// RESEARCH & SCIENCE
AI that develops people,
not replaces them.
The question has never been whether AI can automate tasks. The more important question, the one our founders spent a decade researching, is whether AI can be designed to genuinely develop human competencies. At Magif.ai, every product decision flows from a scientific answer: yes, it can, but only when built with intention, methodology, and a deep understanding of how people actually learn and grow.
- 190
- Leaders in field research
- 4
- International organisations
- 2015
- AI coaching research since
- 2
- Peer-reviewed articles
- 5
- International conferences
- 4
- Book chapters
// THE SCIENTIFIC FOUNDATION
A doctoral thesis built in the field, not the laboratory
Magif.ai's scientific foundation rests on Dr. Olivier Malafronte's PhD thesis, defended at IAE Nice, Université Côte d'Azur: the first doctoral research to explore the mechanisms by which AI coaching conversations develop leader competence in real managerial work situations.
Developing Leader Competence in Work Situations through Artificial Intelligence and Coaching
- Author
- Dr. Olivier Malafronte
- Institution
- IAE Nice · Université Côte d’Azur
- Defended
- June 30, 2025
This thesis explores, for the first time, the mechanisms at play in AI coaching conversations between leaders and AI coaches, understanding how AI Coaching (AIC) contributes to leader competence development within the flow of real work situations.
Four research articles
Competencies Augmentation
How AI-automated coaching questions favour the regulation of cognitive and affective schemas, enabling leaders to augment their competencies in situations.
Schema Change in AIC Conversations
Theorising how AIC conversations influence leaders’ cognitive and affective schemas, making development visible and measurable in real time.
Reflectivity & the Skilled e-Coachee
Identifying the types of reflective conversations that emerge when coachees engage with an AI coach, and what this reveals about leader agency.
AI vs. Human Coaching Conversations
A comparative analysis showing that AI coaching is a distinct developmental category: not a substitute for human coaching, but a different kind of developmental tool.
// FIELD RESEARCH CONTEXTS
Tested across sectors and countries
The research was conducted across four international organisations in France, the UK, and the USA, spanning sectors where leadership development matters most.
Mining & Industry
Complex operational environments requiring rapid situational decision-making.
Software & Tech
Fast-moving organisations navigating digital transformation and team dynamics.
Higher Education
Academic institutions developing future leaders and practitioners.
HR & Professional Services
Organisations scaling leadership development across distributed teams.
// SIX CORE RESEARCH CONTRIBUTIONS
What the science reveals
Each contribution informs a product decision at Magif.ai, from how we design AI agents to how we measure development outcomes.
AI can replicate coaching conversations
AI can replicate coaching conversations and enable cognitive and affective regulation mechanisms which, repeated over time, develop leader competence directly from work situations.
Two relational orientations in leadership
Leaders socialise from an instrumental or empowering orientation. When identified by AI in real time, these orientations become valuable reflective moments for development.
AI coaching augments leader skills
Empirical evidence that AI coaching tools augment leaders’ empathy, work organisation, communication, and confidence: core leadership competencies.
Human vs. AI coaching outcomes
AI coaching achieves some human coaching outcomes while producing distinct ones that human coaches do not, establishing it as a different developmental category, not a replacement.
Scaling coaching knowledge in organisations
AI coaching enables organisations to scale coaching knowledge and tools, making developmental conversations accessible beyond the few who can access human coaches.
AI as a new source of research data
Using AI coaching tools to generate data makes it possible to observe the competence development process in the flow of real managerial situations for the first time.
// THE THEORETICAL FRAMEWORK
How AI coaching develops competence
The research identifies competence not as a fixed set of skills but as a dynamic process: the regulation of cognitive and affective schemas within real work situations. Schemas are the inner lenses through which people interpret their world and translate experience into action.
When leaders engage repeatedly with reflective AI coaching conversations, something measurable happens: they begin to notice, question, and reframe those schemas. Over time, this enables new behaviours, better decisions, and more effective leadership.
Three dimensions of competence develop through this process
Intrapersonal
Self-awareness, identity, self-efficacy, and reflective capabilities that ground effective leadership from within.
Interpersonal
Social and relational competencies: empathy, communication, empowering others, building trust.
Contextual
The ability to read, adapt to, and act within complex, fast-moving organisational environments.
The AIC augmentation loop
Work Situation
A real managerial challenge, a decision, a difficult conversation, a moment of uncertainty, activates the process.
Reflective AI Conversation
The AI coach asks powerful, open questions: the same mechanism at the heart of expert human coaching.
