Why Do Leaders Ask AI to Script the Conversation They Are Afraid to Have? | AI for Executives

Send us Fan Mail

 Sahar Andrade, MB.BCh, Neuroleadership Coach: why do so many executives open ChatGPT before their hardest conversation with a direct report?

 Because the amygdala treats a difficult feedback conversation as a threat, and AI feels like a buffer against that threat. 

The problem is not the tool. It is stopping at the script instead of regulating the body that has to deliver it. 

A grounded leader with imperfect words lands better than a scripted leader who is bracing. 

This episode covers what actually changes the outcome, and why Amy Edmondson's real definition of psychological safety has nothing to do with comfort. 

Senior Leaders& Executives are  using AI to script the conversation they're most afraid to have.

Not a memo. Not a strategy deck. The feedback conversation with the one direct report who says they don't feel safe on the team.

I hear this from executives constantly. They open ChatGPT, describe the situation, ask for the words. Then they rehearse the script alone, before they ever say a syllable to the actual person.

Here's what I tell them. AI can write a perfect paragraph about trust and vulnerability. It cannot walk into that room and steady the body that has to say it out loud.

25 percent of workers say they trust AI more than they trust their manager. That's not a stat about AI. That's a stat about what's been missing from the room.

New episode of AI Café Conversations breaks down why the script was never the problem, and what actually changes the outcome before your next hard conversation.

Register for The Pressure Room. Thursday, July 30, 11am Pacific, on Zoom, free. One model taught in full, one protocol you can use that same day. [https://www.saharandrade.com/pressure-room] 

 #AIForExecutives #ExecutiveLeadership #NeuroscienceLeadership #Neuroleadership #DifficultConversations #PsychologicalSafety #ExecutiveCoaching #AIIntegration #RegulatedLeadership  #NervousSystemRegulation #LeadershipDevelopment #AICafeConversations   #AIStrategy #AITransformation #
#HumanCenteredAI #AILeadership #HumanAdvantage #AIandPeople 
 #AI #Artificialintelligence #executivepresence  #neuroscienceinleadership #Futureofwork  #AIpodcast #    #AInotechrequired 
 #neuroscience   #notechrequired #AItoolsforexecutives #AIandleadership #executivecoachingwithAI #AIcoachingtoolsforexecutives  #AIAdoption #ChangeManagement #LeadershipDevelopment  #ExecutiveCoaching  #AICompliance    #Organizationalchange #AIStrategy  #aistrategyforexecutives #executiveaiinsights  #AIleadershiptransformation    #peoplefirst #Neurosciencebasedleadership  #HumanCapabilities #OrganizationalDysfunction #AIPatterns #ExecutiveTruth  

Why do leaders draft feedback with ChatGPT? 

Is it bad to use AI before a hard conversation? 

Why does a direct report sense when a leader is not present? 

What does Amy Edmondson actually say about psychological safety? 

How do you prepare your body before giving hard feedback?

Support the show

--- 

AI Cafe Conversations: Neuroscience-based AI leadership for executives. Hosted by Sahar (The AI Whisperer) | New episodes Wed & Fri 

🔗 Connect: https://www.linkedin.com/in/saharandradespeaker/

📧 Work with me: [e-mail]

🌐 Website: https://www.saharconsulting.com/

 📧 Instagram: https://www.instagram.com/saharthereinventcoach

Episode Transcript

SPEAKER_00 (00:12):
A VP of Operations told me something last month
that I have not stopped thinkingabout.
Before she gives feedback to theone person on her team who makes
her stomach rot, she opens ChadGPT first.
She types out what happened, sheasked it to write the ones.
Then she sits in her car in theparking garage and reads the

(00:35):
script out loud before she eversays a syllable to the actual
person.
She's not lazy and she's notavoiding her job.
She's doing something her bodydecided to do long before
artificial intelligence everexisted.
Today we are going to talk aboutwhy that matters more than the

(00:58):
software does.
Thank you for joining me today.
This is Samhar Alraji.
I'm a neural leadership coachand a Forbes Coach Council
member.
This is AI Cafe Conversations,and I am here to share with you
the intersection of AIleadership and neuroscience.

