Research Study — Open to Coaches Globally
Does AI-assisted reflection between sessions strengthen coaching outcomes?
A peer-reviewed randomized controlled trial in partnership with Stellenbosch University. Led by Prof. Nicky Terblanche.
The Research Question
A question the field needs answered
Between sessions, clients often lose the thread — insights fade, intentions drift. What if a brief, structured reflection could keep the momentum alive?
This study investigates whether adding a reflective AI between coaching sessions improves client outcomes — goal attainment, working alliance, wellbeing — without disrupting what makes human coaching effective.
The results will be submitted for peer review and published. Participating coaches will receive access to the findings and may reference the study in their practice.
We are investigating whether structured AI-assisted reflection between coaching sessions improves client outcomes — without weakening the working alliance with the human coach.
Study objective — Prof. Nicky Terblanche, Stellenbosch University
Study Design
Two conditions. One rigorous question.
This is a two-arm randomized controlled trial. Coaches are randomly assigned 50/50 to one of two conditions. All clients follow the same condition as their coach — this prevents contamination and preserves methodological validity.
Control arm
Human coaching only*
4–5 sessions over 6–10 weeks. Standard ICF-aligned process. No changes to how you currently work. The control arm is not a lesser contribution — it is essential to the research.
Experimental arm
Human coaching + AI reflection
The same coaching process, plus your client uses a structured CoachBot between sessions. Before each session, you receive a short AI summary. You open with: “What came up in your reflections?”
*Control group coaches will be offered access to the AI assistant after the study concludes to use for same period as experimental group.
Client Experience
The Human Coaches deliver Coaching Sessions and weekly E-mail Reminders for the AI-led coaching support between sessions (depending on which group the client ends up with).
The Client will receive a survey at the end of the program.
What participating in this study involves
A 10–12 week commitment. Designed to be lightweight.
We have designed your participation to be as lean as possible while preserving scientific integrity. Here is what you commit to as a participating coach:
- Bring one new client who has not been previously coached by you and who is willing to engage with AI
- Commit to 5–6 coaching sessions over 10–12 weeks (biweekly schedule preferred)
- Complete a pre- and post-study survey — approximately 5–10 minutes each
- Attend a kickoff briefing session before the study begins and sign the participation agreement
- You may withdraw at any time — there is no obligation to complete the study
The CoachBot
Reflection support. Nothing more.
The CoachBot is not a coaching substitute and does not conduct sessions, onboarding, or contracting. It exists only to help clients consolidate and carry their insights between sessions. Each interaction is capped at approximately 10–15 minutes.
Human coaching sessions are never recorded or ingested by the AI. You remain in full control of session direction throughout.
Borton's structured reflection model
- What stood out from your last session with your coach?
- What did you learn or notice?
- What will you do before your next session?
- What do you want to bring into your next session with your coach?
The CoachBot explicitly references and reinforces you as the coach throughout. If a client gets stuck, it guides them through a simplified CBT reflection flow (ABCDE framework) — before returning to the coach-led path.
What we measure
Rigorous. Published. Yours to reference.
The study measures both client and coach outcomes. Results will be submitted for peer review. Participating coaches receive early access to the findings and may cite the study in their professional practice.
- Goal Attainment
- Working Alliance
- Self-Determination
- Coaching Satisfaction
- Mental Wellbeing
Ethics & data protection
Grounded in research ethics. Compliant by design.
This study operates under a formal ethical framework overseen by Prof. Nicky Terblanche through Stellenbosch University. Registration and data processing are handled by Viva la Coaching Academy GbR — fully compliant with GDPR and hosted on EU servers.
Dual Consent
Both coach and client must provide explicit consent before the first session and before any CoachBot interaction. Either party may withdraw at any time without consequence.
Data Anonymization
All personal data is anonymized prior to analysis. CoachBot conversations are recorded only with explicit consent. No session transcripts are ever ingested by the AI.
GDPR Compliance
All data is stored and processed on EU infrastructure. Participants have full rights to access, edit, and delete their data at any time. Data ownership and deletion rights are transparent.
EU Servers
Registration is processed through Monday.com on EU-hosted servers via Viva la Coaching eGbR. Follow-up communication uses Brevo, a GDPR-compliant email platform.
