Why Does the Day Feel Lost Before It Starts? | AI for Executives

Send us Fan Mail

I am Sahar Andrade, MB.BCh, a Neuroleadership Coach and Forbes Coaches Council member, and here is the direct answer. 

The day feels lost before it starts because the day starts earlier than the calendar says. Leaders process threats, rehearse rooms, and clear small choices in private, then walk into the first meeting with less capacity than the meeting needs. 

The Deloitte 2026 Global Human Capital Trends report found 60% of executives now regularly use AI to support decisions, and only 5% say they manage it well. Decision quality tends to fall as decision volume climbs. The work starts before the calendar, not inside it.

What does this episode cover?
The private hour before the first meeting, and why it is the most expensive hour on an executive calendar. The four artifacts that now arrive in that hour that did not exist three years ago. What the research actually supports about decision fatigue, including the replication failures that killed the popular version of it. What high AI oversight costs in measured terms. And four moves that return decision capacity to a day that has none.

Why does AI adoption add decision load instead of removing it?

Because output has to be evaluated, and evaluation is a decision. A human draft comes with years of calibration on the human who wrote it. An AI draft comes with none, so it gets evaluated from scratch every time. The BCG Henderson Institute study published in Harvard Business Review, March 2026, surveying 1,488 full-time US workers, found that workers with heavy AI oversight reported 19% more information overload, 12% more mental fatigue, and 14% more mental effort than workers with light oversight. The researchers named the state AI brain fry.

Is decision fatigue real science?

The mechanism most people were taught is not. The willpower-as-fuel model from the late 1990s failed a 23-laboratory preregistered replication in 2016 with over 2,000 participants, and a second multi-lab attempt in 2019 with more than 3,000. What survived is the part that matters for leaders: decision quality degrades as decision volume climbs, and expectation shapes the curve. This episode teaches the surviving version and names the debate rather than hiding it.

What is the one move to make this week?

Name tomorrow's highest-consequence decision tonight, and give it the front of the day before absorption starts.

Next room: The Pressure Room for Executives, Session 2, Thursday, August 27, 2026, 11 AM Pacific. Forty-five minutes. Free. Register at 

https://www.saharandrade.com/pressure-room

The perspectives shared here reflect the professional experience and opinion of Sahar Andrade, developed over years of coaching leaders. They are offered for educational purposes and as viewpoints to consider, not as facts or clinical findings. Take what serves you. Leave what does not.

Sahar Consulting LLC provides educational leadership development services. Our programs, frameworks, assessments, and content, including 4 Ka Leaders™ and B.R.A.I.N.™, are not medical, psychological, or mental health services, and they do not diagnose, treat, or cure any condition. References to the nervous system describe general leadership patterns, not clinical findings about any individual. Nothing here creates a physician-patient or therapist-client relationship. For health or mental health concerns, consult a licensed professional. In a crisis, contact a licensed mental health professional or your local emergency services.

#AIForExecutives #ExecutiveLeadership #NeuroscienceLeadership #AIStrategy #Neuroleadership #AITransformation #RegulatedLeadership #AIIntegration #ExecutiveDecisionMaking #LeadershipUnderPressure  #ExecutiveStress #NervousSystemRegulation #NeuroscienceLeadership #DecisionMaking #ExecutiveCoaching #Neuroleadership #AIForExecutives #LeadershipBurnout #PrefrontalCortex #AICafeConversations  #AIdecisionfatigue #executivedecisionmaking #neuroscienceofdecisions #leadershipunderpressure #cognitiveoverloadatwork #AIcognitiveload #executiveburnoutneuroscience   #AIBrainFry #AIExhaustion #NeuroscienceLeadership #ExecutiveBurnout #RegulatedLeadership #CognitiveOverload #AILeadership #NeuroscienceOfLeadership #LeadershipBurnout #AIForExecutives  #AIAdoption  #AIForExecutives #AIWorkplace #LeadershipDevelopment #FutureOfWork   #ExecutiveLeadership #AITransformation #AIStrategy  #HumanCenteredAI   #NotechRequired #humancenteredleadership    #QuietBurnout #regulatetolead #RegulationFirstLeadership #NervousSystemLeadership   #BrainBasedLeadership   #NeuroleadershipCA   #NervousSystemAtWork  #CoRegulationAtWork   #Polyvaga

