, realistic prototypes, service blueprints, OOUX diagrams, visual roadmaps, Figma specs and a little production-ready front-end code.