Jacquard

Prompt Engineering.
Language Tool

Employer

Jacquard

Senior Product Designer

Role

As part of a strategic initiative to enhance personalisation at scale, I led the design of the Prompt Engineering Language Tool, aimed at empowering the Language Team to generate, manage, and approve high-quality personalised language with minimal manual input.

This tool was critical to supporting Jacquard’s Northstar goal: to build scalable infrastructure towards a fully automated service, while delivering actionable, user-guided interfaces for internal teams.

Collaboration

Partnered with engineers to ensure technical feasibility and seamless implementation.

  • Cross-functional: Worked closely with Product Owners, Engineers, and Language Specialists across multiple epics to ensure a cohesive experience.

  • Leadership: Partnered with senior stakeholders (PMs, and Tech Leads) to define requirements, refine success metrics, and maintain alignment on product goals.

  • Technical Partnerships: Engaged engineering teams early to ensure technical feasibility and scalability, particularly around dynamic prompt generation and approval flows.

Results

Successfully launched the internal Language Tool within the project timelines of 6 months.

  • Delivered a scalable design framework that could support 20+ customers using personalisation, 60+ ongoing campaigns, 6,000,000+ variants reviewed.

  • Higher open and click rates for personalised variants vs. non-personalised ones.

  • A Prompts & Variables Management System for scalable language generation.

  • A Language Approval Tool for fast, streamlined variant validation, 75% of recommended combinations were deemed relevant by users.

  • 25% of prompt templates required no re-editing after generation, showing high output quality.

Key challenges included

  • Designing a system that could dynamically generate and approve thousands of language variants at scale.

  • Creating intuitive interfaces for complex tasks such as prompt/variable management without overwhelming users.

  • Managing dependencies across four major epics with interlinked functionalities while keeping timelines on track.

  • Ensuring a balance between human creativity and machine-driven personalisation logic.

Research and Discovery

  • Conducted stakeholder interviews to gather detailed requirements for each epic.

  • Analysed workflows from manual language generation to identify key pain points and automation opportunities.

  • Benchmarked internal processes against industry best practices in personalisation and language tooling.

Design and Iteration

  • Designed wireframes using existing components and high-fidelity prototypes in Figma, aligned to the scalability goals outlined in the project Northstar.

  • Created user flows for major functionalities including recommendation engines, prompt/variable management, generation interfaces, and approval workflows.

  • Collaborated with PMs and engineers through regular design critiques, sprint reviews, and technical refinement sessions.

Validation and Testing

  • Conducted usability testing sessions to validate assumptions around personalised language creation.

  • Conducted feedback review sessions using example POC’s from customers like Adidas to validate our approach with real projects.

  • Iterated designs based on feedback from internal users to optimise for speed, clarity, and scalability.

This 6 month project demonstrated how effective cross-functional collaboration and fast, iterative design can deliver high-impact internal tools within tight timelines. By deeply understanding user needs and prioritising scalability from the start, we built a system that balanced automation with human oversight. The project strengthened our internal design processes for tool-building, and highlighted how even complex, technical workflows can be made intuitive and user-friendly when approached with a clear, structured UX strategy.