Schema Regulation
The leader surfaces, questions, and begins to shift cognitive and affective patterns: the mechanism of real development.
Competence Augmentation
Repeated over time, this builds measurable leadership competencies: empathy, communication, confidence, adaptability.
Repeated over time, this loop compounds into measurable competence.
// PUBLICATIONS
The published record
Peer-reviewed journal articles, book chapters, and international conference papers behind the platform.
Peer-reviewed journal articles (4)
Developing Leaders with Artificial Intelligence Coaching: Competencies Augmentation
Malafronte, O. (2024). Revue Management & Avenir, 4(142), pp. 39–59.
AI Coaching in Leadership: Awakening Competence
Malafronte, O. (2025). Developing Leaders Quarterly, 48.
The Democratisation of Coaching: Artificial Intelligence (AI) Coaching
Burnett, N. & Malafronte, O. (2017). Australian Educational Leader.
Mediating Reflectivity: AI Coaching Conversations as a Reflective Mirror of Relational Orientations
Malafronte, O. & Fatien, P. (2026). Journal of Applied Behavioral Science.
Book chapters (4)
Artificial Intelligence Coaching for Leader Development and Supervision
Malafronte, O. (2026). In Wegener, R., Garcia, T., Terblanche, N., & Grossrieder, T. (Eds.), Coaching in the Age of AI. Springer Nature Switzerland. SpringerBriefs in Psychology.
PocketConfidant
Malafronte, O. (2024). In The Digital and AI Coaches’ Handbook (Vol. 1). Routledge. ISBN 9781003383741.
IA Coaching: Pour Une Approche Responsable et Multiniveaux pour le Développement de la Compétence de Leadership
Malafronte, O. (à paraître). Management Responsable et Performance. L’Harmattan.
Récits de Mer et Inspiration pour le Management
Dumas, M., Dieuze, S., & Malafronte, O. (2025). Littérature, Management et Travail (2). L’Harmattan.
International conference papers (5)
Artificial Intelligence as a Coach for Developing Leader Competence: Exploring Coaching Conversations
Malafronte, O. & Loufrani, S. (2025). 25th Annual EURAM Conference, University of Florence, Italy.
Leadership Learning and Development: Exploratory Comparative Study of AI-Based vs. Human-Based Coaching Conversations
Malafronte, O. & Mbokota, G. (2024). 7th Biennial Conference of the Africa Academy of Management, Cape Town, South Africa.
Developing Leader Competence with Coaching Questions and Artificial Intelligence-Based Tools: Change of Schemas and Multilevel Impacts
Malafronte, O. & Loufrani-Fedida, S. (2023). 34th AGRH Congress, Corté, Corsica.
On Respectful Inquiry and Seeing Human: The Case of Technology-Mediated Developmental Conversation
Cojuharenco, I., Karelaia, N., Malafronte, O., & Potdevin, D. (2023). Academy of Management Proceedings, Boston, USA.
Comment Développer une Compétence de Leader en Situation avec Coaching et Intelligence Artificielle
Malafronte, O. (2022). 33rd AGRH Congress, Brest, France.
Awarded PhD thesis (1)
Developing Leader Competence in Work Situations through Artificial Intelligence and Coaching
Malafronte, O. (2025). Doctoral thesis in Management Sciences. Université Côte d’Azur, Groupe de Recherche en Management (GRM).
// RESEARCH IN PRACTICE
When experts build with science, the research replicates in the field
These two practitioners built AI coaching agents on Magif.ai and produced outcomes that mirror the findings of Malafronte (2025). Not in a laboratory. In real organisations, with real people, in the flow of real work.
Two people. One AI agent. New competencies, replicating research.
How Maxime, a leadership & organisational coach, used a hybrid human + AI coaching approach to develop new competencies in a team leader and a team member, replicating findings from Malafronte (2025). ICF AI Coaching Standards compliant. Opening a 10K€+ organisational opportunity.
Maxime Bui · Leadership & Organisational Coach · AI Agent Creator · Powered by magif.ai
- 10K€+
- Client deal
- ICF
- Standards compliant
- 1 → ∞
- Agency to ecosystem
One agent, three levels of impact
- 1The build
Maxime encoded his coaching method into an AI agent
Using his favoured coaching framework, Maxime created an AI coaching agent on magif.ai: his methodology, his approach, available on demand.
- 2Level 1 · The team leader
The agency director reframed his entire feedback approach
A stuck, unresolved feedback dynamic with a team member. The AI coach session helped the director see it differently, reframing how he gives feedback and how he shows up as a leader.