Research team
Who is behind this study
An interdisciplinary team spanning academic research, coaching methodology, AI design, and platform engineering.
Prof. Nicky Terblanche, PhD
Stellenbosch University
Study design & analysis
Rebecca Rutschmann
Viva la Coaching
CoachBot prompt design
Jonathan Reitz
FLUXIFY
Coach recruitment
Lewin Keller
CoachBot.aI
Platform & Compliance
Common Questions
What is this study actually investigating?
The core research question is: Is human coaching enhanced by an AI reflection tool between sessions more effective than human-only coaching? The study uses a randomized two-group design — one group coaches as usual, the other adds the CoachBot between sessions. We measure outcomes across both groups to see where, if any, differences appear.
Will the CoachBot read my session transcripts?
No. Human coaching sessions are not recorded, ingested, or analyzed by the CoachBot at any point. The bot interacts only with clients between sessions, independently of what happens in your sessions. If you are in the experimental group, you receive a short summary of CoachBot interactions before each session — that is the only connection point.
How should I use the CoachBot outputs in my sessions?
In the experimental group, open each session by asking: “What came up in your reflections?” This keeps alignment strong and prevents derailment. You decide what to incorporate and what to leave out. The AI does not dictate session direction — that remains entirely yours. Coaches in the control group coach according to their current model and preferences, as they normally would.
What if I am assigned to the control group?
You coach exactly as you do today. The control arm is not a lesser experience — it is a scientifically essential part of the study. Your participation is equally valuable. You will still receive the published findings and can reference the study in your practice.
What if the CoachBot feels disruptive to my client?
Previous qualitative work with this design found that some coaches initially felt disruption, while clients broadly appreciated the structured support between sessions. The guidance is simple: open each session with the reflections question, and let the session go where it needs to go. If the CoachBot content benefits the client, use it. You remain in control throughout.
How long should my client spend with the CoachBot?
Each interaction is designed to take approximately 10–15 minutes. This is intentional — enough time for meaningful reflection, not enough to create fatigue or dependency. The design includes a natural stopping point to prevent extended engagement.
What does my client need to do?
Your client must be new to you (not previously coached by you), willing to engage with AI, and able to commit to the full coaching cycle. They will need to give explicit consent before any CoachBot interaction, access the bot via a web browser (mobile or desktop), and complete a short pre- and post-study survey. If a client declines CoachBot use during onboarding, they are removed from the study.
How is my client's data protected?
All data is anonymized and handled in full compliance with GDPR. The platform runs on EU-hosted infrastructure. CoachBot conversations are recorded only with explicit, dual consent from both coach and client. Participants have transparent access to their stored information and can request deletion at any time. No personal session content is ever shared with the research team.
How are coaches assigned to groups?
Assignment is random — 50/50. All clients of a given coach follow the same condition, which prevents contamination between groups. You will be informed of your group assignment, since clients in the experimental arm must be told about CoachBot access as part of the consent process.
How is consent and ethics handled?
Dual consent is required from both coach and client before the first session, and separately before any CoachBot interaction. Ethics oversight is managed through Prof. Terblanche’s board at Stellenbosch University. Both parties may withdraw at any time without consequence. All processes are GDPR-compliant and EU-hosted.
Will the prompts and methodology be published?
The peer-reviewed paper will include the main reflection flow structure for replication purposes. Internal monitoring and quality prompts will not be published. Participating coaches receive early access to findings before publication.
What if I cannot source a client?
The primary design is one coach, one new client. Backup options may be explored, but these can affect randomization and matching. If you are uncertain about sourcing a client, please note this in your registration — the coordination team will be in touch to discuss options.
What recruitment materials will I receive once registered?
Registered coaches receive a one-page study flyer, access to the coach briefing session, automated onboarding emails, consent materials for both coach and client, and email reminders throughout the study period. All communication is handled through Viva la Coaching eGbR via Brevo.
What is the expected timeline?
The study is targeting a Q2 2026 launch. The exact start date is to be confirmed following ethics approval and final logistics. Each dyad (coach + client) runs for approximately 10–12 weeks. You will be notified by email once the start date is confirmed.
