# Peter Bartsch, Staff Product Designer

> I make hard systems usable and get people to actually use them.

These are the systems people have to trust: when the data is uncertain, the
organization is fragmented, or adoption isn't optional. I don't stop at the spec. I
prototype in code and ship design systems.

**Available for Staff and Principal roles: remote, SF, or Chicago.**

## Proof

- **34% → 87%**, Profile completion (Deere)
- **$3.8B**, Connected-services business gated (Deere)
- **500K+**, Users reached
- **1→10**, Design org built (FourKites)

## What I Do

Five patterns, one job, each proven on real work:

1. **Making uncertainty understandable**, FourKites
2. **Driving adoption without coercion**, Deere
3. **Unifying fragmented platforms**, Deere
4. **Finding value inside constraints**, Gogo
5. **Building systems that stay aligned**, Thios

These patterns repeat across agriculture, aviation, logistics, mobility, healthcare
& open hardware.

The AI work is concrete: a published Claude Skill
([`audit-design-token-drift`](https://github.com/peterbartsch/AgentDrift/blob/main/skills/audit-design-token-drift/SKILL.md)),
a live MCP server, and [a product whose front door is an agent](https://petebartsch.com/case-studies/this-trip-btw).

## Selected Case Studies

### Deere: Forced Adoption, Without the Force
Inherited a forced-adoption mandate and made completion something users chose, at
500K-user scale, without breaking trust.
- **Challenge:** Most of 500K+ users weren't completing profiles. The flow
  that gates Deere's $3.8B connected-services business.
- **Approach:** Led design of a multi-channel adoption flow across
  8 product lines, reframing a forced-adoption mandate as
  progressive value exchange.
- **Result:** 34% → 87% profile completion · setup support tickets
  −16%
- Full case study: https://petebartsch.com/case-studies/deere-common-ux

### FourKites: Designing for Uncertainty
Built dashboards people staked million-dollar calls on. Retention up
38% YoY.
- **Challenge:** Enterprise customers were making high-stakes decisions on dashboards
  they didn't trust, because we hid uncertainty instead of explaining it.
- **Approach:** Led the redesign introducing confidence levels that showed
  uncertainty instead of hiding it. "Never hide uncertainty."
- **Result:** Customer retention up 38% YoY · repeated account expansion
- Full case study: https://petebartsch.com/case-studies/fourkites-tracking

### Thios: One Source of Truth, Six Surfaces
Built and shipped a product line solo. Six surfaces, one source of truth,
AI-assisted engineering.
- **Challenge:** 6 sub-brands across 4
  production sites plus an Onshape CAD source. Drift is the default, not a possibility.
- **Approach:** Six canonical surfaces (DESIGN.md ↔ tokens.json ↔ main.css ↔
  design-system.html ↔ Onshape ↔ Figma) kept aligned by agent-driven audits.
- **Result:** 6 surfaces in sync · 212 design tokens · 1
  designer
- Full case study: https://petebartsch.com/case-studies/thios

### this trip, btw: Designed for the Agent Era
A trip map an AI can build and a human can open. No account, no trackers, nothing
fabricated.
- **Challenge:** AI assistants are a new front door to travel, but they can't make
  accounts or pay, which locks them out of nearly every travel tool.
- **Approach:** Designed the account out of the product entirely. Trip data rides in
  the URL fragment; an MCP server and Skill let an agent hand over a drawn map.
- **Result:** No-account share loop · privacy as architecture · open beta on real
  trips
- Full case study: https://petebartsch.com/case-studies/this-trip-btw

## Contact

- **Email:** pete@petebartsch.com
- **LinkedIn:** https://linkedin.com/in/peterbartsch
- **Résumé:** https://petebartsch.com/resume
- **Portfolio (PDF):** https://petebartsch.com/Peter-Bartsch-Portfolio.pdf


---

*This is the Markdown representation of this page, served to clients that send
`Accept: text/markdown`. Humans get the same content as HTML at the same URL. The
approach (and why an enterprise portfolio ships an agent-readable surface at all)
is written up in the case study "this trip, btw: Designed for the Agent Era":
https://petebartsch.com/case-studies/this-trip-btw*