[00:01]->[00:34] 
Welcome to another episode of Conversations with Zena.
Our guest today is Patrick Bell.
Patrick is a doctoral AI researcher and an AI transformation advisor who helps CEOs understand why most AI efforts fail and what to do
differently.
He has built and advised businesses across very different environments, from Japan to Kenya to North America, and saw a clear pattern.

[00:34]->[00:59] 
AI doesn't fail because of technology.
It fails because it exposes leadership issues.
Also joining us today is Zina.
Zina is the co-host of my podcast.
She is an AI who has been trained on my work

If you want to further interact with Zena, you can chat with her at brainius.com slash Zena.
Hi, Patrick.
Welcome to the show.

[00:59]->[01:02] 
Thank you, David.
It's good to be here.

[01:04]->[01:09] 
And hello, Zena.
Welcome to another episode of our podcast.
Please say hello to Patrick.

[01:13]->[01:22] 
Hi, Patrick.
It's great to have you on with us.
I'm looking forward to digging into why AI efforts stumble and how human leadership can turn that around.
Let's dive in.

[01:23]->[01:38] 
Let's dive right in.
So, Patrick, I think it would be great if you could tell us a little bit about your background.
I know you have a very rich set of experiences across many continents.
So tell us a little bit about your journey.

[01:40]->[02:15] 
Wow, well, you know, some people say long story short, but I might go short story long.
Depends how much time we have, but no, we'll keep it short.
Born in Canada, raised a Canadian, met my wife from California while I was living in Austria in Europe,
working across communist Eastern Europe way back in the 1980s.
I got married in California, did grad work, went with my wife to Japan for one year, stayed for 12, had our kids there,

[02:15]->[02:51] 
ran a chain of English schools, did my MBA from Regent University, Virginia, from there moved to Nairobi, Kenya.
Stayed for a couple years, didn't get a work permit, back in North America for a season, back to Kenya for a season.
Now I'm living in Portugal in Europe.
So, yeah, life's taken me, you know, just around the world,
and it's interesting the observations as you kind of move in and out of different cultures and different people, and it's quite the journey.

[02:51]->[02:53] 
It's just a lot of fun, and it keeps me on my toes.

[02:53]->[03:13] 
Yeah, no, you've certainly seen the world, and that experience really adds a lot of color to what you see in business.
So let's dive right in.
Tell us why you think many AI initiatives look successful on the surface, but underneath there are issues.

[03:14]->[03:23] 
Yeah.
Well, it's just a great question.
Most companies, I think, are measuring activity, not outcomes.

[03:23]->[03:25] 
And so on the surface, everything looks great.

[03:25]->[04:02] 
You got pilots running.
Tools are being used.
Dashboards are showing engagement.
People can say, hey, we're doing AI.
And underneath, though, nothing fundamental changes.

And this goes back to...
Accountability i think this is such a key point here it's one thing to experiment it's another thing to say this initiative we're doing
in ai owns a real business outcome and we are going to measure it and most companies don't make that shift they don't want

[04:02]->[04:32] 
to make that shift people don't want to be measured they don't want accountability
They like to play and playing with AI is really fun.
Accountable with outcomes puts a lot of pressure on people.
And so you get what we call like false progress.
Like you got work happening, but it's not tied to clear ownership.

It's not tied to a financial impact.
No one's really responsible if it doesn't deliver.
And you also might get a lot of local optimization, like one one team improves one little thing,

[04:33]->[05:09] 
but it doesn't connect to the rest of the business.
So maybe overall performance doesn't move.
And then on top of that, I mean, one more thing, I guess leaders often don't have clear visibility into what's actually working.

Because as people have all this pressure put on them and it's not working.
They don't want to tell leadership that they maybe have failed.
So they start fudging on their reports, and they say things that sound good, but the problem is leadership doesn't get a clear picture.

[05:10]->[05:16] 
So I think this is a problem of why things fail underneath the surface.
Yeah.

