Raf Laus.Lead product designer.
Working at the intersection of design systems, research practice, and applied AI for enterprise and infrastructure products. Eight years across healthcare, retail, animal health, and connectivity — most recently the experience architecture for IoT at T-Mobile and the design system + AI tooling for Elanco Knowledge Systems.
T-Mobile · IoT Connectivity.
The framework that drives every enterprise IoT deal — from Einstein/Pardot lead scoring through a nine-section intake to contracting with Legal — redesigned end-to-end out of a Salesforce dual-list-picker hellscape.
I reframed the whole pre-sales arc as one decision system — explicit stages, recoverable branches, visible authorship at every handoff. The nine sections then drive custom and special pricing (including connected-vehicle and SASE bundles) before routing into contracting.
images/iot-connectivity/iot-quote.pngElanco Knowledge Systems.
A two-year program treated as one body of work: discovery, a shipped design system, applied AI tooling, and the operational infrastructure that tied them together. The work that holds my lead-level claim together.
Rather than four isolated efforts, the umbrella case study reads as a single operating model — a system the rest of the org could build on. The design system and ElancoGPT below each stand on their own.
Elanco Design System.
Tokens, type scale, naming, governance — shipped at scale across enterprise UX, AI tooling, and ops surfaces. The chassis everything else ran on.
Three-tier tokens — primitive · semantic · component — public decision logs, and codemod-backed deprecations. Authored as the system the rest of the team builds on, with governance run in the open.
Read the case study →
images/elanco/design-system.jpgElancoGPT · applied AI.
Enterprise AI with a designed epistemic structure — fact, inference, risk, action — over federated LLMs. A decision surface, not a generic chat shell.
Designed for the decision-maker — structured output, recoverable branches, visible authorship after automation fires. Evaluation loops measure whether the recommendation was taken and worked.
Read the case study →
images/elanco/elancogpt.jpg