Introduction: From Vision to Reality
Seven years ago, I began working on an idea to create the next generation of learning products. My vision: short, personalized learning paths that would adapt to each learner’s needs and be accessible on any device.
I called it ARI — Adaptive Response in Real Time — after both its core principle and my daughter, Ari.
Back in 2018, building even basic prototypes was unaffordable for a small team. Today, with AI’s maturity and the availability of adaptive technologies, ARI is no longer a dream. It is now a working prototype: customizable, scalable, and designed to bring personalization to life.
Why Personalization Matters
Traditional eLearning has always struggled with one problem: it’s one-size-fits-all. Content is often static — every learner gets the same module, the same quiz, and the same feedback. But each learner is different. They bring unique goals, knowledge gaps, and motivations.
Evidence consistently shows the power of personalization:
- 30% higher test scores in personalized learning programs
- 8-point gain in math and 9-point gain in reading
- 15% lower dropout rates and 12% higher attendance
- 75% student motivation rate in personalized settings vs. 30% in traditional classrooms
Source: Matsh. (2024). Statistics on personalized learning effectiveness.
AI is the missing ingredient to scale personalization. Without it, individualized learning is possible only in small classrooms or tutoring sessions. With AI, adaptive learning becomes scalable across organizations and geographies.
The ARI Loop: Adaptive Learning in Action
Every learner creates a profile: goals, prior knowledge, role, language, preferences. ARI uses this information to adapt the training proposal — content, sequence, and the conversation itself. Two learners might watch the same video, but they won’t have the same experience: questions, scaffolds, and follow-ups change on the fly based on what each person enters and how they respond.
The ARI system follows a five-step loop to make personalization scalable and human-like:
Profile → Propose → Interact → Assess → Adapt
1. Profile
Learners create a personal profile, including goals, preferences, background, and prior knowledge. This serves as the foundation for adaptation.
2. Propose
ARI generates a tailored learning proposal: individualized course content aligned with learners’ goals, competencies, and progress.
3. Interact
Learning becomes a dialogue. Learners engage, answer, reflect. At every step, ARI gives contextual, conversational guidance. SMEs can review/approve rules and exemplars.
4. Assess
Quizzes, exercises, and reflective questions dynamically adjust to skill levels, closing knowledge gaps immediately.
5. Adapt
The system updates the path in real time, ensuring content stays aligned with the learner’s progress, competencies, and evolving needs.
The result: a living learning ecosystem that grows and adapts with each learner
Why Now: The Market is Ready, and So Is the Technology
Recent research highlights why 2025 is the right moment for ARI:
- McKinsey (2024–2025): AI adoption accelerated sharply. In early 2024, 65% of organizations reported regular use of generative AI, and 72% used AI in at least one business function — signaling widespread integration. By mid-2025, 90% of employees were using Gen AI at work, with 21% being heavy users. In the L&D context, McKinsey now advocates for personalized, data-based development plans and rotational apprenticeships.
- OECD (2019): The OECD Learning Compass 2030 defines student agency as the ability to set goals, reflect, and act responsibly to shape one’s own learning and that of society. It identifies agency as essential for helping learners navigate complexity and uncertainty.
- World Economic Forum (2020, 2024): The Future of Jobs Report 2020 projects that 39% of workers’ core skills will change by 2030. In 2024, WEF’s Shaping the Future of Learning: The Role of AI in Education 4.0 highlighted the potential of AI to personalize learning and support large-scale reskilling initiatives.
- Association for Talent Development (2025): ATD’s AI in Instructional Design report found that 80% of instructional designers are already using AI in their workflows, with many reporting improved course quality and faster design cycles.
- Bond Capital (2025): The Trends in Artificial Intelligence report highlights the unprecedented growth of AI adoption, the rapid rise of multimodal models, and the falling cost of inference — conditions that make real-time, scalable applications increasingly feasible.
Together, these findings show that both market demand and technological readiness have aligned in 2025.
Governance: Building Trust in Adaptive AI
Personalized learning cannot come at the cost of trust. ARI is designed with governance and ethics at its core:
- Accuracy & SME checkpoints: Human review ensures content quality.
- Copyright & IP controls: Source tracking for AI-generated assets.
- Data privacy: Secure handling of learner data with opt-in transparency.
- Bias mitigation: Regular audits and fairness testing across learner profiles.
By design, ARI is not just adaptive but also transparent, ethical, and safe.
Beyond Content: Emotional & Multimodal Personalization
The future of ARI goes beyond static course content:
- Multimodal support: Personalization through sound, visuals, and interactivity, including better support for neurodivergent learners.
- Emotional engagement: Systems that adapt tone, pacing, and cognitive load based on learner context and state.
- Global reach: Multilingual delivery that adapts while maintaining tone and context.
Conclusion
The ARI system brings together personalized learning, real-time adaptation, and ethical design. By combining technology with trust, ARI makes personalization scalable, equitable, and truly learner-centered.
The future of learning isn’t more content. It’s better conversations with learners. ARI is where that future begins.
References
Association for Talent Development. (2025, August). AI in instructional design: Transforming workflows and content creation. Association for Talent Development. https://www.td.org/product/research-report--ai-in-instructional-design-transforming-workflows-and-content-creation/792512
Bond Capital. (2025). Trends in artificial intelligence. Bond Capital. https://www.bondcap.com/report/pdf/Trends_Artificial_Intelligence.pdf
Matsh. (2024). Statistics on personalized learning effectiveness. Matsh. https://www.matsh.co/en/statistics-on-personalized-learning-effectiveness/
McKinsey & Company. (2024). The state of AI in early 2024: Gen AI adoption accelerates (McKinsey Global Survey on AI). McKinsey & Company. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024
McKinsey & Company. (2025, March 12). The state of AI: How organizations are rewiring to capture value (Global Survey on AI). McKinsey & Company. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
McKinsey & Company. (2025, June 17). Reimagined: Learning & development in the future of work. McKinsey & Company. https://www.mckinsey.com/featured-insights/people-in-progress/reimagined-learning-and-development-in-the-future-of-work
McKinsey & Company. (2025, July 9). The learning organization: How to accelerate AI adoption. McKinsey & Company. https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/the-learning-organization-how-to-accelerate-ai-adoption
OECD. (2019). OECD future of education and skills 2030: OECD learning compass 2030. Organisation for Economic Co-operation and Development. https://www.oecd.org/education/2030-project/
World Economic Forum. (2020). The future of jobs report 2020. World Economic Forum. https://www.weforum.org/reports/the-future-of-jobs-report-2020
World Economic Forum. (2024, April 28). Shaping the future of learning: The role of AI in education 4.0. World Economic Forum. https://www.weforum.org/publications/shaping-the-future-of-learning-the-role-of-ai-in-education-4-0/
To cite this paper:
Villar, M. A. (2025). A.R.I.: Adaptive Response in Real Time – Creating an ecosystem for personalized learning. Zenodo. https://doi.org/10.5281/zenodo.17025683