- 3Level 2 · The team member
A team member in shutdown recovered, stayed, and wants to go deeper
Not feeling good. Shut down. Ready to leave. Through an AI coaching session he got clarity on his situation, recovered his motivation, and stayed. He now wants to build on those outcomes.
- 4The hybrid model
The AI agent expands Maxime’s human coaching, it doesn’t replace it
Maxime coaches the director directly. The AI agent extends his reach to the team, available when Maxime isn’t, at the employee’s pace, with no time pressure. Human depth plus AI scale.
- 5Level 3 · The ecosystem
The director now wants this in every agency in his group
One conversation. Two results. Now a rollout conversation across other agency directors, with a confirmed budget of 10K€ minimum.
Replicating research, in the field, in real time
A team leader who developed new feedback and relational competencies. A team member who recovered from shutdown and built self-awareness and clarity. Outcomes that directly replicate findings from Malafronte (2025), produced by an AI coaching agent, in the flow of work, compliant with ICF AI Coaching Standards.
In their own words
“I preferred having the time to work with the AI. I didn’t want to get a time rush with a coach having to close after an hour.”
“I prefer this AI expertise coaching tool to ChatGPT, for the power of the method behind it.”
What AI coaching measurably moves
Engagement
Employees who feel genuinely supported show up differently. Measurable at team level.
Retention
One employee who nearly left, didn’t. Multiply that across an organisation.
Productivity
Clarity unlocks performance. A clear employee is a productive employee.
From hours lost to clients won, in one day.
How Roy, an expert in critical thinking, scaled his knowledge with AI coaching agents, without being a tech person.
Roy van den Brink-Budgen · Critical Thinking Expert · AI Agent Creator · Powered by magif.ai
- 1 day
- Time to first client
- < 1h
- Agent built & live
- 0
- Tech skills required
“Roy is a critical thinking expert, not a tech person. Before magif.ai, he was losing hours on his computer, unable to scale. Then everything changed.”
One day. One agent. A new business.
- 1Day 1 · Morning
Built his AI coaching agent on magif.ai
No coding, no technical training needed. Roy designed an agent that embodies his expertise in critical thinking.
- 2Day 1 · Afternoon
Got his business offer and a shareable link
magif.ai generated a ready-to-send offer and a direct link. Roy shared it with a client the same day.
- 3Day 1 · Evening
Client tried it, loved it, and transacted
The client experienced the agent, saw its value, and converted, all before the day was over.
- 4Days after
Client came back, ready to expand to his ecosystem
One satisfied client became a multiplier, opening the door to a whole new audience beyond Roy’s direct reach.
Beyond the platform, a new way to grow
Learning built into the tool
Roy didn’t buy AI or coaching training separately. He learned directly through magif.ai: Q&A on ethics, pricing, and content creation, anchored to his product and audience.
Scalability without losing the human touch
Roy reserves his human time for what he loves: sharing ideas, chatting, reading, and thinking with clients, while his agent handles scale.
Knowledge personalised for any audience
His expertise becomes applicable and personalised for each client, not a one-size-fits-all course.
Reach that wasn’t possible before
From one client link to an entire ecosystem. AI lets Roy reach many more people with the same depth of expertise.
What the research predicts, what Roy experienced
The research (Malafronte, 2025) demonstrates that AI coaching scales expert methodology beyond what direct human reach allows, making knowledge personalised and applicable to audiences who would otherwise never access it. Roy’s case is a direct real-world replication: expert knowledge, encoded in AI, reaching new people with no loss of methodological depth.
Key takeaways
Expertise first
Deep knowledge matters. AI amplifies it, it doesn’t replace it.
Time reclaimed
Hours once lost to admin are now spent on meaningful human connection.
Compounding growth
One happy client expands into an ecosystem. Word-of-mouth at scale.
// WHY THIS MATTERS
AI is not here to replace the hard work of becoming better.
There is a widespread and understandable fear that AI will take over the cognitive and emotional work that defines growth: the reflection, the discomfort, the moments of honest self-examination. Our research shows the opposite can be true.
When AI is built by experts who understand development, not just automation, it becomes a tool that holds space for that hard work, makes expert methodology accessible to those who would never otherwise reach it, and meets people precisely where they are, in the situations that matter most to them.
Magif.ai was founded on a single conviction: that the greatest application of AI in human development is not to delegate thinking, but to make the best thinking of the best experts available to everyone.
Build AI that develops people.
Join the coaches, trainers, and consultants turning their expertise into scalable AI agents, grounded in research, with their knowledge, brand, and clients under their control.