# Strategic AI Adoption: Beyond Tool Selection

**Podcast:** Engineering with AI
**Published:** 2026-05-04

## Transcript

We were helping another organization with Vision AI and building platforms for data scientists to effectively have these robots that weld these massive structures.
But sometimes when these robots start the welding process itself, the parts that they're trying to weld would move.
And if they move for safety reasons, the robot would have to stop, in which case mission control would be alerted.
And somebody would have to be a human in the loop.
Somebody would have to actually like go in and try to fix things so robots can continue the well.
Alina, Haney, Ibrahim, and I were all lucky enough to be attending a client conference near Palm Springs.
When we realized we all had enough time together to record an episode, we decided to roll with it off the cuff.
Part of this, we recorded outdoors at the beautiful Grand Hyatt in Indian Wells, in the other part, in a conference room.
On this special episode of the Engineering of AI podcast, we get to sit down with three members of Navalya.
the place where I work, and we get their takes on the industry.
Navalya, as you guys know, is our founding sponsor.
We're very grateful to them.
One of the things I'm interested in asking is more about Navalya.
Let's hear about the founding story and what they do, but also what are their thoughts, right?
Like these are folks who have been in the industry for a long time and have a lot of experience.
What do they see?
Are we ready yet?
What's next?
Anyhow, thanks for joining us on this special episode of the Engineering with AI podcast.
Alina, can you introduce yourself?
Hi, Ken.
So hi, I'm Alina Murphy, as Kyle says, with Navalya, but my background is about 10 years in professional services.
And I think what I like to say is I'm a professional amoeba.
And then I really geek out on the people process side of things.
I think any shiny tool can be implemented poorly.
And we'll probably get into today.
I think what makes or breaks things are the people, the mindset, and the processes around it.
Very cool.
Awesome.
I'm Hanyo Amari.
I'm one of the co-founders of Navalya.
And probably much like you, Kyle, I spend most of my time riding the elevator.
You're familiar with that analogy, Gregory Hope.
Working with the executive leadership team, but most importantly, just rolling up the sleeves and working with product engineering teams, trying to build more effective software and help our clients achieve the outcomes via trying to organize them.
in the right ways so that they can actually achieve the outcomes that they want to achieve.
Got it?
Yeah.
Ibrahim Taha, one of the co-founders for Navalya as well, mainly focusing on the operation for Navalya and also acting as a delivery principle where needed and where applicable on certain accounts.
Coming from a QA background back in the days.
So I think we shared that foundation together, moving to the project management, the delivery principle role, and then all the way till we...
co-founded Navalia together.
Looking forward for this conversation and talk about the future regime of AI.
Amazing.
Could you guys say a little bit more about Navalia?
When did it get started?
Sure.
So Navalia is a fairly young company.
We're about to turn five years old.
We started in 2021.
And Navalia in general is a global technology consultant.
The sweet spot for us, as you already know, is sort of that intersection between strategy and delivery or strategy and execution.
We sort of want to, you know, don't want to just like, you know, talk to talk, but we want to walk the walk as well.
And that tends to be, you know, where the bulk of our engagements.
We have a lot of key shaped individuals, so it's not like we, you know, we have a niche in a particular area.
We have a lot of folks who are experts in QSR industry, retail, airlines, financials and financial institutions, fintech.
So sort of general purpose consultancy, but we really try to align ourselves with the goals and the outcomes that our clients want to achieve.
And we go in there and partner with them and help them achieve these goals.
Awesome.
So lots of custom software delivery strategy, technology strategy, and these kind of do we do?
Absolutely.
And a little bit of work design as well, because, you know, as you might already know, some of the problems that we see often at organizations is that they are organized according to functional areas.
Yeah.
And oftentimes we want to sort of employ Conway's law or reverse Conway maneuver as just to try to help them organize in a way that sort of aligns with the outcomes they want to achieve.
So oftentimes that.
you know, appears to be domain-like teams or domain-oriented teams, Jonathan.
And just to clarify, I mean, about the initiation of Navalya, Hannes started it in 21.
I joined in 22 as a co-founder.
He gave me that look for words.
But again, I mean, one of the things that, just to add to Hannes' points, is what we focus on is our interview process, right?
That's why we have, like, great people like Alina working with us, right?
We focus on not only the technical skill set, but also on the actual mindset and the attitude of the individuals throughout the interviews.
At the end, you can be the most genius technically, but if you don't have the right mindset and the right attitude, you're not going to be compatible with the culture that we were able to foster in general.
I know for a fact that lots of our Navalians joined us because of our culture.
