OIAI, from the Otermans Institute, did something amazing in 2021. They created the first AI teachers. These smart digital teachers helped students learn in many places, even in refugee camps.
Now, their AI, called OIMISA-7B, teaches different subjects all over the world, in universities and companies. Their journey is inspiring. They worked hard to make learning better and more fun with technology. Their courses are so engaging that many more students finish them compared to other online courses.
This story isn't just about a company. It's about making a big difference in education with new ideas. It shows us how technology can change learning for the better.
Let's hear directly from them!
Dev was awarded by the UK Prime Minister for his work and has been a Young Global Innovator and Under 30 Social Entrepreneur (Mint). For research work in AI, he was recognised by Innovate UK and secured $250,000 in funding just in 2022.
He is an AI expert with research experience at The Alan Turing Institute and Brunel University London and his work is currently operating in 15 countries.
Dr. Pauldy Otermans is a female tech leader in the UK. She is a neuroscientist and psychologist by academic background and a female leader of AI technology.
She was awarded as the inspirational woman in the Tech Industry by Hustle Awards, was named one of the ‘22 most influential women in the UK of 2022’ by Start-Up Magazine UK, and has also been awarded by the UK Prime Minister in the UK in 2021, and globally for her work.
She is currently the topic Editor for Frontiers in AI: The Role of Conversational AI in Higher Education, and Chair of AI in Education forum London.
We released the world's first digital human teachers completely powered by AI (our own model) in 2021 (pre-ChatGPT) that taught learners including at UNHCR-supported refugee camps. Since then, we have produced other iterations of the product and our pedagogy which earned USD 300K revenue.
The company also taught learners in 15 countries to date. Our current in-market product is version 2 of the product and version 3 of our language model OIMISA-7B; a specialised fine-tuned 7 billion parameter language model specialised for teaching and learning. The AI teachers of this current version taught entire courses for corporates and higher education institutes from Q4 2023 - first in the world.
The avatar used to humanise our AI are also proprietary technology developed by the company. Our AI teachers are content agnostic and can teach any lesson or entire course (e.g., currently teaching statistics at a University, teaching Health Sciences, conducting teacher training, doing compliance training at corporates, etc.).
For solution-market fit, we have extremely positive outcomes where even our free courses without accredited certificates show over 64% completion rates as compared to 7-14% globally in digital learning. Also, in every one-hour lesson, a unique user asks 4 unique questions to the AI.
We make every piece of learning outside of the classroom two-way conversational.
This means we provide a personalized AI teacher to all students of Higher Education institutions and all employees of corporates that can teach them, test them, and answer their queries and clarify their doubts in real-time.
Simultaneously, we provide intricate data from each learner’s learning journey to institutions and corporates to enable them to increase their support at a highly personalized level.
We released the world's first digital human teachers completely powered by AI (our own model) in 2021 (pre-ChatGPT) that taught learners including at UNHCR-supported refugee camps.
Since then, we have produced other iterations of the product and our pedagogy which earned USD 300K revenue. The company also taught learners in 15 countries to date.
The AI teachers of this current version taught entire courses for corporates and higher education institutes from Q4 2023 - first in the world. Similarly, it teaches and trains employees for corporate clients including a FT500 company.
Finding solution-market fit: Using Generative AI to create something cool does not lead to product-market fit. Our current product is a result of several iterations and is a version 2 offering.
Understanding the exact pain points of our customers, HE institutions and corporates, was the biggest challenge. In this process we discovered that, while the AI teacher was extremely beneficial for the users (i.e., students and employees), the ability to edit AI-generated content and getting personal level learning journey data points of each learner or employee was of paramount importance to the customer.
This discovery took 6 months to complete and involved interactions and trials with several universities and corporates.
We spoke to existing fund owners, current and former entrepreneurs to firstly gather feedback on our pitch deck and secondly to get warm introductions to investors.
We also generated a list of 386 Angel and early-stage investors from LinkedIn and called and contacted them via LinkedIn.
It is extremely difficult for first-time founders to reach the current investors for their startups. This process is usually faced with several hurdles like artificial gatekeepers, the need for warm introductions, and lack of direct contact information for many funds.
To tackle these, we spoke to existing fund owners, current and former entrepreneurs to firstly gather feedback on our pitch deck and secondly to get warm introductions to investors.
We are making parts of our product offering Edge-AI enabled so that we can further improve user experience and data through lower latency, protect part of our user information by allowing them to remain within the device, and make our cost structure more competitive.
Within the next 12 months, we are growing our new customer onboarding 4x.
We are most excited about three points:
- Edge-AI versions of LLMs that can enable part of large models to run directly on devices.
- The growth of open-source generative AI models and tools that are greatly accelerating early-stage start ups in this space.
- The rate at which cost of training LLMs is falling year on year; 60% annually. We are greatly concerned about large closed-source generative AI companies who are utilizing massive funds to limit the market of key supplies like A100 GPUs.
Simultaneously, we are very concerned about companies like Microsoft completely uprooting top talent from competing, well-funded AI companies like Inflection AI and Stability AI.
- Focus on solving one problem instead of many at a time.
- While building products that become a fit for the market, hire ambitions generalist and pro-active problem solvers instead of specialists
or
Wix
Resources
Startup Events
Live Chat