The business determined Google Cloud was the best fit for its needs. “Google Cloud provided the easiest management and maintenance, and the partner we selected to help deliver the project was extremely comfortable with its selection,” said Newton. “Furthermore, when we started talking to the Google Cloud team, we realized its people and services could support us across the entire process, which was quite complex and required a lot of thought. In fact, Google Cloud could provide us with far more support than we had ever experienced before.” 

iCanStudy and addaxis.ai began working with Google Cloud in May 2023 on elements of the architecture for its new platform. By September the same year, iCanStudy had completed this work and in December that same year, publicly released the platform running on Google Cloud. The business has also developed two key AI products for iCanStudy customers: a Deep Search Engine and a generative-AI-based Meta Analyser. iCanStudy’s Deep Search engine enables users to write questions that aim to improve their cognitive skills and, based on program progress, provides an answer customized to their stage of skill development. For example, learners can ask cognitive questions like "My astrophysics exam is in three days and it has lots of procedural knowledge. How should I revise?" To emotional questions like, "How do I adopt a growth mindset when I always self-sabotage myself?" The answer will be personalized to their current skill level, while recommending specific resources to view next.

As a world's first, the Meta Analyser evaluates a learner's cognitive learning processes and diagnoses issues based on 100 identified patterns, which completely revolutionizes how to train one's ability to learn. Depending on the thousands variables considered in its analysis, users can submit a photo of their mind map notes or a reflection on an experience, and the Meta Analyser will provide step-by-step instructions on how to improve their next learning session, essentially acting as a coach who can see how your mind is thinking and address the areas for improvement.

The business undertook AI development, testing and prototyping in Vertex AI, while the Gemini large language model powers its proprietary learning products. 

The learning platform is run by a resilient, scalable cloud-based architecture, comprising services such as Firestore data storage, VPC networking, Cloud Run serverless containerization, and Compute Engine virtual machines. 

“We have become the first edtech business in Australia to build on Google Cloud using generative AI through Gemini and Vertex AI,” explained Tsai.

With Google Cloud, iCanStudy is now well positioned to achieve its target of more than one million active users, a 200X increase from its previous capacity. “The infrastructure provided by Google Cloud enables us to scale as we need to,” said Newton.

The AI-powered products developed in Vertex AI have eliminated the need for iCanStudy to invest in 1,000-plus hours training (human) coaches before they can teach students. Additionally, user feedback is received 288X faster than before, enabling the business to respond quickly to any perceived issues. As well as increased uptime, CPU utilization has fallen from about 94% on average to just 16%, accelerating load times and improving the online user experience. 

“Our partnership with Google Cloud aligns perfectly with Google's mission to make data accessible and useful for everyone,” said Tsai. “As we continue to advance our Learning Experience Platform and AI capabilities, our goals are to harness this wealth of data to revolutionize education and empower learners globally.”