AI-Driven Product Designer

Design with AI in the loop

I’m a Senior Product Designer who builds AI into how products actually get made — custom agents, AI-assisted research, automated usability scoring — applied to fintech and enterprise systems where a wrong decision costs real money.

The result isn’t faster slide decks. It’s a 321-screen credit platform shipped at 86% average usability score, UX writing reviews cut by 67%, and research synthesis that took days now landing in hours.

About Me

Marcelo Pampanini on a train through the Swiss Alps, looking out at the mountains
  • AI-assisted discovery and synthesis
  • Custom agents and knowledge bases
  • Automated usability metrics
  • Fintech and financial platforms
  • Enterprise and B2B systems
  • Data-heavy, high-stakes workflows
  • Design systems at scale

Hi

I am

Marcelo Pampanini

9 years designing complex products — fintech, enterprise, B2B. For the last 3, AI hasn’t been a demo in my work. It’s tooling I build, configure, and hand to the teams around me.

At Zup and Itaú BBA that meant custom agents on StackSpot AI grounded in real domain context, a specialized knowledge base for the design team, and a UX Writing command that cut daily review time from 3 hours to 1. On a 321-screen rural credit platform, I used AI to cluster interview findings and built an agent that standardizes usability scoring across 9 squads.

AI compounds where work is repetitive and verifiable — synthesis, scoring, documentation, consistency. It doesn’t decide what to build. That still takes evidence, business context, and the judgment to catch a model being confidently wrong.

Work Experience

  • Senior Product Designer @Zup Innovation / Itaú May 2021 › Present
  • Senior Product Designer @Facily Jan 2022 › Mar 2023
  • UI/UX Designer @Venturus / Apple May 2020 › Dec 2021
  • UI/UX Designer @Assertiva Feb 2019 › Mar 2020

Education

  • Digital Product Leadership @TERA 2023
  • Product Discovery @PM3 School 2021
  • UX Strategy @MIAMI AD School 2020
  • Bachelor's in Advertising @PUC Campinas 2010 › 2014

Tools and Skills

Figma
Maze
Notion
Figjam
Adobe Softwares
ChatGPT
Claude
AI Agents
Figma MCP
Product Discovery Research and Prototyping Problem Solving Cross-functional Collaboration Ownership Prompt Engineering AI-assisted Research
0 Years as
Product Designer
0 Years as
Visual Designer

What I love to do

01

AI-Assisted Product Design

End-to-end design where AI carries the repetitive load — clustering research, drafting microcopy variants, checking consistency across hundreds of screens — so the thinking time goes to structure, risk and trade-offs.

02

Custom Agents and Knowledge Bases

I build agents grounded in a specific domain, not generic prompts: an expert agent for a business vertical, a knowledge base with traceability and access control, a tone-of-voice command that matches a client’s guidelines. Tools a team actually keeps using.

03

Research and Validation at Speed

Interviews turned into actionable themes in hours instead of days. Usability scoring standardized by an agent so results stay comparable between versions and squads. Faster loops, same rigor — the evidence still comes from real users.

04

Bringing Teams Along

Adoption is the hard part. I run workshops, set up shared agents, and show the work in practice — because a tool nobody trusts delivers nothing, however good the demo was.

Mentoring designers, contributing to design systems, and raising organizational design maturity.

Contact

Looking for a designer who already works this way? Tell me what you’re building.

Please enter your name.

Please enter a valid email address.

Please write a short message.