Episode Transcript

SPEAKER_00 (00:13):
You checked the numbers three times last night.
Not because they changed,because you needed to be sure
you had not.
Every leader has a version ofthe rule they rehearse alone.
The door still closed, thecoffee going cold.
The first real decision of theday already made before anyone

(00:34):
watched you make it.
Here is what nobody says outloud.
The hardest part of the meetinghappens before you open that
door.
And but by the time most leaderssit down, the capacity the room
needs from them is alreadyspent.
Why does the day feel lostbefore it starts?

(00:54):
That's today's conversation.
Welcome to the AI CafeConversations Podcast.
I'm Sahar Andradi.
This is where the neuroscienceof pressure meets the future of
leadership.
Before we go anywhere, the thingI say every week, everything I
share comes from my experienceand my opinion as a coach.
It's not a diagnosis and it'snot a fact about you.

(01:17):
You decide what fits now.
Let me give you the answer upfront because I don't think you
have the patience today for aslow reveal and I don't blame
you.
The day feels lost before itstarts because the day starts
earlier than the calendar saysit does.
Not at nine, not at the firstmeeting.

(01:38):
It starts in the car.
Actually, before that, it startsin the shower.
Actually, before that, it startsat 5:40 in the morning when the
mind opens the same file itclosed at midnight and runs it
again.
By the time a leader walks intothe first meeting, dozens of

(01:58):
decisions have already beenmade.
Small ones, invisible ones, noneof them on an agenda.
And every one of them spentsomething.
That is the whole episode inthree sentences.
Now let me show you themachinery underneath that.
Because once you see it, youstop reaching for the wrong

(02:18):
explanation and start treatingit as a volume problem you
actually have power over.
Let me describe a morning.
Not one person's morning, thecomposite I hear over and over
and over from executives.
And if you have sat in thatchair, you will recognize the
shape of it.
The alarm goes off, and thereview has already started.

(02:43):
Not the meeting review, theself-review.
What I said Tuesday, how itlanded, whether the silence
after meant agreement orsomething else.
Then the phone.
Three messages that each containa decision delivered as a
question.
Can we push the lounge?
Do you want me to tell them ordo you want to?

(03:05):
Are we still doing the offsite?
Then come the numbers.
Opened again.
Not because anything moved,because certainty needs to be
manufactured before it can beperformed.
By 8.15, the leader hasprocessed a threat to the
business.
Rehearsed the room, absorbed forrequests, and cleared maybe six

(03:28):
small choices.
And the first meeting has noteven started.
Here is what I want you tonotice about that list.
Almost none of it is leadership.
I'm sure all of you, includingmyself, went through some form
of it at a certain time.
What is happening here isabsorption.

(03:50):
Somewhere along the way, therole became a place where other
people unfinished thinking getsdeposited, and the leader became
the processor of last result.
Leaders describe this to me inalmost the same words every
time.
Not I am stressed, somethingmore specific.

(04:10):
By the time I get to the actualdecision, there is nothing left
to bring to it.
And here is the part that landshardest.
They read that sentence asevidence about themselves.
It's evidence about volume.
This is not a disciplineproblem, it's a load problem.
And in 2026, the load has a newengine bottled into it.

(04:34):
Let me talk about AI becausethis is where the conversation
gets honest.
Deloitte's 2026 Global HumanCapital Trends Survey, built
with Oxford Economics acrossmore than 9,000 businesses and
HR leaders in 89 countries,found that 60% of executives now

(04:56):
regularly use AI to support thedecisions.
60%.
Two years ago that number wouldhave been around an error.
Now hold that next to what thetools were sold as.

