AI assistant doubles completion rates in tech skills training
New study shows that embedding AI-powered support into online tech skills courses can double completion rates and boost grades by 15%, highlighting AI’s potential to improve learner outcomes across both education and workforce training settings.
New research from Codio, an AI-enhanced learning experience platform, shows that embedding AI-powered learner support into online tech skills courses can double completion rates and improve grades by 15%.
The large-scale study, covering 1,800 learners across 39 computer science courses, evaluated the impact of Codio Coach, an AI assistant built on Anthropic's Claude model. The assistant, used in both higher education and professional upskilling contexts, provides plain-language explanations of programming errors, Socratic-style hints and summarised assignment briefs.
Error explanation was the most-used feature, making up 56% of interactions. Guided hints (27%) and assignment summaries (17%) followed. The tool saw highest usage in introductory and intermediate courses, key stages for attrition and learner disengagement.
'These findings reinforce how AI, when embedded thoughtfully, can drive significant learning impact,' said Philip Snalune, CEO and Co-Founder of Codio. 'As demand for tech skills grows, scalable and effective learner support becomes essential.'
Launched in April 2024, Codio Coach has logged over 9,000 usage events within three months and is used by a mix of academic institutions and workforce training providers. The research, based on course data from 2023 to 2024, will be presented at the 2025 ASEE Annual Conference in Montreal.
The findings support a broader pattern: generative AI in learning platforms is increasingly used for real-time learner support, identifying disengaged users, delivering adaptive feedback and assessments, and providing personalised guidance. When thoughtfully designed, for example using scaffolding rather than giving answers, these tools can enhance comprehension without compromising learning integrity or trust in the process.