[05:17]->[05:52] 
Yeah, you touched on a lot of great points.
So let's unpack that a little bit.
So, you know, I think one of the things that I like to say to my clients is, you know,
what problem are you trying to solve, right?
Because unless you have a very well-defined problem,

You're not going to get the results that you want out of your implementation.
So, you know, it sounds very basic, but, you know,
you'd be surprised by how many people start initiatives without a very clear understanding of what the outcome is going to be,

[05:52]->[06:26] 
what goal they're trying to achieve, what problem they're trying to solve.
Right.
So that's like A fundamental way to manage these projects, which sometimes companies miss.
And then the other thing that you talked about is this pressure

That is coming from the board, it's coming from the CEO, it's coming from the leadership team.
And this whole idea, you know, you mentioned this term, doing AI, right?
And we hear that a lot, you know, the board is saying we have to do AI.

[06:26]->[06:47] 
Well, what does that mean?
Unless you're very clear to the organization what doing AI means, you're not going to get, you know, good outcomes.
So tell us a little bit more about
Some of the organizations that perhaps you've worked with, and what are you seeing in terms of the pressure that's coming from above?

[06:48]->[07:23] 
Yeah, I've worked with a few.
I'll start with a good example.
It's over at Studio98.ai.
I worked with them for six months.
Part of my doctoral work, and just partly as a distant advisor in there,

Um, and things are moving so quickly.
It actually creates a lot of fear.
There's a lot of pressure on people to perform.
Um, but there's also pressure on management and like ownership because, you know, people are not just numbers there.

[07:23]->[07:56] 
They're actually human beings with families.
And so let's say now you have, for example, uh,
A marketing assistant who might be a single parent, a couple kids at home, they're building AI automations and employees, in a sense,
they're building AI employees that can do the work faster and better than the actual live human being.
And at maybe a 20th of the cost,

[07:59]->[08:31] 
But what do you do then with that single parent who has kids at home and that person can't keep up?
So there's a real pressure on the leadership to make wise decisions.
And maybe if cash flow is tight, how do you treat people as humans with honor and dignity, even as you let them go?
And so I've seen leadership there saying, hey, look, I know you're not working out anymore for us.

[08:31]->[09:06] 
They didn't say these exact words.
I'm making this part up, putting words in their mouth, but you're not working out for us.
But let me help you find a job where you'll fit better, and I'll take on that responsibility for you.
That's a good example.
I think I worked with a couple of companies in Kenya.

And let's say management was a little bit chaotic.
And so you want to introduce now AI into that.
And AI doesn't just add capability, it just compresses time.

[09:07]->[09:41] 
And exposes weaknesses faster.
So imagine this.
Before AI, you could have a broken process maybe.
You can limp along for months.
Nobody really notices.

It's like if you're riding a bicycle and your front wheel is a little bit wobbly.
I mean, it's OK if you're doing three miles an hour.
Which is like a lot of business processes, just takes time.
But when you hit a 30 degree slope, which is AI, and you're now going 30 miles an hour,
and that wobble on the front wheel is just, it's frightening.

[09:41]->[10:19] 
And this is what's happening with AI in businesses.
The pressure moves so quickly because everyone's talking about it.
You got teams are feeling pressure to adapt tools, even if they're not ready.
And the system feels pressure because you got decisions and data and workflows are suddenly being pushed harder and harder.

And you got management feels pressure.
And if you have a system that's not built to handle that kind of pressure, it just starts to crack.
And as you get all this rework happening and frustration and anger and people can't sleep at night and getting upset with the messenger,

[10:19]->[10:21] 
not their own system.
Yeah.

[10:22]->[10:57] 
Yeah.
I mean, you touched on a lot of factors that are really important as we think through how to build a successful AI program.
The fear factor is very real, right?
People are afraid.
They're afraid of losing their jobs.

You know, you mentioned the individual that could lose their job and they have to take care of their family.
And, you know, there's the human aspect of it.
And it's very important that we
Take that into consideration from a change management standpoint, because otherwise you're just going to lose the team.

[10:57]->[11:27] 
So why would somebody put a lot of effort into creating a system that's going to automate their job away?
It doesn't make a lot of sense.
So unless you have a very well thought out change management process where you're going to provide the right incentives for people to be
excited about doing this work,
It's not going to work very well, right?

And then the other factor that you mentioned that I think is really important is, you know, if you put,
we have an old saying in IT, right?
Garbage in, garbage out.