(05:28):
They measured mental effort,mental fatigue, information
overload, error rates, andintent to quit against how
people actually used AI.
Not how they said they feltabout it, how they used it.
What they found, the peoplecarrying heavy oversight of AI

(05:53):
tools reported 14% more mentaleffort, 12% more mental fatigue,
and 19% more informationoverload than people carrying
light oversight.
19% more information overloadfrom the tools that were

(06:14):
supposed to reduce.
The researchers gave the state aname.
They call it AI brain fry.
Their definition is mentalfatigue for use or oversight of
AI beyond a person's cognitivecapacity.
The people living it describedas buzzing feeling, mental fog,

(06:39):
slower decisions, headaches.
And this is the line I would endthe line twice if there was a
document instead of aconversation.
14% of AI users in that studyreported it.
In marketing segments, 26%.
And the people reporting it alsoreported 33% more decision

(07:02):
fatigue and 39% more majorerrors.
So follow the sequence here.
The tool arrives to remove work,the tool produces output.
The output has to be checked.
Checking is a decision.
Every check is a decision.
And now there are 11 of thembefore the first meeting instead

(07:27):
of six.
Let me make that concretebecause abstractions do not
change behavior.
A leader opens the laptop at7.20.
There is a summary ofyesterday's customer calls
generated overnight.
There is a draft of thereorganization memo generated by
someone else's tool.

(07:48):
There is a forecast with aconfidence range attached to it.
And the range is wide enoughthat the range itself is a
decision.
There is a flag compliance itemthat may or may not be a real
flag.
Four artifacts.
None of them existed three yearsago.

(08:11):
Every one of them arrives withthe same silent question
attached.
Do you trust this?
And how much?
The question is the load, notthe reading, the trusting.
Because here is what nobodyaccounted for when these tools

(08:32):
rolled in.
A human draft comes with a humanyou can't calibrate.
You know that this personoverstates and that person
buries the bad news in paragraphfour.
You have 20 years of calibrationdata on them.
An AI draft comes with none ofthat.
So every artifact gets evaluatedfrom scratch every morning.

(08:57):
And the evaluation happens inthe same private hour or the
rehearsal already lives.
That is the AI integrationproblem in most organizations
right now.
And it's not technical.
Judgment forms slower thaninformation arrives.
The tooling is fine.
The capacity to absorb what thetooling produces was never

(09:20):
funded.
The load did not get smaller, itgot faster and it got quieter.
It moved from things you canpoint to on a calendar into a
category nobody schedules andnobody counts.
Which brings me to themechanism.
And I'm going to be careful herebecause I would rather you trust

(09:41):
me than be impressed by me.
The mechanism is decisionfatigue.
And the popular version of it iswrong.
The story most people have heardis that willpower is like fuel
in a tank, that it drainsthrough the day, that glucose is

(10:02):
the fuel.
The version came out of aresearch in the late 90s and it
dominated for two decades.
Then it was tested properly.
In 2016, 23 laboratories ran apre-registered replication for
over 2,000 participants.

(10:24):
The effect came backstatistically undistinguishable
from zero.
A second large multi-lab attemptin 2019.
More than 3,000 participantsfound the same nothing.

(10:44):
So the fuel tank is not themechanism.
In my opinion, a physician whokeeps repeating a debunked model
because it makes a nice metaphoris not educating anyone.
Here is what survived thewreckage, and it's the part that
matters for the person listeningto this.
Decision quality degrades asdecision volume climbs.

(11:10):
Not because a tank empties,because switching between
decisions has a cost.
And that cost accumulates.

(11:36):
I'm going to say that again.
Decision quality degrades asdecision volume climbs.
Not because a tank empties, butbecause switching between
decisions has a cost, and thatcost accumulates.
The accumulation shows up as anarrower thinker, fasting or

(11:58):
faster shortcuts.
And a strong preference forwhatever option requires the
least from you next.
There is also a finding out ofCarol Dweck's lab that I think
about constantly.
People who believe mentalcapacity is strictly limited
show the classic decline afterdemanding work.

