CODIO ANNOUNCES THE LAUNCH OF "COACH," AN AI-POWERED LEARNING ASSISTANT THAT UNLOCKS LEARNER POTENTIAL WITH AUGMENTED ERROR MESSAGES AND HINTS

Codio leverages generative AI to tackle the critical barriers to learner success in tech skills and computer science education.

CAMBRIDGE, Mass., April 3, 2024 /PRNewswire/ -- Codio, the leading specialist platform for computer science and tech skills education, today announced the launch of Codio Coach, a conversational AI learning assistant that leverages generative AI to provide personalized learning assistance through targeted guidance and error explanations. Learners can continue to make progress on their learning journey outside of office hours, or when instructor/mentor support might not be available. Coach delivers carefully calibrated support that fosters learning without providing direct answers, ensuring an authentic learning experience. Any language supported by Codio is compatible with Coach.

Coding learners spend on average 25% of their hands-on time trying to resolve errors in their code, and whilst programming error messages have been researched for decades, until now, attempts to scale augmented error messages to accelerate learning have fallen short as LLMs, without the necessary context, proved to provide accurate explanations in no more than 50% of cases. "In Codio's recent survey of student use of generative AI, 92% of students perceived error messages as a barrier to progress," explains Mohit Chandarana, Data Science and AI Specialist at Codio.

Now with Coach, Codio has used real-time context-augmented generative AI to optimize the output of a large language model, so it has all the relevant information it needs along with a rich set of guidelines before generating a response. "Coach leverages not just the error message but the related code files and assignment text to provide more accurate and meaningful error explanations, enabling learners to go from confused to coding," Chandarana adds.

Coach is powered by the Claude-3 models from Anthropic, for their long context windows and exceptional performance at handling large, structured prompts.

In addition to error messages, Coach breaks complex tasks into smaller steps, providing summarized requirements and hints on next steps, helping learners move from stuck to unstuck, and creating a more approachable learning experience. The generated hints, modeled on the Socratic method, are contextual, and phrased as questions or suggestions, that guide students towards making progress by themselves, without giving away code or solutions.

"We've grown Codio by obsessively focusing on the needs of our customers and closely collaborating with instructors and researchers," says Phillip Snalune, CEO of Codio. "Our earlier research highlighted the challenges educators face in scaling instructor support for computing and tech skills programs. By embracing generative AI with the launch of Coach, Codio is tackling critical gaps in the learning journey, empowering educators, reducing learner attrition, and driving better completion rates."

Codio has a long history of implementing evidence-based interventions to assist teachers to teach more effectively and to create a more engaged experience for learners. In that vein, Coach is extensible and configurable, complementing a teacher's pedagogical approach and providing complete control over how AI assistance is provided. Codio believes this human-in-the-loop approach makes for a more authentic and ethical use of AI.

About Codio

Founded in 2015, Codio is the leading specialist platform scaling accessible, high-quality learning experiences in computing and tech skills education.

Leveraging computing education research to provide educators with an easy to use and scale tool set for teaching tech skills and computer science at scale, Codio is the platform of choice powering the leading computer science schools, bootcamps and workforce development partners globally, empowering educators, and boosting learner engagement and completion rates. Codio's tech stack scales a wide variety of hands-on learning experiences from introductory programming through to advanced courses in AI, ML, Cybersecurity and Data Science, with uniquely interactive learning experiences and unparalleled course management and learning insight tools for educators.

For inquiries, please contact: [email protected]