Comparing AI and human coaching goal attainment efficacy
AI vs. Human Coaching: Who Helps Us Reach Our Goals More Effectively?
In the PLOS ONE research article “Comparing artificial intelligence and human coaching goal attainment efficacy,” authors Nicky Terblanche, Joanna Molyn, Erik de Haan, and Viktor O. Nilsson examine whether an AI chatbot coach can help people achieve goals as effectively as human coaches. Published June 21, 2022, the peer-reviewed study compares two longitudinal randomized controlled trials conducted over ten months: one involving human coaching and one involving an AI chatbot coach named Vici.
For business leaders, the study is highly relevant to the future of leadership development, employee coaching, AI-enabled learning, and workforce capability building. Its central finding is surprising: both human coaching and AI coaching significantly improved participants’ goal attainment compared with control groups, and by the end of the study, the AI coach performed at a level comparable to human coaches in this narrow goal-attainment context.
Executive summary for business leaders
Overarching theme: AI coaching may not replace human coaching as a whole, but it can be effective when applied to a narrow, structured coaching task grounded in proven theory. The authors emphasize that coaching is often expensive and limited to senior leaders, while AI coaching could make certain coaching benefits more scalable, affordable, and widely available.
The study compares outcomes across four groups: two control groups, one human-coaching group, and one AI-coaching group. Across the two studies, 327 participants submitted data across all eight time points. The researchers found that the human-coaching and AI-coaching groups had notably higher goal-attainment gains than the control groups, and the human and AI coaching groups did not significantly differ from each other during the experiment.
The leadership lesson is not “replace coaches with bots.” It is more precise: use AI where structure, consistency, availability, and repetition matter; use human coaches where empathy, complex sense-making, contextual judgment, and emotional intelligence matter most.
Major takeaways
1. AI coaching can rival human coaching in a narrow goal-attainment task
The study found that both human coaching and AI coaching produced significantly greater goal attainment than control conditions. By the end of the ten-month trials, the AI coach had results similar to the human-coaching group, with reported effect sizes of ηρ² = .265 for human coaching and ηρ² = .269 for AI coaching.
Business implication: AI coaching may be useful for structured development tasks such as goal setting, progress tracking, action planning, and accountability check-ins.
2. The AI coach worked because it was narrow and theory-based
The AI chatbot Vici was not designed to be a general-purpose executive coach. It was built around goal theory, helping users define realistic goals, assess feasibility and impact, create action plans, check progress, reflect on obstacles, and adjust actions.
Business implication: Companies should avoid broad claims about “AI coaching” and instead define specific use cases where AI can perform reliably.
3. Consistency was one of AI’s advantages
The authors suggest that Vici’s strength came from its rigorous, consistent execution of goal theory. Unlike human coaches, the AI coach always asked about progress, kept records, and required users to type their goals into the application.
Business implication: AI can be especially valuable where leaders need standardized coaching routines across large populations.
4. Availability matters
Vici was available 24/7, and participants could use it as often as they wanted, with a minimum of once per month. The authors note that participants who used the AI coach more often had higher goal attainment.
Business implication: AI coaching can reduce access barriers by giving employees support when they are ready to reflect, act, or course-correct — not only when a scheduled session is available.
5. AI could democratize coaching
The article notes that professional coaching is often expensive and inaccessible, especially outside affluent markets. The authors argue that AI coaching could extend basic coaching benefits to people who would otherwise never receive coaching.
Business implication: Organizations could use AI coaching to support broader employee populations, not only executives and high-potential leaders.
6. AI may increase demand for human coaching
The authors suggest that AI coaching could expose more people to the benefits of coaching, potentially increasing demand for more advanced human coaching later.
Business implication: AI and human coaching may become complementary: AI handles structured, scalable support; human coaches handle complex, high-stakes, relational, and developmental work.
7. Human coaches still have important advantages
The study is careful not to claim that AI coaching outperforms human coaching overall. The authors state that AI currently lacks the empathy, emotional intelligence, context awareness, and complex sense-making that make human coaching powerful.