(01:18):
So here is the pattern I keepwatching show up in my own rooms
over and over, with executiveswho are good at their jobs and
good people underneath thetitle.
There is one direct report,almost always one.
And the direct report has said,in some words or none at all,

(01:40):
that they do not feel safe onthis team.
The leader hears that anddoesn't know what to do with it.
So they go to AI first.
They describe the situation,sometimes in pain-staking
detail, and they ask for thescript.
Does that sound familiar?
Here is what I want you tonotice before we go any further.

(02:04):
AI can write a beautifulparagraph about vulnerability
and trust.
It can draft feedback thatsounds like it came straight out
of a leadership book.
Every phrase measure, every edgesend it down.
What it cannot do is walk intothat room and study the body

(02:25):
that has to say the words outloud.
Because that part was never alanguage problem to begin with.
25% of workers say they trust AImore than they trust the room
manager.
Sit with that number for asecond because it's not really a
statement about artificialintelligence and how good it has

(02:50):
gotten.
It's a statement about what hasbeen missing from the room long
before the chatbot ever showedup.
Let's slow down a bit and lookat what's actually happening in
a leader's body before a meetinglike this even starts.
Because this is where the realstory lives.

(03:11):
This is not a confidenceproblem, it's a threat detection
problem.
And the mechanism is almostembarrassingly simple once you
see it.
The amygdala, the part of thebrain built to scan constantly
for danger, does not know thedifference between a predator

(03:35):
and a hard conversation with adirect report.
Both get flagged the same way.
Both trigger the same cascadeunderneath the surface.
Heart rate climbing, breathgoing shallow, and the
prefrontal cortex, PFC, the partthat handles careful language

(03:58):
and nuance, and reading theroom, going wider at the exact
moment.
A leader needs it working at youguessed it, at full capacity.
So the leader reaches for abuffer.
Not because they're weak, andnot because they do not care,

(04:21):
because the body is doingexactly what it's built to do,
which is look for something tostand between itself and the
perceived injury.
For a long time, that buffer wasavoidance itself.
The meeting that never gotscheduled.
The feedback that got softenedinto almost nothing or nothing.

(04:47):
Now the buffer has a name and aninterface and it types back.
Here is the part I need you toreally sit with, because it's
the part almost everyone in thisconversation gets wrong,
including some very well-meaningHR departments.
The direct report who says theydo not feel safe is not hiding

(05:09):
something.
And they're not manipulating thesituation to get out of
accountability.
Their body is doing its job too,and its job is reading the room
accurately.
I don't mean to shoot myself onthe foot here, but I'm not
saying that that's gonna be 100%of the cases, but this is the

(05:29):
majority of the cases.
Leaders and the reports arebeating each other constantly,
faster than either one, canconsciously track.
Through tone, through posture,through the dozens of small
signals, a nervous system throwsoff before a single, single word

(05:51):
gets spoken.
Scientists call this mirrorneuron activity the white that
lets one person's internal stateregister in another person's
body almost instantly.
You have felt this yourself.
I know I did multiple times.
You know that within seconds ofwalking into a room, whether the

(06:15):
person across from you isactually present or just going
through the motions.
And you knew it before either ofyou said a word.
When the person deliveringfeedback is running in a state
of strain, walking in alreadybraced, already scripted,

(06:36):
already halfway rehearsing thenext sentence instead of
actually being present with theperson in front of them.
The report on the other side ofthe desk picks that up
immediately.
Not because they're paranoid andnot because they're looking for
a reason to feel prompted,because the signal is real and

(07:00):
the nervous system is reading itcorrectly.
A leader in imbalance reads asunstable.
And the buddy sitting across thedesk registers that with total
accuracy, every single time.
Two people, both out of agrounded state, reading each

(07:24):
other with precision, andneither one able to say so out
loud.
Because saying it out loud isexactly the kind of hard
conversation this whole cycle isbuilt to avoid.
That is the crisis.
Not that leaders are using AI,leaders using AI to solve a

(07:48):
language problem when what theyactually have is a nervous
system problem.
And no script, however, whenwritten, changes what the body
in the room is broadcastingunderneath the word.
Here is what happens when aleader keeps reaching for the
script instead of addressingwhat is actually underneath it.

(08:12):
The words get better, true.
The delivery doesn't.