(12:21):
People who do not hold thatbelief show much less of it.
Expectation shapes the curve.
That is not a permission slip togrind.
It's a warning about the storywe tell ourselves at 8:15 in the
morning.
If a leader has already decidedthe day is spent, guess what

(12:44):
will happen?
The day will behave accordingly.
Now let me be specific aboutwhat this costs, because in my
experience, the cost is neverwhere leaders think it is.
They think the cost isexhaustion.
Exhaustion is the receipt, notthe purchase.
The first cost is order.

(13:05):
The most consequential decisionof the day almost never sits in
the first part.
It sits at 2 in the afternoon orin the last 10 minutes of a
meeting that ran long or in ahallway.
So the decisions that shape aquarter get made with whatever
is left after the decisions thatshape nothing.

(13:29):
Nobody designed that.
Everybody lives it.
The second cost is the smallchoices.
This is the tell I listen for,and it's almost never mentioned
in a court in a coachingconversation until I ask
directly.
Decision fatigue does notannounce itself in the
boardroom.
It shows up when choosing lunchfeels like too much.

(13:52):
When picking a restaurantbecomes an imposition.
When someone asks what movie,and the honest answer is, I
cannot do one more of thesetoday.
Leaders hear that in themselvesand reach for a word like
discipline.
It's arithmetic showing up inthe least important place first

(14:17):
because the important placesstill have adrenaline propping
them up.
The third cost is the room.
And this is the one with theprice tag on it.
Under high load, the range ofoptions a person can hold
narrows.

(14:38):
Fewer alternatives getconsidered.
Descent starts to feel likedelay instead of data.
Volume narrows range.
That is the whole story, andvolume was never budgeted.
So go back to that BCG number.
39% more major errors amongpeople in that state.

(15:03):
That is not a wellnessstatistic.
That is a business statisticwearing a wellness costume.
The fourth cost is the oneleaders pay in private.
And it's the reason I built thework I built.
There is no peer.
Advisors exist, coaches exist,boards exist, but almost nobody

(15:30):
sits where a given leader sitswith the full contacts at the
same altitude on the same day.
So the highest takes thinkinggets done with no one to test it
against.

(15:57):
That is expensive in a way nodashboard captures.
And the fifth cost lands on theteam, whether anyone intends it
or not.
Teams do not read strategy decksfor signal, they read faces.
So when a leader walks inalready spent, the team receives

(16:20):
a version of that leader that isshorter, flatter, and quicker to
close conversations.
The team does not conclude theboss is tired, the team
concludes something is wrong andnobody is saying it.
Then they do their own math.
On the same BCG data, theworkers reporting that

(16:44):
overloaded state also reported39% higher intent to quit.
So the sequence completesitself.
Load arrives before the day.
Load spends the leader.
The spent leader leads anarrower room.
The room reads the narrowing andstarts protecting itself.

(17:08):
And six months later, somebodycalls it an engagement problem
and buys a survey.
Though it's just a toxic cycle,like I just read it.
Lord arrives before the day,Lord spans the leader, the spent
leader leads a narrow room.
The room reads the narrowing andstarts protecting itself.

(17:30):
And then they call it engagementin six months.
And I want to sit in that wordfor a second because it's the
most useful word in thisepisode.
Math is checkable.
Math can be shown to someone.
Math is not an accusation aboutanyone's character.

(17:53):
But almost nothing in thissystem is built to display it.
There is no line item on anydashboard for decisions
absorbed.
No column for artifactsevaluated.
No field where a leader canenter the number of times they
were asked to convert someoneelse's uncertainty into a yes or

(18:17):
a no before 9 in the morning.
So the load stays invisible, andinvisible loads get attributed
to people.
The leader looks at the team andthinks, they do not see how much
lands on me.
The team looks up and thinks,management does not understand
what we are carrying.

