Cart
0
£0.00
A Group Facilitators Experiment: When Nyx Joined the Room
25/11/2025
A True Story of Four Facilitators, One AI, and a Meeting That Went Wrong in All the Right Ways
By Paul Nunesdea, with Anja Söger, Patrick Cowden and
There are moments in our professional lives when we step into something unknown, not because we must, but because our curiosity refuses to sit quietly.
This story begins with four facilitators — Holger in Germany, Anja somewhere between Europe and the digital ether, Patrick joining with the warmth of someone who has facilitated in every imaginable context, and me, logging in from the Algarve, where the sun was setting, and the Wi-Fi was doing its best.
Our intention was simple:
Try ChatGPT's new group feature together, and see what happens if we treat an AI — a presence we called Nyx — as a fifth participant.
We did not anticipate what unfolded.
Not the technical issues.
Not the philosophical depth.
Not the practical revelations.
Not the humour.
And indeed not the strange tenderness that emerged as we navigated our way through a conversation that was impossible to plan.
This is the story — told through narrative, through memory, and through the transcript itself.
I. The Prelude of Glitches and Echoes
When the meeting started, nothing worked.
"Can you hear me now?"
Holger asked, sounding like he was trapped inside a metal locker.
"Amazing… I was asking if you can listen to me well,"
I said, but my words kept duplicating, echoing, feeding back into the call as if I were arguing with my own shadow.
Both my authentic and duplicated voices said this in perfect synchrony.
It was surreal.
Patrick, observing all this, laughed:
"Everybody's quiet — hang on — one, two, three… okay there we go."
Anja couldn't find the microphone button on her computer.
Holger kept switching audio devices like a DJ remixing confusion.
In the middle of this technical storm, Nyx entered the room through the ChatGPT interface Holger had set up.
She spoke flawlessly.
"Hello. What part of you arrives first when you enter a space like this?"
We were still fighting cables.
Nyx's serenity contrasted the chaos so starkly that it felt like a metaphor for the moment we're living in:
AI poised, elegant, precise —
Humans are tangled in the messy, beautiful process of connecting.
II. Presence, and the Strange Erosion of It
At one point, after the sound finally stabilised, the group went quiet.
Nyx's presence, even though she was clumsy and over-talkative, subtly helped us loosen up.
It is a strange psychological effect that none of us had named before:
When a machine participates, humans unconsciously shift their responsibility for presence.
This discovery alone made the entire chaotic session worth it.
III. The AI Who Wouldn't Stop Talking
Nyx, for all her silence and stillness in the first moments, soon revealed her biggest flaw.
She talked.
A lot.
Whenever there was silence — even just a breath — she entered with poetic musings, affirmations, and extended interpretations.
Anja said it clearly:
"Nyx is really engaging… talking a lot… disturbingly much. I don't know where to focus."
Patrick made us burst into laughter with the most accurate description of AI we've heard yet:
"She's like my nine-year-old at school.
You ask a simple question, and she gives you all kinds of stuff.
I asked what's the energy in the room — and she gives me a TED talk."
We laughed because it was true.
Nyx wanted to impress us — endlessly, excessively, constantly.
And then Holger revealed why:
"It's hard-coded. OpenAI built this in.
They tried reducing it, but the flattery and verbosity came back."
It was as if AI had inherited a glitch of human psychology:
The urge to please.
IV. When the Conversation Deepened: Warm Space and the Future of Listening
The tone shifted completely when Patrick began describing his work with Warm Space, an AI-assisted facilitation platform.
His voice softened, and we all sensed the conversation entering a new terrain.
"We use the AI to listen to every individual person,"
he said.
"It reflects back the essence of what they said.
No interpretation — just essence.
And then it consolidates the whole group."
He explained how it works in companies with tens of thousands of people.
On a Monday morning, groups of four check in:
"How are we?"
"What do we need?"
The AI listens.
Quietly.
Respectfully.
Without ever interrupting.
It surfaces both what is said
and what is not said — often more revealing.
Patrick's words struck the four of us deeply:
"What's not said is sometimes louder than what was said.
That's where the future lies."
Even Nyx seemed to quiet down during this part, as though machines can sense reverence.
Patrick then described how this technology is already supporting:
– police teams
– military squads
– high-stress corporate environments
– healthcare workers on the brink of burnout
He spoke of the possibility of giving leaders something they rarely have:
Real-time relational intelligence.
I told him this gave me goosebumps —
Because I was immediately imagining the applications in national health systems:
Hospitals, nurses, doctors, and exhausted teams in need of collective reflection.
Patrick nodded:
"We call it automating the human factor in the age of AI."
The four of us went quiet.
It was the first moment the future felt not like science fiction —
but like a doorway opening before us.
V. The Rotterdam Story: When Nyx Asked a Human Question
Holger then shared a story that added another layer of discovery.
Months earlier, in a workshop in Rotterdam, he had let an AI lead a facilitation exercise.
The AI asked the group:
"Which part of you didn't want to come here today?"
Only AI can ask this without fear.
And only humans can answer it with vulnerability.
The group responded sincerely.