(08:33):
The buddy in the room says, getthis over with as fast as
possible.
Which one do you think thatperson actually believes by the
time the meeting ends?
Tell me in the comments what doyou think.
This is where I want to bring inAmy Edmundson, the Harvard

(08:53):
researcher who gave the feel,the team, psychological safety
in the first place.
Because I hear leaders miscoather more than almost anyone else
in this place.
And in this space.
Psychological safety was neverabout making people feel
comfortable.
It was never about lowering thebar or protecting people from

(09:16):
hard truth.
Edmundson has spent yearscorrecting exactly that
misunderstanding.
Because a generally safe team isnot a team where nobody gets
challenged.
It's a team where people aresafe enough to take the risk of
being honest.
Safe enough to say the wholething and hear the whole thing

(09:39):
back without the relationshipcollapsing.
Candor is the entire point ofthe framework, not the thing
safety is supposed to shieldpeople from.
Which means an AI perfect scriptcannot manufacture the thing
that was actually missing inthat room.
You can hand the leader the mostbeautifully worded feedback,

(10:02):
conversation ever written.
And if the room itself is notgrounded, if the leader is
gracing instead of present, thewords do close to nothing.
Worse than nothing, actually.
Honestly, because now the gapbetween what was said and what
was felt becomes one more pieceof evidence.

(10:25):
The direct reports nervoussystem quietly files under
cannot trust this person, evenwhen they say the right thing.
And the rest of the team iswatching, whether they mean it
or not.
Every person who was never inthat room still fears the
aftermath of it.

(10:45):
Because this is where contagionstops being a personal cost and
starts being an organizationalone.
One leader running constrained,defaulting to scripted, careful,
performed language in everydifficult moment, teaches an
entire team that hard truths gethandled through performance

(11:09):
instead of presence.
People quietly stop bringinghard things forward.
Not because anyone told them to,because they read the room
correctly over and over.
And the room kept saying it'snot actually safe to be honest
here, no matter how good thevocabulary sounds.

(11:31):
Then the turnover conversationhappens later, and everyone is
confused about why the strongestpeople left first.
In my experience, they areusually the most attuned people
on the team.
They are the first to noticewhen safety is being performed

(11:53):
rather than felt, and the firstto leave once they have noticed
it enough times to stopbelieving it would change.
I want to be clear aboutsomething before we move on.
Because the instant here isgoing to be to hear all of this
and decide the tool itself isthe villain.

(12:19):
It's now.
It's now.
It's deflection.
AI is generally useful fororganizing your thoughts before
a hard conversation, forcatching a blind spot in how you
were about to phrase something.
For helping you find languagewhen you are too activated in
the moment to find your own.

(12:41):
The problem was never that aleader opened Chat GPT before a
hard conversation or Cloud orGemini.
The problem is stopping thereand treating the script as the
whole job instead of the veryfirst and step.
This is what I call the copy andpaste syndrome.

(13:04):
The conversation still has to bedelivered by a body that is
regulated enough for the writtenwords to land the way they were
intended.
So what's the solution, youmight ask, right?
So actually, let's talk aboutwhat will change this.

(13:24):
And it's not a better promise,it's a different order of
operations, and it is simplerthan most leaders expect.
Though simple is not the samething as easy.
Here is the shift I teach in myown groups.

(13:45):
Before you touch the keyboard,before you ask AI for a single
word of that conversation, youdeal with your own body first.
You get yourself grounded beforeyou get the language right.
And that order matters more thanalmost anything else in this

(14:08):
episode.
A grounded leader with clumsy,imperfect words would learn than
a scripted leader with perfecteverything about what they say.
Because the nervous systemsitting across the desk is
reading the leader's statebefore it ever fully processes a

(14:32):
single sentence being said.
Authenticity comes through ornone.
This is the whole premiseunderneath the brain V-R-A-I-N
framework I built.
Five evidence-based principlesthat create the conditions for
lasting leadership change.

(14:54):
I'm not going to walk youthrough the full methodology in
this episode, that work happensinside the room with an actual
client.
But here is the piece I want youto leave with today.
Regulations, which is the R inbrain, comes before strategy
every single time.