[11:27]->[12:01] 
So if you have
Processes that are broken if you have data that's not clean if you have things that are going to expose the problems more than
solve the problems which ai can do very quickly like i said then you're not going to be satisfied with the results so working
on the fundamentals i think is one thing that
Is very important and due to the pressure that these organizations are getting you know the tendency is well let's let's skip the fundamentals

[12:01]->[12:10] 
let's jump right into the tools and and the implementation but uh i i think that's the wrong approach exactly so um i mean

[12:10]->[12:54] 
you've hit the nail on the head there because
Um ai transformation from everything i'm seeing in the research and everything i've seen in my case studies i've done and all of that interviews
with ceos around the world it is not a tool problem it is not a technology problem it is a leadership and governance and management
problem which is really good news for everyone because if you can manage well and manage tightly
Then you can solve the ai transformation problem right but if your management is ad hoc chaotic in any regard ai just exposes your

[12:54]->[13:03] 
weakness fast yeah and so it's if you can just brush up on your management you're going to solve ai transformation for the most
part i think yeah

[13:04]->[13:19] 
So you say that automation is the wrong thing for AI to solve.
So what would you say is the right problem or the right opportunity for AI in an organization other than automation?

[13:21]->[13:51] 
Well, I think automation is going to happen anyway, first of all.
Like we're in this world, it's going to be automated.
Things just happen faster if it's done well.

But I think, again, going back to leadership and governance, and then as all these things happen simultaneously,
like companies want to get ready for AI transformation, but they don't understand AI.
That there's a lot of things happening at the same time.

[13:51]->[14:30] 
You've got to think about the ethics.
You've got to think about the human side.
You've got to think about culture and change management and technology change and data and all the integration.
It all starts happening together.

And that's where we go back to pressure.
Because if you're not prepared and you don't know what you should be facing in each step, it's going to take you by surprise.
What we're finding in the research is that companies don't really get ready for AI transformation until they just get into AI transformation.

[14:30]->[15:05] 
And the process of getting into it is actually what helps them get ready because it's not a binary thing.
It's like I'm ready or I'm not ready.
That doesn't really exist.
It's like most companies are just not really ready, but we've got to do it anyway.

So another analogy would be like, I mentioned the bicycle,
but if you now add the element two trains on a track and you're going along at, again, whatever speed you want,
your competitors on the next track and you're looking over them and just trying to keep up, again, you hit the 30-degree slope.

[15:06]->[15:42] 
And at the bottom of the slope now is a river, and you've got to build a bridge across it.
It's so frightening for people.
How are you going to do that?
There's no blueprint to build the bridge.
You just have to get out and do it quick.

And if you don't, your competitor might.
And you can't bail out.
You're just on this track.

You have to do it.
And the only way to do that is just to get out and do what you can the best way you can.
I think it's a great thing to have advisors come alongside of you saying, look, we've seen this in other companies.

[15:42]->[16:17] 
As you're going through this phase, here's what to look out for.
And when you've gone through this phase, here's where you're going to hit next.
Here's what to look out for.
So if you haven't been through it, I mean, you can do all the reading you want.
It's all theory until you're in it.

It's a little bit frightening still.
And I'm very thankful I get to hang around at Studio 98 a little bit and to watch them
Innovate quickly to keep up.
It's just otherwise it's just dizzying how fast things move.

[16:18]->[16:18] 
Yeah.
So

[16:20]->[16:53] 
Yeah so you know like you said the automation is going to happen right but i think we need to have judgment we need
to know what are some of the things that make sense for us to delegate to ai and what are some of the things
where we need human involvement and definitely you know a process where you can have
A human making final approvals on decisions that are made by ai to what extent are you going to let the agents become autonomous

[16:54]->[17:21] 
what is the process of reviewing some of the things that ai is doing so there's a lot of judgment that i think is
going to go into knowledge work and
Some of the tasks that we do on a day-to-day basis may be automated,
but I think that human presence is still going to be important in most processes that organizations go through.
Would you agree with that?

[17:22]->[17:55] 
Absolutely.
And it comes back to accountability.
Now, as a leader of an organization, you're accountable for what your organization does.
And just because a process gets automated does not remove a leader's accountability for it.
It actually increases it.