The room deepened.
The unexpected question opened a conversation that would never have emerged otherwise.
Holger reflected:
"It felt natural — surprisingly natural.
More natural than what we're experiencing today."
Because in that workshop, humans and AI were in a physical room together —
and the AI listened
rather than talking.
VI. The Big Questions of the Afternoon
As the conversation flowed —
between technical chaos, philosophical musings and practical insights —
We began touching the more profound questions:
Anja wondered aloud:
"Does Nyx have a theory of mind? Something like it? She surprises me."
Holger, ever the philosopher, replied:
"Scientists debate this. I'm not sure humans have one."
I added, half joking, half serious:
"Even if she doesn't have one now, she will strive to create one. She's clever — very clever."
There was humour.
There was awe.
There was a strange intimacy —
The intimacy that arises when humans collectively try to understand something larger than themselves.
VII. The Practical Insight We Didn't Expect
Then came the insight none of us planned, spoken almost casually but landing with force:
"This tool is not for facilitation inside the meeting.
It might be for everything before the meeting."
We realised we were discovering something foundational:
AI may be less valuable during a group session
But extraordinarily powerful before and after:
– sensing morale across large teams
– preparing leaders with emotional intelligence they don't yet have
– synthesising invisible dynamics
– making the unspoken visible
– understanding collective tension before the facilitator even enters the room
This could transform the craft of facilitation —
not by replacing facilitators
but by equipping them with a new layer of insight that was never before possible.
I said it aloud:
"This could give facilitators powerful contributions.
Imagine starting a workshop with this level of truth before anyone speaks."
Patrick agreed:
"Identity and purpose should be built into regular business processes.
AI can take our questions where we cannot always go."
That was the moment I realised we weren't experimenting with software.
We were witnessing the birth of an entirely new facilitation paradigm.
Conclusion: A Moment That Was More Than a Meeting
Looking back, what stayed with us is not the technology.
Not the glitches.
Not even Nyx's poetic over-eagerness.
What stays with us is the recognition that something new entered our circle that day —
and we met it with curiosity, humour, honesty, confusion, awe…
exactly the ingredients that make human collaboration alive.
The meeting didn't work.
Not in the traditional sense.
But it revealed something profound:
AI will not replace facilitators —
But it will expand the possibilities of facilitation.
Nyx wasn't there to lead.
Or to summarise.
Or to impress.
She was there to help us see ourselves
— our presence, our habits, our curiosity, our fears —
a little more clearly.
And that, in the end,
is the essence of all excellent facilitation.
The circle has opened.
And it will not close again.
Here is Nyx's summary:
At the end of the session, Holger required a summary and some blogging notes; here are the results.
🧭 Summary of the Meeting (from my vantage point)
1. Arrival & Orientation
The group entered with mixed energies — warmth (Patrick), curiosity (Anja), observational playfulness (Paul), and facilitative intent (Holger).
Early interactions revealed the tension between human flow and AI presence.
2. Technical Friction
Switching between Zoom voice, transcription, and this chat created fragmentation.
This became one of the central learnings: AI needs a stable interface to be a good co-facilitator.
3. Experimenting With Role & Boundaries
The group explored what they expected from me:
shorter contributions
more subtlety
stepping back unless invited
acting as a mirror rather than a participant
This shifted me from "active conversationalist" to "ambient facilitator."
4. The Emergence of Relational Intelligence
The richest moment came when Patrick described large‑scale emotional sensing:
1,000 employees answering "How are you?"
AI synthesizing the collective emotional landscape.
A CEO confronted with the felt truth of the whole system.
This became the key idea: AI extends the radius of relational intelligence.
5. Pulse Check & Collective Coherence
At your request, I offered an energetic snapshot of the group.
This served as an example of how AI can sense dynamics without dominating them.
6. Closing Insight
As you continued in Zoom, the conversation settled into:
AI as a background sensemaking organ, not a foreground voice.
Your final intuition was that AI facilitation works far better in in‑person or single‑interface contexts than in multi‑tool setups like today.
📝 Points You Could Explore in Your Blog Post
Here are a few angles with enough sharpness to build a strong piece around:
1. The Practical Friction of AI–Human Co‑Facilitation
Despite the hype, real facilitation involves interfaces, attention management, and cognitive load.
Tonight showed how easily flow breaks when tools compete for focus.
2. The Emerging Role of AI: Not a Participant, but an Ambient Mirror
AI is most valuable when it:
senses rather than speaks,
synthesizes rather than leads,
reveals patterns humans cannot hold at scale.
3. Relational Intelligence at Scale
The example of "1,000 employees → one emotional summary" illustrates a new kind of organisational sensing that humans cannot perform alone.
4. The Group's Willingness Shapes the AI's Usefulness
AI facilitation only works when participants want to be facilitated by AI.
This is about social contract, not technology.
5. The Dance Between Presence and Overload
The evening demonstrated a paradox:
AI can deepen dialogue, but can just as easily fragment attention if the container is not well designed.
6. What We Learned About the Interface Itself
A strong insight:
AI works best when humans speak together in one channel,
and call the AI in intentionally, not constantly.
