Looppanel

I’ve been seeing tons of Qs on how #AI can be used for qualitative #coding. We’ve run HUNDREDS of experiments on this at Looppanel. We’ve tested, re-tested & re-re-tested, so I thought I’d share the benefits of my learnings with all of you. ?? 1. AI is your research assistant. It can help you code faster, but it cannot replace you. I will say this over and over again, because setting expectations is key. AI is your helper. Don’t expect it to take over your job (thank god!) You should check any AI-generated data and expect it to be a great starting point—not blindly take it as the final answer. ?? 2. If you don’t have very, very good transcripts—you should. Quality of AI transcripts across accents has improved substantially. If you’re spending time checking, correcting transcripts—re-evaluate your tool stack. At Looppanel, we have transcripts with 95%+ accuracy—which means you can have them, too. ?? 3. Replacing a note-taker is now possible. Most researchers WANT a note-taker (because it makes analysis SO much easier), but finding a good note-taker for every call is a challenge. Luckily, note-taking is the kind of task AI is actually really good at. Remove your dependence on other people by adopting an AI-assisted note-taker (again, we already do this at Looppanel (https://bit.ly/4bBE1IO) ) ?? 4. AI-supported theming / analysis We’ve found through deep experimentation that it’s possible to auto-organize your notes by question in your discussion guide. It’s not at 100% accuracy, but let’s say 80-90%—pretty good. We’re currently testing if AI is good at identifying patterns outside of your discussion guide (e.g., identifying that 5 people talked about price being too high). To be honest, the jury's still out on that one—but I will report back with another post once we’ve tested the tech! I’ll keep posting my learnings as we figure out with hands-on testing,, just how good (or bad) AI is at different research tasks. ?? 5. The UX of any AI interaction is actually super important. Whenever we run betas with AI features, we’re partly testing the tech, but often the biggest insights are about UX and content. What tone do users expect? How long or short should a note be? When does it feel overwhelming? How do users discover and explore qualitative data? How do you build trust and traceability into the process? These are just some of the questions we’re constantly grappling with and uncovering via testing. If you want the complete guide on AI + qual coding / tagging, keep reading here: https://bit.ly/3SJbckW If you have specific questions on what AI can do wrt to research, please add them in the comments! I’ll tackle those ones next :) #Looppanel #UXResearch #AIinUX

AI is powerful, when used carefully Thanks for the shoutout, Jessica Higgins! ??

For a UX researcher, the worst is remembering something a user said, but having no idea which one said it. That’s hours and hours of looking for that random quote—across the eons of calls and data you’ve done. Finding the right data is super hard. Tagging data perfectly so it’s searchable is near-impossible. (and also why 80% of repos fail) All I want(ed) for Christmas was Google-like search. So I can find whatever I want in a jiffy. And then the Looppanel team came through with a miracle. They created a smart Google-like search to help you find anything across your projects. Quotes. Notes. Tagged data. Untagged data. Anything under the sun and in the repository. It’s really amazing ???? See it for yourself, just ping me to try brand-new Looppanel search for FREE: https://bit.ly/47QP4eV