They want to be part of it and a collaborative innovative.
transparency kind of culture and they want to continue into that kind of mode.
Yeah.
So aptitude, attitude.
Exactly.
Yeah.
Yeah.
And I also just want to chime in.
I feel like one of the things that sold me in one of the first conversations was not just pushing for blind growth.
I liked the intentional slow growth because I think that's where a lot of people burn out in professional services because it feels like who upstairs is selling this thing so out of touch with people on the phone.
I really feel that was a big job of the volume.
Yeah, yeah, absolutely.
And sort of, you know, it's actually wonderful hearing you say that because sometimes it feels like chaotic growth.
It feels like we've grown like too fast.
But you're absolutely right.
I think it's like very, it's been very deliberate.
And I think that's part of the advantage of being self-funded, self-bootstrapped.
And, you know, you're not beholden to specific goals.
We collectively and not just Ibrahim, myself, or, you know, Smith, our co-founder.
But collectively, we get to talk and set certain directions for the company and say, are we comfortable with this or are we not?
Yeah.
Do you find that these days, like a lot of your customers are talking about artificial intelligence?
Is it coming up?
Like, it seems to be coming up a lot in many rooms, but not every room, right?
Like, what are you finding with the different people you're talking about?
How often is it going off?
So, I mean, it's not only the customers, by the way, like who's talking about the eyes.
tell you my three years old nephew, he's also talking about the company nowadays, right?
So it's been a trend.
As a matter of fact, on my flight into Palm Springs here from Chicago yesterday, the passenger right next to me, he's also a founder of a cosmetic company.
And he was asking lots and lots of questions about the eye and how does it impact and affect his business in general, right?
I mean...
So to your point, yes, of course, clients have been asking questions.
And I know many organizations have been reaching out to kind of help them out with, hey, help us understand AI.
Help us basically get to AI.
What is AI all about?
Yeah.
Does it, like, will it eliminate, like, certain roles?
Will it save us money, save us time?
Things like that.
So that's been the trend in general.
To talk about the AI, we cannot just like conclude it into like a half an hour or an hour kind of like podcasting.
No, it's right.
But there are lots of things that need to be kind of defined in order to understand how to answer the need and the ask or the objective of the question.
Yeah.
What are you trying to do?
What are you trying to do?
Like what exactly are you trying to accomplish?
Right.
I can tell you like this is what AI is all about.
Yeah.
But does it help the purpose of answering your question in this case or?
or serve your purpose of what you're trying to accomplish.
I love the story of the person on the plane next to you, right?
Like Ibrahim and I were talking before you guys came down, like New York Times, The Daily, you know, one of the biggest podcasts in the world was basically about Vodkos this week.
I spent an hour just talking and, you know, not like...
This is the news.
This is the news.
It's everyone.
everybody gets what AI is all about until like you get to the specifics of what exactly are the objectives.
Otherwise, it's just like a loose term that could be used anywhere.
But yes, we have been seeing this across our clients and even non-clients organization.
I've been hearing a lot that people are reaching out for like, how can you help us out and like converting into an AI shop, for example, right?
And as you guys know, it takes a lot.
If you remember back in the days when Agile started off, I was part of a couple of engagements at least for organization transformation to move organizations from all traditional model into Agile.
And we've seen lots of stories and examples where basically some organizations were thinking just because they had two developers on the same machine called pair programming or just because basically they have a two-week sprint.
they call themselves officially an agile shop.
So we get in there and then we start basically trying to convince like, no, this is not agile, right?
Agile is much more than what we're talking about.
And this is where we had to go through basically like a gradual implementation of our transformation and recommendations and convince them this is the value of this and that's the value of this.
You can easily get some sarcastic kind of comments in lots of cases.
I remember like one of the developers at this engagement, he said, sarcastically, of course, like, what do you mean by pair programming?
Like a developer will be typing on a keyboard and the other one is helping them out like with the mouse.
Right?
It's a waste of time.
Yeah, exactly.
So you need to absorb and like, you know, take on these kind of sarcastic comments nicely.
And sure enough, every change you introduce, no matter what it is, no matter how small it is.
it will reflect into the acceptance or resistance phase, right?
So how good you are at basically managing this acceptance phase, this is where it's going to take you either north or south.
So managing through that makes a huge difference in general.
My point is AI currently is, I think, is following the same pattern.
You can just buy certain tools where you think like, those AI tools can basically do me one, two, three.
and claim that I'm an AI shop in this case, right?
It takes much more than this.
So our recommendation is always to say like, hey, you know, first help us define your objective.
What exactly are you trying to accomplish?
And it better be a clear, straightforward kind of an objective of what you're trying to accomplish in this case, right?