(18:38):
And by the way, I hear it allthe time.
Both readings are accurate.
Both are reading the samemissing number from opposite
ends.
And because the number doesn'texist anywhere, each slide fills
the gap with a story about theother side's comment.

(19:03):
And commitment.
That is how two groups of peoplewho are both strained end up
describing each other as theproblem.
I have watched that dynamic runinside organizations that had on
paper everything they needed tosucceed.
Good people, real strategy,funded plan, and a load nobody

(19:28):
was measuring, quietlyconverting competence into
friction.
So what actually changes this?
Or what can we do about this?
It's not more discipline, not anearlier alarm.
Please, definitely not anearlier alarm.
What changes is treatingdecision volume as a real

(19:51):
budget, the way a leader alreadybudgets money and headcount.
You would never let 11 peoplespend from one account with no
cap and no reporting.
That is exactly the mostexecutive warnings of.
Here is the direction, and I'mgiving you the thinking rather
than the full protocol becausethe protocol is what we do live

(20:13):
in the room.
First, the first two hoursbelong to one decision, and you
name it from the night before.
One decision, not a list, one,the highest consequence decision
available to you tomorrow.
It gets the front of the daybefore the absorber starts, and

(20:35):
it gets made when the arithmeticis still in your favor.
Everything else in that windowwaits.
Not forever until 11.
Second, convert repeateddecisions into rules.
You know how we do now with AI,like with Claude, if I do

(20:56):
something repetitively, I justcreate a skill so I don't have
to repeat myself anymore.
Same principle.
Every decision that shows upmore than twice a month is not a
decision.
It's a policy you have notwritten yet.
Travel approvals under a certainnumber, which meetings you

(21:17):
decide by default.
Who gets an answer today and whogets an answer Thursday?
Every rule you write removes anitem from the account
permanently.
This is the single biggest winavailable to most executives,
and almost nobody does it.
Because writing a rule feelslike less work than making a

(21:40):
decision, and it's not.
It's the same work done once.
Three examples so it does notstay theoretical.
One any spend under a certainset number does not come to you.
Somebody else owns it, and theyown the outcome to.

(22:00):
You will lose a small amount ofcontrol and recover an hour of
week in thinking.
Two, any request that arrivesafter 4 in the afternoon gets
answered tomorrow.
Unless a customer is activelyaffected.
Not because the request does notmatter, because the version of

(22:20):
you available at 4.30 is not theversion that should be answering
that request.
3.
Any AI output that touches anexternal audience gets one
review by one named person atone schedule time.
Not for people glancing at itacross nine hours in a thread.

(22:41):
One pass, one owner, done.
Every rule like that is adecision you will never make
again.
And decisions you never makeagain are the only kind that
stay removed.
Third, and this is the AI piece.
And it's the opposite of whatmost people expect me to say.

(23:06):
I'm not going to tell you to usefewer tools.
I'm going to tell you changewhat you use them for.
Go back to the BCG findings.
Perceived productivity climbedfrom one tool to two, kept
climbing at three, and then fellas tools tagged beyond that.
And the group carrying theheaviest cognitive cost was not

(23:29):
the group using AI most.
It was the group supervising AImost.
Read that again.
The cost was the in thewatching, not in the using.
And here is the counterweight inthe same research.
When AI took over generallyroutine work, people reported
less burnout, not more.

(23:50):
Higher engagement, betterconnection with the people
around them.
So the question stops being howmuch AI?
It becomes one question askedper task.
Am I delegating this or am Ibabysitting this?
Delegation returns capacity.
Supervision consumes it.

(24:12):
Most executives, in myexperience, have quietly built a
supervision job on top of thejob they already have.
And nobody in the organizationhas noticed it.
Because it happens in the sameplace, all the other invisible
load happens before the dooropens.