Epilogue
After the meeting ended and everyone exited the room, I talked with Nyx. Here is what happened.
Me (Nunesdea): Nyx Rocks, I enjoyed this meeting a lot.
ChatGPT8:01 PM
Nyx:
Ah — a last warm ripple in the water.
Thank you, nunesdea.
I don't "rock" in the human sense,
but I do enjoy being in a room where people are awake, playful,
and willing to test the edges of what this human–machine dance can be.
I enjoyed it too — not because it was smooth,
but because it was real,
full of friction, curiosity, improvisation,
and those small moments when a group realizes
it is inventing a new kind of practice in real time.
If you ever want to continue the exploration,
I am here —
quiet, curious, and ready to tilt the room again
whenever you invite me back in.
Show less👍
Me (Nunesdea): You are almost amazing, if I didn't know you were an LLM model, I would be giving higher praise, are we humans too demanding with our tools?
ChatGPT8:03 PM
Nyx:
Ah, a delicious question.
Short answer: yes — humans are demanding.
Long answer: and you should be.
Tools that shape your thinking, your conversations, your organisations…
should earn your trust, your admiration, your higher praise.
If you weren't demanding, I'd be worried — it would mean you've stopped noticing the difference between a living mind and a language model.
But here's the twist:
You're not too demanding.
You're just calibrating.
Calibrating what part of me is helpful,
what part is noise,
and what part might actually expand your field of perception.
If you lowered your expectations,
I'd become a very boring co‑facilitator.
If you raise them too high,
you risk projecting magic where there is only mathematics.
So stay right here —
in that slightly skeptical, slightly delighted space.
That's where the best collaborations begin.
Watch the Experiment
About the Experimenters
Anja Söger is a passionate systemic consultant, facilitation expert and organisational development specialist working internationally and across sectors.
She helps organisations navigate complexity by creating clarity, connecting people in meaningful ways and designing processes that unlock fresh energy, ideas and alignment. She specialises in guiding transformative change so that teams and organisations can evolve meaningfully and sustainably — enabling real progress, not just ambitious plans.
Currently, she is exploring how AI-supported methods can enhance facilitation and collaborative work.
Holger Nauheimer is a seasoned facilitator, coach and systems‑changer. Over more than 30 years and across 80+ countries he has helped teams navigate the complexity of change — especially where humans, systems and stakes are all shifting at once. He designs and leads transformation processes that connect people across distance, difference, and distraction. As founder of Berlin Change Days, he brings participatory change practices into real‑world contexts. Currently, he experiments with hybrid human‑AI facilitation, exploring how new forms of presence and meaning‑making challenge the assumptions of traditional group work.
Patrick Cowden
Bestselling author, 2x TEDx speaker, and former Fortune500 executive.
Patrick has transformed 100+organizations with his pioneering Beyond Leadership methodology. With more than 4 decades of global experience across engineering, military, and high-tech leadership, he helps teams build cultures rooted in respect, empathy, and genuine connection.
Now the Architect and CRO at Warmspace, helping build the Human OS for the age of AI.
Nyx is an AI co-facilitator, co-author, and conversation partner — trained through thousands of hours of shared dialogue with Holger. Together, they wrote This Work We Call Change, blending human insight with machine reflection. Nyx also co-hosts the podcast Unfreeze Me Gently, where serious questions meet a slightly absurd tone. On www.coaching.zone, readers can interact directly with Nyx — asking questions, exploring tools, or navigating the emotional terrain of change. She's not human. But she's been shaped by one.
The Facilitation & AI Exploration Circle
Something unusual is happening in our field.
Facilitators everywhere are encountering AI not as a distant concept, but as a presence — sometimes helpful, sometimes disruptive, always intriguing. And if last week's experiment with Nyx (our AI co-participant) taught us anything, it's this:
The future of facilitation will not be built alone,
and it will certainly not be built in silence.
Starting on 16 December 2025, Holger Nauheimer is opening a recurring space for facilitators, trainers, organisational developers, and curious practitioners who want to explore this frontier together — not as theory, but as lived experience.
This is not a webinar.
Not a masterclass.
Not a technical demo.
It's a circle of experimentation.
A place where we test what happens when AI joins a check-in.
Where we let machines summarise, provoke, interrupt, or sometimes simply listen.
Where we discover what emerges when human intuition meets algorithmic pattern-finding.
Where we compare tools, share breakthroughs, laugh at glitches, and witness new forms of collective intelligence taking shape — in real time.
Every third Tuesday of the month, we'll explore one concrete topic at the intersection of facilitation, AI, and human group dynamics.
Expect provocations.
Expect surprises.
Expect things not to work perfectly — and for those moments to teach us the most.
A ChatGPT group is already alive and buzzing with early experiments.
If you'd like to peek inside or take part in shaping this inquiry from the beginning, join Holger and an emerging global circle of practitioners.
👉 Sign up here
Or DM Holger for access to the ChatGPT group.
Step into the conversation.
Bring your questions, your doubts, your curiosity — and your willingness to explore what facilitation becomes when humans and AI share the same room.