Summarizing user research interview transcripts
Tools: Microsoft Copilot, Claude
I have used Microsoft Copilot and Claude to help analyze usability interview transcripts. The AI anonymized transcripts (using “Interviewer” and “Participant” instead of names) and summarized each question and response while preserving the interview structure. This removed filler words, repetition, and small talk, making the transcripts significantly faster to review.
I then fed the cleaned transcripts back into the AI and asked synthesis questions, such as “What did the participant say about the images on the website?” These prompts allowed the AI to pull insights across multiple responses. I also requested direct quotes for reporting, which helped speed up stakeholder-ready summaries.
Exploring new areas of expertise
Tools: Claude, ChatGPT
When starting work in an unfamiliar domain, I use AI tools to quickly build foundational understanding. For example, I asked Claude to explain SEO considerations for LLM-based search, including concepts like schema, GEO, and how these overlap with traditional SEO practices.
I’ve also used AI to learn the basics of deal registration and bidding in sales contexts. When possible, I ask the tools to cite sources so I can validate information against trusted references, helping me get up to speed without hours of unstructured searching.
Creating draft research plans, surveys, and interview scripts
Tools: Claude
Claude can generate an initial research plan based on unstructured notes, providing a useful scaffold when no template exists. I’ve also used it to expand draft interview questions into full scripts, including introductions, instructions, and contextual explanations for participants.
This approach speeds up early drafting and allows me to focus my time on refinement.
Finding websites associated with properties owned by a management company
Tools: Microsoft Copilot
To identify stand-alone websites for properties owned by a specific property management company, I used Copilot to generate lists of potential URLs. While the results were not always accurate, the process was faster than manual searching.
Copilot provided a list of clickable links that I could quickly validate, reducing overall time spent gathering the information.
Analyzing the audience for a website
Tools: ChatGPT
I provided ChatGPT with websites from different property management companies and asked it to assess who the homepage content appeared to be targeting. The tool was able to accurately summarize likely audiences and identify whether shopper-focused content was present.
Importantly, it explained which elements it used for its assessment, allowing me to evaluate whether those elements really supported shopper needs or were being interpreted incorrectly by the AI (e.g., property listings that were not actually designed for shoppers).
Generating realistic dummy content for mockups and wireframes
Tools: Claude
While designing a marketing dashboard, I needed realistic placeholder content related to joint business plans and marketing development funds. Claude generated a sample report that reflected the type of data the dashboard would contain.
This saved time on research and brainstorming and allowed me to focus on design using context-appropriate content.