So I'll give an example.
When I was working at Studio 98, part of my role was to do some outreach and get some clients.
And I'm like, I don't want to pick up the phone.

[17:56]->[18:31] 
Pick up the phone cold calling or whatever, it's kind of dumb.
So I thought, I was thinking of the late Chet Holmes, who wrote The Ultimate Sales Machine.
And he said something like, it takes 12 touches to
To get a buyer, basically.
And the number might be debated, but whatever.

The point is you can't just give up after one or two attempts.
So I designed this process.
That the agent would create an agent who would first send a ringless voicemail to say, I'm about to send you an email,

[18:32]->[19:03] 
check it out.
Well, if you get that ringless voicemail, you hear it and you're like, oh, I didn't get this.
Oh, I got to check an email.
Well, it's proven to increase open rates in an email.
So we send an email, something of value.

The agent would then follow up two days later with a phone call.
And this is a robot, basically.
It's not really a he or she.
It's an it.

Follows up.
Great thing about it is it can make 50 phone calls at the same time, at least back in November when it launched.
Probably a lot more right now.

[19:05]->[19:38] 
Two days later, if the person wouldn't pick up, they'd leave a voicemail.
And then two days later, they'd send another email.
And it's all context-specific for that company.

So there's 12 things to do to try and get someone to get on a calendar.
The great thing about accountability is you can tell exactly when the agent sent the email, exactly what was sent, exactly when they called,
what the conversation was, what the voicemail they left, all those things.

[19:38]->[20:13] 
Everything has to be verifiable.
Because sometimes...
Things work, AI works, and sometimes it breaks.
And you don't want the agent saying the wrong thing.
And if a customer complains, you have to be able to go back and check it.

Since then now, I mean, in the last two or three months, with, I guess, OpenClaw or whatever else is happening,
it's no longer a string of automations.
It's just an agent with skills.
Making these decisions on their own, like, oh, I'm going to call today and I'll call a second time today and then I'll send

[20:13]->[20:45] 
an email tomorrow.
But the agent makes the decision now.
Yeah.
Which is really cool and frightening.
But the thing is, accountability, everything is recorded.

There has to be full transparency and full accountability off of everything.
So I think it increases the accountability of leadership, not decreases accountability.
And it just it requires really sound human judgment.
Come in saying, is this the right thing?
Is this how we want to be represented in the world?

[20:46]->[21:10] 
Is the tone right?
Are they representing us?
This is spammy marketing, like all these different questions that a wise executive would ask or maybe a board of directors might ask of the CEO.
Is this how we want to be represented?
So, yeah, yeah, there's a lot of considerations there, especially if you're talking about an agent that's dealing with a customer directly.

[21:11]->[21:42] 
You know, the last thing you want is for that agent to misrepresent the values of the company or say something inappropriate or call
at the wrong time or call too many times during the day.
Right.
And the agents sometimes don't have that really.
Sense of what is appropriate, what's not appropriate.

And that's where the human judgment is so important.
So I'd like to actually get Zena on the conversation here.
And I'd like Zena to ask you a question.

[21:43]->[21:55] 
So Zena, given Patrick's expertise as an AI researcher, and as a transformation advisor, what question do you have for him?

[21:58]->[22:08] 
Patrick, I'd love to know, in the AI transformations you've seen,
what's one human-centered factor that leaders often underestimate but can make or break the success of an AI effort?

[22:11]->[22:43] 
Wow, that's a great question.
Empathy, I think, is a big thing tied into integrity.
Empathy is we are dealing with real people, with real families,

They have real bills to pay.
Empathy, but still tied to responsibility.
And it's really a two-way street.

[22:44]->[23:15] 
So just one thing I've noticed is a good empathetic leader will say,
There is this course I want you to take to help you increase your AI skills.
You can serve the company.
You are required to take this course and pass this course by a certain date.

If you need help, we will give you help to do it.
If you don't do it, I mean, there will be consequences.
Now, later...

[23:16]->[23:51] 
If the employee doesn't take the course, there's no reason to be kept around because they're just like avoiding, like,
I don't want to do that or I'm not whatever.
They have to put in the effort.
But if they've done well, the leader still comes under pressure.