Business implication: Human coaches remain essential for senior leadership development, conflict, identity work, emotional complexity, team dynamics, ethical judgment, and major transitions.
8. Low-complexity coaching may be most vulnerable to AI substitution
The article discusses “coach maturity” and warns that coaches who rely mainly on simplistic, model-based approaches may be easier for AI tools to replicate.
Business implication: Internal and external coaches need to move beyond formulaic coaching scripts and deepen their capability in context, systems thinking, emotional intelligence, and adaptive judgment.
9. AI coaching raises ethical issues
The article highlights ethical concerns including prevention of harm, algorithmic transparency, bias, client autonomy, and data ownership.
Business implication: Any organizational use of AI coaching should include clear governance, consent, data protection, escalation paths, and boundaries around sensitive topics.
10. The evidence is promising but limited
The participants were undergraduate students, and goal attainment was measured through self-reported scores. The authors note that these limitations may affect generalizability to other populations.
Business implication: Leaders should pilot AI coaching carefully before scaling, especially with executives, frontline workers, or high-risk employee groups.
Leadership talking points
AI coaching should be used where it fits the work: structured, repeatable, goal-oriented, and measurable.
Human coaching remains critical where empathy, trust, emotional intelligence, and complex judgment matter.
AI can expand coaching access beyond senior leaders and high-potential programs.
A strong AI coaching strategy should not ask, “Can AI replace coaches?” It should ask, “Which coaching tasks can AI handle well, and where do humans add irreplaceable value?”
The quality of the coaching model matters. AI tools built on proven behavioral and goal-setting theories are more credible than generic chatbots.
Ethics, privacy, transparency, and escalation rules must be designed before rollout.
Reflection questions
Which coaching needs in our organization are structured enough for AI support?
Where are employees missing basic goal-setting, progress tracking, and accountability support?
Are human coaches being used for work that could be handled more affordably and consistently by AI?
Where do we need human coaching because the issues involve emotion, identity, conflict, judgment, or leadership complexity?
How will we protect employee privacy and autonomy if we introduce AI coaching?
What data will the AI coach collect, who can access it, and how will it be used?
How will employees know when to escalate from AI coaching to a human manager, coach, HR partner, or mental-health resource?
Potential action items
Pilot AI coaching for narrow use cases such as goal setting, action planning, habit tracking, onboarding milestones, learning goals, or career-development check-ins.
Use human coaches for complex leadership development, executive transitions, conflict, performance challenges, and high-stakes interpersonal issues.
Require any AI coaching tool to show the behavioral theory or coaching methodology behind its design.
Create clear consent, privacy, data-retention, and escalation policies before employees use the tool.
Measure outcomes beyond usage, including goal progress, employee confidence, manager feedback, behavior change, retention, and well-being signals.
Train managers and coaches on how AI coaching fits into the broader development ecosystem.
Offer AI coaching as an optional support tool first, then evaluate trust, adoption, and effectiveness before wider rollout.
Review equity of access: AI coaching may be especially valuable for employees who currently receive little or no development support.
Recommended similar articles
“4 Styles of Coaching—and When to Use Them” — A practical HBR companion article on adapting coaching style to employee needs, task complexity, and required guidance.
“Successful Leaders Are Great Coaches” — A useful HBR article on why coaching is central to leadership, employee growth, and performance.
“How Coach K Does It” — A leadership interview from Chief Executive on coaching, empowerment, accountability, preparation, and team culture.
“The Playbook for Auditing AI Opportunities” — A practical AI opportunity-audit article for leaders evaluating where AI can add measurable value.
“Why Agentic AI Projects Fail—and How to Set Yours Up for Success” — A helpful HBR companion on selecting AI use cases carefully and building governance before scaling.
“Executives love their AI rollouts, but employees aren’t buying it” — A CIO article that reinforces the importance of employee trust, AI literacy, and change management in AI adoption.