We caught up with Emily Blumenthal, Research Director at Yabble to get her take on why for both quantitative and qualitative projects you’re missing a trick if you’re not using Yabble Gen.
Q: What makes Gen different from the average chat bot?
A: I don't even put Gen in the same class as the average chatbot. Gen's been trained by some of the best researchers in the industry. Gen’s more like an advanced research assistant/research executive. It’s had a lot more specific training and learning to deliver answers that are considered and accurate.
Q: How do you see Gen changing the way that quantitative and qualitative researchers work?
A: I think Gen's going to have a big impact on the research process and I think it's really positive. I think the first area is around the way that we ask questions. Obviously as researchers, that's key - being able to ask a good question! But I think with the use of AI tools like Gen, it's going to be taken to the next level, because essentially the better your questions, the more you probe, the richer and the better quality your insights. The other area that's going to be interesting is AI tools like Gen are going to democratize the research process. The ability to run and analyze research has become easier with Gen. I think overtime, we are going to see a lot of other roles and functions being enabled to do their own research and a lot faster as well. Tools like Gen can do a lot of the heavy lifting that traditional researchers may have done in the past.
Another area obviously is speed. Gen is super-fast. Yabble’s AI tools on average are thousand times faster than a qualified human. A recent example of that – I had a client who was on deadline to deliver a voice of customer report for that day, she hadn't even started it – you know the scenario, a client that is increasingly under the pump. With Gen she was able to quickly start asking questions of her data and was able to have a written report in about an hour. Her quote: ‘Gen saved me that day’! So, she's a real advocate. We're going to see the speed at which insights can be created revolutionized with tools like Gen. That's going to really speed up the whole process.
Q: What kinds of projects does Gen work best with?
A: Gen really comes into its own on projects where you've got a lot of verbatim comments. For example, for quantitative data, voice of the customer programs, brand trackers, review data, call centre transcripts, social media data. And then for qualitative, you've got interview transcripts, focus group transcripts, e-qual boards or even podcasts – for those of you who aren't familiar with what e-qual boards are, they are sort of like a private social media group. You've got 20 to 30 respondents posting videos of themselves talking about different questions, writing comments and answering each other's questions. So, they're quite complex and they're incredibly rich as far as the data collected. For projects with high volume open ends, comments or conversations Gen is a superpower for sure.
Q: How do I know my data is secure when using Gen?
A: Gen sits within the Yabble platform. An AI platform that is designed specifically for insights. Yabble knows how important data security and privacy is in the insights industry and you can think of the platform like a walled garden. That means any data you upload or connect into the Yabble platform for analysis stays 100% private to you, it does not get used to train any large language models like ChatGPT.
Q: When using Gen, what are some key things to think about to maximise and optimise your outputs?
A: I think it's important just to think about what you're trying to get out of the project that you're working on with Gen. What are the kinds of key objectives that you want to cover. Because it's all really in the way that you ask those questions. It's also having a bit of a framework, so what are the questions that I want to ask? And then, as you go through and those responses come up, probing Gen a little further. So, you know, you talked about this theme can you tell me a bit more about that? And, really challenging Gen to show you, where did Gen come up with that information from. So, you just go deeper and deeper and end up getting rich insights out of your data.
Q: Where have you seen Gen come into its own use case wise?
A: So, the most recent example came this week – Gen is now able to interrogate qualitative transcripts. We had an e-qual board for a retail customer - theirs was a very comprehensive board, it ran for five days, with 30 respondents - so huge amounts of data that we needed to sift through. We simply ran some questions over Gen and the quality of the answers was incredible, and the researchers saved many, many hours of manual analysis through leveraging Gen.
Q: How can researchers trust what Gen says?
A: Gen has been built specifically for insights and works in a deterministic way. That means Gen only creates insights from the data in your project and doesn’t hallucinate. We’ve also designed Gen so you can see which comments or interviews Gen has used to create the insight so you can have confidence in the output.
Overall, though, I think with any new technology in the data space, you've got to do your own testing. So, I'd encourage any researcher, maybe pick a past project that you've done your manual analysis on, and then, follow a similar process with Gen and just have a look at what the different quality of the outputs are and decide for yourself. But certainly, the feedback that we're getting, is, Gen is doing an incredible job.