Your discovery research. Accurately themed. In minutes, not hours. What would that mean to you?
A look at Discy's new Smart Tag feature
Is this scene familiar?
You’ve done your interviews.
You’ve got your notes / recordings.
You’ve a decent set of survey responses. Not quite the thousands you expected... but still a lot to go through.
There’s also the 15 relevant research articles you found in your firm’s knowledge drive on Google.
It’s a lot.
5-10k lines worth of text... and you’ve got deadlines.
So while you know the headlines, you can’t just rely on the headlines. You need the detail..
Which means you (and team) reach the point where you need to assemble the data in one place. Then go through it. Picking out relevant text and arranging them in themes.
“This won’t take long”.
But it does.
Days in fact.
You’ve no guidebook for your qualitative theming. And after a few hours, you are already going word blind.
No matter how interesting the topic, moving text and images around is dull and cumbersome.
It’s also textbook conditions for confirmation bias!
The unspoken flip side of us needing to be experts... is the ironic limit it puts on our growth mindset. Our 'beginner' mindset.
So then, what’s a smart tag?
We want to help fix this pain with our new thing.
It’s a thing that enables you to become the robocop of theming. (Or tagging as we call it in Discy).
You capture your interview notes in Discy (or upload a survey/ recording/ project file etc). Then the AI trawls through and marks the most salient sentences and sentiment for your project’s tags.
What makes it smart(er) is the use of AI trained on consulting-wide concepts that apply across firms. And/ or vocabulary and frameworks specific to your firm. So now you have your codebook for qualitative theming.
More efficient and less wizardry needed than ‘zero-shot’, ‘few-shot’ prompt engineering.
Picking up language and themes in your project research, that generic AI would miss. Let alone tired human minds reviewing thousands of words!
You just pick the combination that makes sense for your project.
And whichever combination you choose, at the end of a few minutes of trawling you get this….
Then you say whether you agree with the tag suggestions or not.
So you can spend more time on what matters. Digging into what the data tells you, and the depth / relevance of those insights.
Cut the faff.
Limit the human bias.
Keep your creativity levels up.
And, build themes that are precise and data-led, rather than high-level generic.
All aboard for Deeper insights.
Wonderful you say, but what’s the impact?
Over the past month, Smart Tag has been put through its paces on a few projects.
First examples include...
In short, it’s promising.
Deeper project insights is our north star.
And enabling you to confidently theme your data at speed is a big old building block for our big dream.
A key waystation on the path to making data science more accessible.
It’s what we get out of bed for. It’s our (insert your preferred big aim cliche here).
It’s why bit by bit we are going after every part of the manual, painful, bias prone qualitative analysis process.
So yeah, we think it’s cool.
We’re pretty sure you’ll think it is too.
At the very least, you'll appreciate spending less days doing this manually...