|
Michelle Jackson-Blackwell
|
Certified UX Writer, Certified Usability Architect, Certified Business Analyst
|
|
I led AI content strategy for a personalized experience
|
|
|
|
The challenge
AI content at scale. Alignment around the leadership team's vision for "Experience Collections" -- an AI content-led, personalized shopping experience for interests, hobbies, and goals.
Executing on the company's bet on AI-powered content for a discovery-based shopping app built for exploration; encouraging shoppers to explore related items that were not on their radar.
How do we make meaningful collections, with human-centered AI-powered content at scale, that engages, inspires, and converts?
|
|
|
|
The vision
Short-term goal. Experiment with 300 content-led experience collections centered around specific interests. From search to feeds to videos, these modules help inspire shoppers to discover more.
Long term goal. Scalable AI content frameworks. All content-led, interest-based collections throughout the app, with AI-powered content, influencer content, videos to trigger inspiration, user-generated content (reviews), & collections of items needed for interests, hobbies, or goals.
|
|
Personalized AI-powered content
|
AI/LLMs: This was the companies first (ever) project using LLMs (ChatGPT and Claude) to generate AI content for the app -- requiring prompt engineering and AI content frameworks.
Experience Collections: The Wish app was built for discovery and exploration of unique, niche items and affordable shopping made fun. The "Experience Collections" go beyond just showing a product feed of WHAT to buy to providing context and inspiration for WHY to buy, connected to a specific interest, hobby, or goal. Displaying the experience collections is personalized per shopper based on data analytics.
|
|
|
|
|
|
My role: Content Design Lead
What I did
Cross-functional collaboration: I joined a team to explore integrating AI content into the user experience. I worked closely with executives, product managers, & product design. I communicated the project goals to the content design team. My team worked together to craft 300 collections for the company's bet on AI content-led, interest-based experience collections.
Led content design & AI content strategy: I led a team of UX writers on a project to integrate AI content into the app experience to promote personalized discovery and exploration. Leading a team on the first (ever) AI project meant learning (and teaching) prompt/context engineering, AI editorial oversight, documenting guidelines, & frameworks.
I put together an alignment guide for the team with AI content strategy and framework -- including a prompt library, custom AI prompts, legal constraints, content standards, style guidelines, best practices for prompt writing & working with LLMs, and more.
Prompt engineering: I wrote and refined AI prompts, evaluated AI output, set standards, and communicated how to execute to the team. We used the prompts to generate content at scale for 300 experience collections, each with a specific interest, for an experiment. I used AI to research interests to see what was meaningful, refined prompts to improve response quality, and edited AI content.
Systems & strategic thinking: I proactively worked with the product designer to come up with future-state solutions for connecting AI-powered content across flows, features, platforms, and experiences to create maximum value for decision making and leverage AI content at scale. We presented our ideas to leadership and influenced product & design strategy.
UX writing: We wrote microcopy for the entry point tiles/modules (placed throughout the app) and the headers on each of the curated sections within each of the collections.
|
|
|
|
|
|
|
|
|
Prompt engineering. I crafted and refined AI prompts, managed AI/LLM editorial oversight, and documented guardrails, guidelines, & frameworks. I wrote custom AI prompts with content guidelines to generate AI content within brand voice, refined prompts to improve response quality, and created a scalable content framework.
Context engineering. I put together an AI content strategy with an alignment guide for the team -- including best practices for prompt writing, content standards, style guidelines, legal constraints, and more. I evaluated AI output against content guidelines, guardrails, and brand identity -- editing content for style, repetition, hallucinations, and anything inappropriate.
The results. Meaningful AI-powered, human-centered content that positively impacted engagement & conversion. Plus, AI frameworks and content governance for creating content at scale to test leadership's vision for an AI-assisted, personalized shopping experience for interests, hobbies, and goals.
|
|
|
|
|
|
Collaboration with product, design, & engineering. I collaborated with the product designer and engineers on feasible options for a solution aligned with the product vision. I worked with the Product Designer to define a meaningful MVP that helps users explore possibilities and ideas. We provided design options with flexible components.
Together with the content design team, we created 300 collections for an experiment. Each collection has an interest tile/module pointing to a landing page with AI-powered content introducing the intent (interest per collection). The tiles/modules are strategically placed throughout the app (product feed, search pages, Wish Clips videos, Wish Assistant chatbot, etc.) to inspire customers to discover products, goal-related topics, hobbies, or new ideas.
|
|
|
|
Cross-functional collaboration
Teamwork. I set up a collaborative process for the content design, legal, globalization, & merchandising teams to work together to map, review, critique, & approve our microcopy.
We wrote microcopy for 300 entry point tiles written to inspire shoppers to discover items related to interests, hobbies, or ideas to explore. Which tiles to display throughout the app is personalized per shopper based on data analytics.
|
|
Localization
Translation for globalization. Wish is available in 62 countries and 34 languages. I worked with the Globalization team to ensure context is maintained for the global audience and text expansion doesn't break the design.
|
|
Systems thinking for UX content
How might we? Connect AI-powered content for experience collections across flows, features, platforms, and experiences to create maximum value for decision making
|
|
Partnering with UX research
User testing. We used quantitative and qualitative data from user testing and experiments to iterate on the designs. Prototypes were shown to 10 test participants who bought on Wish and provided feedback.
|
|
|
|
|
|
Outcome for MVP experiment
AI-powered content at scale.
Measurable business outcomes for going beyond showing shoppers WHAT to buy, to providing context for WHY to buy to complete a goal (Jobs-to-be-Done) related to an interest.
6.5% lift in product boost
~ 5% lift in GMV / AOV
AI frameworks with context engineering
Prompt library
|
|
What I learned
Human-centered design matters.
AI content MUST be edited for context, accuracy, and alignment with strategy and brand identity. Knowing your personas, target audience, and brand voice remain important, as well as advocating for a content design point of view.
You'll also need to help cross-functional teams understand UX writing fundamentals like clear, concise content, character limits, reusable patterns, chunking for scanning.
|
© 2026 jacksonblackwell.com
|