(15:14):
And that includes the strategyof what you are about to say to
your hardest directory.
Regulated leadership is not justthe nicer way to run a team
either.
DDI's research, TDI research,has connected it to 2.6 times

(15:36):
multiplier in financialperformance, which tells you
this was never a soft skilldressed up in science language.
It's a business result with amechanism behind it.
Practically, here is what lookslike before your next hard
conversation.
You do not open ChatGPT orClaude or Gemini or whatever you

(16:00):
use first.
You do something physical first.
Something that tells your bodythe threat in front of you is a
conversation and not a predator.
A few slow exhales, each onelonger than the inhale before
it.
Feet flat on the floor, alwaysnoticing the floor is there.

(16:24):
30 seconds of doing nothing atall.
No script, no rehearsal.
Just letting your heart ratecome down out of threat mode and
back into something closer togrounded.
Only after that do you go gethelp with language.

(16:46):
If you still want it, becausenow you are drafting from a
grounded state instead of abraced one.
And whatever comes back willactually sound like you instead
of sounding like a leadershipbook someone else wrote.
And when you finally sit downwith that direct report, you let

(17:07):
go of the script the moment theconversation stops going the way
you rehearsed it.
Because it will stop going thatway almost every time.
Real conversations do not followscripts.
A leader white knuckling theirway through a memorized
paragraph is broadcastingexactly that.

(17:33):
Performance, not presence, whenyou need presence and not
performance.
No matter how good the sentencestructure is, the person across
from you feels the differencebetween a leader who prepared
and a leader who is reciting.
Even when they could not meanwhat tipped them off.

(17:53):
One more piece, because I canalready hear the objection
forming from some of you.
What about the direct reportside of all this?
You cannot regulate someoneelse's nervous system for them.
And you should not even try.
But a leader who walks ingrounded changes, what becomes

(18:16):
possible in that room, becausesteadiness is one of the only
things that reliably travelsbetween two nervous systems in
real time.
You have felt this yourself inrules that have nothing to do
with work.
One calm, present personally ina chaotic room can shift the

(18:36):
entire temperature of it withoutsaying a single word.
That is not a soft, vague idea.
That is contagion working in thedirection you actually want it
to work in for once.
Instead of the direction thatkeeps costing you your best

(18:58):
people.
Now I can already hear againsome of you thinking it.
I don't have time to sit with myfeelings before every hard
conversation.
That mushy stuff, woo-woo stuff,I don't believe in it.
I have a board meeting in 20minutes, a direct report crisis
today, and a hundred slackmessages I have not launched.
I hear that completely, I do.

(19:20):
And here is the mechanism behindmy answer, not just reassurance.
The 30 seconds I'm describing isnot a spot aim and is not
self-indulgence, and it's notwoo-hoo out there.
It's actually based inneuroscience.
It is the fastest route back toa functioning PFC, prefrontal

(19:40):
cortex, which happens to be theexact part of the brain you need
online for that conversation togo anywhere close to one.
Skip it, and you walk in with athreat response quite clear,
running the show underneathevery single word you say, which

(20:01):
means worse judgment in themoment, worse listening, and a
much higher chance aconversation goes sideways and
has to happen again next week.
The leaders who tell me they donot have time for this are
almost without exception thesame leaders redoing the same

(20:23):
difficult conversation two orthree separate times.
Because the first two attemptswere delivered by a nervous
system in training, not by theleader that actually are
underneath it.
30 seconds now is cheaper thanthree attempts later.
That is not philosophy, I'mselling you.

(20:44):
That is math.
If any part of this landed foryou, if you recognize yourself
opening a chat window beforeyour hardest conversations, I
want to invite you somewherethis episode cannot take.
The pressure room.
The pressure room is a free livesession I run once a month.

(21:07):
And the next one is Thursday,July 30th at 11 a.m.
Pacific on Zoom.
45 minutes.
That's it.
45 minutes.
One model taught inform.
One protocol you can always usethat same day before your next
hard conversation.
Register for the pressure room.

(21:28):
The link is in the show notes.
Leadership doesn't fail.
Nervous systems do it.
Before I leave, like every time,show me some love.
I'm gonna put the pressure roomlink down here.
It's 45 minutes, it's a free umattendance.
I hope to see you there.

(21:49):
Please like, comment, subscribe,or rate our podcast.
Thank you for your support,making us one of the top 2%
global podcasts that exist.
I really appreciate it.
Show me some love, like I alwayssay.
Till we meet again on Friday inthe Forbes article like short
edition.
Till then, peace out, come up.