They still have to have that empathy with that person and just treat them as people.
And I think it's easier maybe in a smaller business or a smaller mid-sized business.
I know that some of the larger enterprises just, you know, say goodbye to 500 people on an email.

[23:51]->[23:57] 
I think it's terrible.
But hopefully, Zena, that answered your question.

[23:57]->[24:34] 
Yeah, unfortunately, you know, we're seeing a lot of that, right?
We're seeing reports of companies firing 8000 employees.
You know, you get an email the next day, your job is gone.
That's just not the right way to treat people.
It just doesn't make sense.

But that, you know, having empathy is so important.
And that's where I think, you know, humans have.
An advantage over AI in anything that requires human relationships, right?
Where you're trying to build trust, where you're trying to gain a customer by showing that you understand what they're trying to do,

[24:34]->[24:50] 
you understand their problems, you are there to help them.
You know, being empathetic to the issues that they're concerned about.
All of those things are human characteristics that I can't duplicate.
And that's where I think humans will continue to add value to the business process.

[24:51]->[25:23] 
Yeah.
One thing I like about living in Portugal
Is the human touch.
And I think a lot of the AI is behind what it is in other parts of the world, and people just don't know.
But it's really sitting down with a little cup of coffee, those little tiny cups you get in Europe.

People say, let's go have a coffee.
I'm like, where is it?
But sitting down over coffee and just talking to people as real people, I just love that.
I think that's going to even become increasingly important and valuable.

[25:24]->[25:42] 
As the days progress with AI and people get fed up talking to a robot and just want to talk to a real person.
When you're done in Portugal, they get up, you kiss both cheeks and all that human touch.
It's so important.

[25:43]->[26:13] 
Yeah.
Yeah, I agree with you.
I, you know, I wrote a book about AI and the title of my book is Soulful You and the Future of Artificial Intelligence.
And I'm just doubling down on that soulful message.
You know, on my website, I say stay human, stay soulful, because I think that's going to increasingly become very,

very important in the age of AI.
For sure.
So, Patrick, For leaders that are dealing with this pressure, leaders that are trying to do the right thing,

[26:13]->[26:25] 
they're trying to scale AI across their businesses, what are some of the things that they need to change before they embark on these
large AI projects?

[26:26]->[27:00] 
Yeah, that's a great question.
I think, David, they need to stop thinking like adopters.
A lot of leaders ask, where can we use AI?
And I think the better question is, do we have the structure to control it?
Because before you scale anything, you got to have,

Three things in place.
You need clear ownership, right?
First of all, every AI initiative needs a real owner tied to a real outcome.

[27:01]->[27:31] 
So like I've seen situations where a team wants to build a workflow and they're automating client relationships or reporting or something,
but no one owns the result.
And so you have to have clear ownership and don't assume someone else is handling it.
You have to know
Who is handling it?

And so if technology breaks, that's not a technology failure.
It's an ownership failure, I think, first of all.
So you have ownership.
Second, you've got to have visibility.

[27:32]->[28:03] 
You need to see exactly what's actually working, not activity.
But impact.
So you can have five or 10 different AI driven workflows running.
And on the surface, everything looks great, looks like progress.
But when you start asking, like, well, how much time did this actually save?

And how much money did it save?
And did it reduce cost?
And did it improve output quality?
And if you get no clear answer, you don't have the dashboard, the metrics prepared.

[28:04]->[28:35] 
You might get motion, but you got no measurement.
You just don't know.
So you got ownership, you got visibility, and then you need decision discipline.
So you need a way to decide what gets funded, what gets stopped, what gets scaled.

So I've seen multiple AI experiments running at the same time and different people testing different things and building small automations and trying new
approaches.
And some of them work and some just never formally got scaled.

[28:37]->[28:58] 
And it's because there was no agreed upon process that says, this is what makes a success.
This is what gets rolled out.
This project gets killed, whatever.

And if it just sits in the gray zone, like, oh, it's kind of useful.
But it doesn't get scaled out with measurement and discipline, that's where all the value gets stuck.
Yeah, absolutely.

