Over the past few weeks, we published a series of blogs about how Generative AI can positively impact qualitative data analysis. We spoke about
In a nutshell, the aim was to discover, together, how generative AI is shaping new age QDA. But what of the impact? On the whole, we should be very postive about these rapic developments. I am very convinced that it is a promising tool that promises to take away the tedium, the onerous tasks at the very least. And hence help us focus on second and even third level QDA.
However, we must remember that many of the tools out there are in ‘Beta’ or limited release stage. Consequently, the output often feels incomplete, often tangential to what you want, and requires preparing the input material so that the tool can work with it.
The widely adopted tools are offering typically summarization options to users using AI (and there is a Pandora’s Box here that I don’t want to open up, yet). The output is mixed bag, at least in my testing.
At this point, I am sure you are thinking, so what? ????
What we have shared so far is a humble attempt to make current information, mostly what we have researched first hand, available to all the readers of this blog. However, the proof of the pudding is in the eating. If you want to really explore all of this yourself (and I think you should), then here are some steps that can get you started off.
Easy, right? Now imagine doing this at scale for a single project! This can be of such great help for you in eliciting insights from the transcripts/project. Think about a prompt for ‘finding the 3 most mentioned positive attributes of Brand A, as mentioned in the following transcript’.
I believe that while the experiment will give you first-hand knowledge of the advantage you can gain, you are going to be challenged when doing this at scale.
At flowres.io, we decided to speak with qualitative researchers about their expectations, anticipated challenges, and experience with existing tools, specifically those that use Generative AI. Here are some preliminary insights that came out from our conversations so far.
And our all-time favorite - it better be easy to use!
This is where we realized that we need a capable solution (or partner even) - keeps you in charge while supporting you and sharing the workload intelligently. And guess what, we had one right in our platform, and did not realize it.
Our transcription and content analysis practice leverages an application that we have built in-house, for tagging and data sorting. It is a focused type of CAQDAS, which slots into our workflow from getting the input media to generating the analysis. And has a plug-in that will directly allow you to converse with the data.
We realized that this application potentially can deliver immediate and powerful benefits to researchers and analysts. And guess what, we did! We showed it around, got great feedback, encouraging adoption by researchers, and so we decided to release it to a wider audience.
Obviously, it addresses some of the points raised by researchers we spoke to. Not all of them, mind you, but certainly the ones that make a viable application for a researcher to use.
With their own transcripts … and their own analysis headers!
If you want to try it out, you can head here and do so
So just like our QDA tool, there are many more tools out there. Apart from the tagging of data, they also provide significant amount of tools for visualization, and for the sorting process. And have varying degrees of adoption across user segments.
Here is where there seems to be an overwhelming need for a co-pilot emerging. That is, someone who understands the fast evolving AI space, can connect the dots from an applicability standpoint, is able to generate and analyze output, and can also understand the risks involved in using any or all of all the tech out there from a data security and privacy standpoint.
Wouldn’t it be ideal to work with a partner who can be a capable co-pilot? Reach out if you want to talk.