Problem

Students often have to spend a lot of time and effort manually taking lecture notes, which distracts them from absorbing the information.

The client decided to test the hypothesis that a speech recognition service could help automate note-taking and make life easier for students.

Solution

To test the idea, we created a minimum viable product (MVP). The program interface is as simple as possible: press a button to start and stop recording, after which the audio is sent for processing by the neural network. Within a few minutes, the user receives a text transcription in their email inbox.

Why is the result sent via email rather than displayed directly in the service?
At this stage, it allowed us to save time and resources, as implementing features like storage, display, and search through recording history requires significant effort. This approach made it possible to develop and deliver the product to the first testers in just 7 days!

To enable access to audio file conversion, users need to complete a simple registration—this is where we collect the email address to send the data.

Technical Side

The Elevate service was chosen for speech recognition as it meets the key requirements:

• support for English and Japanese languages,
• high recognition accuracy: during the lecture, the teacher may move around the classroom, turn away, speak louder or quieter—this should not significantly affect the result,
• low usage cost (in this case, there is even a free option with a limit on the number of requests).

Text pages of the service

After the first phase of idea testing is complete, the plan is to implement a paid subscription model with three packages offering different amounts of minutes for processing. Integration with a payment service for online transactions will also be added.

Resume

With the growing popularity of AI, new services using neural networks emerge every day, making speed of execution crucial. A well-coordinated team and the right prioritization allowed us to deliver a working version in a short time.

The MVP is already being tested in the target market in Japan, and we are receiving the first positive feedback from satisfied students.