Birdhouse was an internal project featuring connected and interactive birdhouses. Initially launched during Prototyp Week 2022, the project remained incomplete. In 2024, a new team expanded the product with advanced AI functionality. Birdhouse AI was born.
Press to view video
When the Birdhouse project was initiated during Prototyp Week 2022, the primary goal was for employees to record their own messages in the office birdhouse and send them to other offices. However, due to a compatibility issue, the project was never completed.
The challenge for the new team at the 2024 hackathon was to identify what had gone wrong previously and incorporate AI to improve user experience.
This time, the developers focused on two primary features named Bird-to-Bird Communication (B2B) and Bird Oracle (BO). Each feature was activated by its own button on the birdhouse.
The B2B feature is straightforward: the user presses a button to activate the recording function and sends the message to birdhouses at other offices. A B2B light blinks to indicate a new message is waiting to be heard − similar to how voicemail systems worked before mobile phones.
The BO button is even more intriguing, as it allows the user to ask questions to the mighty Bird Oracle. The team used AI tools such as Whisper for speech recognition, ChatGPT for generating answers, and ElevenLabs for text-to-speech synthesis.
Development was carried out in Python, which the team described as “equal parts gentle breeze and compatibility chaos.” Most issues revolved around deciding where and how the project should run, with a large portion of debugging focused on ensuring all components worked under the same configuration. The solution involved REST APIs to enable seamless communication between systems.
This time, the team successfully completed the project, albeit with only 20 minutes to spare before the demo presentation. A key factor in their success was building upon the previous team’s development environment, which saved considerable time.
Having two specific and well-defined features also made it easier to divide the workload. Each participant quickly took ownership of their part, enabling efficient and focused collaboration without bottlenecks.
The project offered several valuable lessons on how even seemingly simple features can create headaches during development − but that problems can almost always be overcome with collaborative efforts.
The team sees potential in further exploring additional AI functionalities for the birdhouses. Future iterations could enhance the ability to predict answers based on past questions.
There is also great potential in adding sentiment analysis, a branch of natural language processing that would enable the AI model to better understand the user's tone, emotions, and mood.
We use cookies to give you a better experience when visiting our website. Read more about how we handle cookies