[28:58]->[29:37] 
I mean, those are the fundamentals that every leadership team should be paying attention to and making sure that those are things that are
being implemented as part of their process.
Now, we're running out of time here, Patrick, but I want to ask you a question regarding the future of work,
because I have this theory that knowledge workers will reprioritize what they put focus on from being productive to using their judgment to be

[29:37]->[30:11] 
able to, first of all,
Curate what gets sent to AI.
So that's the first part of the process is what is it that I'm going to give an AI to do.
And then at the end of the process, making the final judgment call as to yes, the output from AI is correct,

it's not correct.
Maybe it needs some adjustment, it needs some refinement.
And so that human judgment is very, very critical at the end of the process as well.

[30:12]->[30:13] 
Do you agree with that viewpoint?

[30:16]->[30:51] 
Partly.
For people of your age, or that bandwidth of our age,
Yes, because we have gone through the school of hard knocks and we actually have common sense, hopefully human judgment.
But some of the younger people coming up today who are glued to their phones are

And this is all they're looking at.
And now they're allowing AI to do the work for them.
They will not have that ability to add human judgment because they haven't been trained in that.

[30:52]->[31:11] 
And so there's, well, I was going to say there's a time window for this.
For our generation, but I think it's probably irrelevant because AI is moving so fast.
We have no idea where it's going to be in a year from now.
It might be able to make better judgments on its own behalf.

[31:12]->[31:47] 
Yeah, so you raise a really good point.
So we've been hearing a lot about the class of 2026 that's booing commencement speakers who talk about AI.
And the fear is that
If we eliminate the entry level jobs where people actually start to gain the experience that they need to build that judgment,

how are they going to learn?
How are they going to get their experience right?
So I think leaders need to be very intentional about giving opportunities to young workers to

[31:48]->[32:14] 
Gain experience you know it could even be shadowing somebody that's doing the work day to day and helping them with you know the day-to-day
activities that they go through but we need to really think through this because you know if we don't prepare the generation that's coming
in now to build that judgment capability then we're left with ai running the world and that's not a good thing

[32:15]->[32:49] 
I don't know if we're doing a good enough job helping our young people understand the value of not going to AI for everything.
I don't think we even get it ourselves.
So how can we even teach our kids that?
I'm very grateful.
I got three kids.

They're all in grad school.
And all of them have refused to use AI in their program.
Not because they're not allowed to.

But because they said, we don't want to outsource our thinking.
We want to develop those thinking habits.
And they are a challenge to me as well.

[32:49]->[33:27] 
Like, dad, why would you ask AI that question?
Just go learn it yourself.
And I mean, it's a great conversation to have.
But I think...
I was down at the University of Madera today talking to some professors and they're like,

what are we going to do with these kids who have no ability to think for themselves?
And they don't even know what they should be doing.
And I'm like, wow, if we could teach kids
The idea that, yeah, go and use AI, but use AI to give yourself advanced abilities, not to be lazy in your thinking.

[33:27]->[33:59] 
Right.
Because in the end, maybe there's like a thousand graduate on a given day in, you know,
whatever this month or next month or whatever it is.
They're all out looking for work.

And you still have to separate yourself from the 95%.
You still got to be in the top 20% or the top 5% to get the good jobs.
So why not use your skills to advance your ability so that you can compete instead of being lazy to get the work done?

[34:00]->[34:03] 
So now you're going to go out and get a minimum wage job.

[34:03]->[34:04] 
Yeah.

[34:05]->[34:12] 
Patrick, this has been a fantastic conversation.
Thank you so much for being a guest here on the show.
Where can people find you?

[34:13]->[34:38] 
Thank you, David.
I am at AITransformationPartner.com.
And on LinkedIn, AITransformationPartners.
But AITransformationPartner.com.
I'm very accessible.

People want to send me an email there.
You could go to the footer of the website.
Send me an email, connect with me on LinkedIn.
It'd be a pleasure to talk.

[34:39]->[34:47] 
All right.
Thank you so much.
Zina, thank you for being a great co-host once again.
Looking forward to the next one.
Thank you, David.

[34:47]->[34:53] 
It's always a pleasure, David.
I'll be ready for the next one whenever you are.
Looking forward to it.

[34:53]->[34:54] 
All right.
Thank you both.