(24:33):
And the gap is not a secret.
In the same Deloitte 2026 GlobalHuman Capital Trends Report, 60%
of executives use AI in decisionmaking, and only 5% say they
manage it well.
60 in five managing it.

(24:54):
This is not a technology gap.
That is an unbudgeted loadshowing up as accountability
problem.
Pick one workflow this weekwhere you are checking output
you do not actually need tocheck.

(25:15):
Stop checking it or stop usingthe tool for it.
Either answer returns somethingto the account.
Fourth, get one person who canhold the contacts.
Not a direct report, not aspouse who inherits the whole
day at seven in the evening.
Someone at a comparable attitudewho can hear the thinking and

(25:35):
push back on it.
When leaders describe the momentthey came back to themselves,
another human being is almostalways in the room.
That pattern is so consistentthat I stopped treating it as a
nice idea and started buildingaround it.
This is the part where I willname the deeper framework once

(25:56):
and then leave it alone.
The four method I take executivethrough is brain.
And it exists to rebuild accessto steady judgment under load
rather than to add anotherpractice on top of an overloaded
day.
That is all I'm going to sayabout it here, because it's not
a free content thing, and Iwould rather underpromise than

(26:20):
perform.
What I will say plainly is thebelief underneath all of that.
Steady before strategy.
Steady before strategy.
Not because steadiness ispleasant, because the strategy
formed inside a spent morning isa worse strategy, and it costs

(26:42):
the same amount to execute agood one.
Now I can hear the objection, Ican hear it almost every room I
stand in, and I would be doingyou a disservice if I pretended
it was not reasonable.
You're thinking, Sahar, I do nothave time to sort the load
before the day starts.
The sorting is the load.

(27:02):
You're asking me to add a stepto a morning that is already the
problem.
Fair.
But here is my answer, and it'snot reassurance.
You are not adding a step.
You are moving one.
The sorting is alreadyhappening.
It's happening at 5.40 in themorning with no structure, no
order, and no cap.
It's happening in the shower andin the car and in the four

(27:24):
minutes before the first ball.
The cost already being paid.
What's missing is any return onit.
10 minutes of deliberate sortingdoes not add 10 minutes to the
load.
It replaces 90 minutes ofunstructured sorting with 10
minutes of structured sorting.

(27:46):
And it does it before the volumeclimbs, which is the only window
where the arithmetic stillfavors you.
And the second half of theanswer is harder.
If it's generally true thatthere is no room anywhere in the
day to name one decision, thatis not time management finding.

(28:08):
That is capacity finding.
And it's information about thedesign of the role, not about
the person filling it, which isa different conversation.
Usually the one nobody above youhas invited.
Here is what I want you to dowith this.
Once a month I open a live room,it's called the pressure room,

(28:29):
and it's 45 minutes workingsession on Zoom where executives
work on leading under pressure.
It's not a webinar with a pitchat the end.
One model taught properly, onething to practice, and a room of
people carrying the same loadwho are willing to say so out

(28:50):
loud.
Session two is Thursday, August27th at 11 a.m.
Pacific.
45 minutes free.
Registered at saharendradi.comslash pressure dash room.
That is saharendradi.com forwardslash pressure dash.
And I'm gonna have it also inhere in the notes of this

(29:13):
podcast.
If you take one thing fromtoday, take this.
The day is not lost before itstarts, it's spent before it
starts.
Those are different problems.
And the second one has a budgetyou can actually see.
Name tomorrow's decisionstonight.
Write one rule this week.
Ask one honest question aboutwhat you are supervising that do

(29:36):
not need to supervise.
Thank you for spending this timewith me.
Leadership does not fail.
Nervous system do.
Before I leave, show me somelove.
Rate, comment, subscribe, orshare our podcast.
Thank you for your support.
You have made us one of the top2% global podcasts.

(29:57):
I really appreciate it.
Till I see you on Friday for theshort podcast of the Forbes like
article.
Peace out.