AI is the new buzzword to throw into tech conversations. But how can this machine learning help the end user? Recently I spend two days in Dublin at the UXDX annually conference. To keep track on the presentations I decided to test Otter AI, a note taking app that turn voice conversation into notes.
Quick note: I was not paid by Otter AI to do this story. Everything written is my own opinions and tests.
So first things first you have to download the app. But don’t worry, Otter A is available in App Store and Google Play Store. The App uses tooltip onboarding which is a decent method if you only have a few steps.
Once I created my free account I jump right into note taking. A very minimalistic overview that’s easy to understand.
To put the app on the line I wanted to test a couple of different situations at the conference:
- Perfect environment close to an amplified microphone with an English native speaker.
- Noisy environment sitting in the back, far away from an amplified microphone and with an English native speaker.
- Non native environment, same conditions as the perfect environment but with a non native English speaker.
First test was Alan Klement from Revealed. This test actually went quite well. I was following along the transcript and the words appeared quite fast in the app. It felt like a really fast typewriter doing the job for me.
Reading through the text afterwards revealed a few hiccups here and there, but I still understood the overall message. A reoccurring problem was product names. Klement was talking about a camera named Arlo that Otter kept naming Allah.
Overall, the perfect environment was a really good experience with note taking. I felt more in the moment of the presentation and I didn’t have to think about writing everything down.
With the second test I was quite surprised about the outcome. Speaker was Conor Walsh from Duolingo. For this test I sat all the way in the back of the room on an elevated platform close to the ceiling. This resulted in the recording being very bad quality, noisy with a lot of reverb.
The app still managed to do most of the note taking correct but still with a lot of problems when mentioning products and company names.
Reading it back I still managed to understand the general context of the presentation.
You probably already know what’s coming. Unfortunately the app did not perform very well here. I tested it out with Marja Ojala, Lead Designer at Finnair and Sebastien Fabre, Head of UX at Lego. Non of the speakers had a very harsh accent and both very easy to understand. But Otter AI had a really hard time of figuring out what the speakers was saying.
In these two cases i ended up taking handwritten notes on my iPad.
Through the entire conference, Otter had set up a screen where you could follow the text from all speakers. In this case it seemed like the note taking worked in all cases. To be fair, they’ve got a setup that most people haven’t (microphone hooked up the the recording service). So this was more like a WOW-factor and was actually why i downloaded the app in the first place.
So here’s the final judgement. I actually really liked using the app and was surprised at first how well it performed. The UI and UX was very simple and easy to understand, and that made the app so much more enjoyable to use.
Even though i love to work less, I don’t think that Otter AI is quite there yet. If I had to use the app in the future, I would make sure of an environment in the best condition possible, so the note taking would be done correctly. I think everyone would be kind of bummed out, recording an entire interview or speak only to realise that the notes was like reading Esperanto.