And then based on that, we can take it and start like, you know, break it down into how can we help you take it to the next step?
Right.
So let's identify all your gaps and all your strengths as well.
Yeah.
You know, I mean, you might have strengths that you're not aware of.
That's our job and responsibility to kind of like highlight them for you.
And for sure, you'll have some other gaps.
Like, do you have the right competencies, the right resources?
Do you have the right tools in place, for example?
What about AI regulations and guidelines?
Right.
That do you have like, you know, any like specialists that can take care of that?
And then maybe the recommendation in a nutshell would be like, let's take it gradually in a controlled fashion.
But as you are moving step by step into the end goal, you always have to put the objective in front of your eyes.
If you're moving to the right, then that objective has to follow and align accordingly.
If you're moving to the left, likewise in this case.
Start with a low risk project with like a smaller team and start scaling.
to other higher risk projects or bigger teams accordingly and take it to the next step.
On some levels, it almost reminds me of like big data, you know, and everybody was like, oh, we have to have big data.
But, you know, don't worry, we bought a big data.
tool.
Okay, good.
Yeah.
Yeah, very buzzy.
Not even sure what they have or why they have it or if it fits them.
And we were talking about this yesterday that any tool, any process you could do poorly, they're just a tool in your toolkit.
And you need to see if we think of people and companies as an alphabet.
And let's say AI is Z.
If someone's at A, you can't shove them over to Z.
So you can help get them to C.
And it's not very sexy.
And that's not what they want.
They want to get to Z right away.
But there's a way to do it properly where you don't Frankenstein or just have AI in name.
But it's not actually helping you.
For sure.
And to tie off to what Alina and Ibrahim said, it's sort of a spectrum from like what we're seeing, right?
Like you have.
Some folks who are at the end of the spectrum who've adopted a tool and said, okay, we're done with, like we've adopted AI or what you've done.
Right.
Yeah.
So that's like one, you know, one end of the spectrum.
Then there's the other end of the spectrum, which is basically this thing is a massive security hole or like we really need to understand the regulations around it and like, you know, governance around AI.
So we are not going to take a step until we figure this out.
And that's not a judgment on this end of the spectrum or that end of the spectrum is just basically a comment that says, well, you're not quite done yet, right?
Like it's great that you're adopting that position, but there is more.
So if you're on this end of the spectrum, you've adopted a tool, great.
Let's now build an AI strategy.
If you are on that other end of the spectrum, super admirable that you are, you know, having security top of mind.
But, you know, there's a whole lot of white space between where you're at and sort of like experimentation with AI.
Examples, we see some folks in the middle who might classify themselves as like AI native organizations.
And there are certain truths to that.
There are much smaller, much more nimble organizations who've probably started more in the data science world.
But I would say they're also not done yet because.
Their background is more data science and scripting.
And oftentimes what they need now in the age of AI is more thinking about platformization and engineering principles to couple with what they're actually doing with AI.
Yeah.
And it is interesting.
I do feel because AI is so radically different than some other trends and waves that have come before.
It is more black and white thinking that we're seeing out there.
And like Haney pointed out, there is this whole middle spectrum where is kind of where the secret sauce is, where you could get into.
And so we're trying to help people who are so afraid to take a step, get there, or who have gone way too fast the other way.
And it's a mess and maybe are not even aware of what they're doing and what that looks like in the future.
Maybe take a step back into the white space, into the center, which is just more experimentation, thought.
I am a big pro of being.
Not a naysayer, but I'm a QA background as well as Ibrahim.
So, I mean, we were just those why kids.
Why?
Why should we do it that way?
Could we do it another way?
Just poke, poke, poke, poke, poke.
And I think that that's helpful as well.
Like what about the applications of it?
Are we seeing interesting applications of AI that we think are really great?
Do we suspect that we know where some are?
Is there anything that we admire out there in terms of using AI to accelerate business and not?
by building the new digital platforms and not by doing software engineering?
One of the things I'm most interested in is just the ability to find patterns in data and things that it would take humans so long to be able to pull out, but then also where we have blind spots.
So I'm really interested in more the, how it's going to apply to the science field.
How are we curing diseases?
I mean, I think that that's such a cool application.
Anthropa has talked a little bit about how like maybe medical is next.
And there have been startups that are like found new proteins and things like that.
Yeah, for sure.
It's happening.
Yeah.
There are some use cases that we have been engaged in with, you know, with some of our clients.
They are narrow use cases, but they are quite good.
So it's like I wouldn't say that, you know, these enterprises have adapted AI at scale.
But for example, one of the QSRs that we help.
And we've helped them successfully launch a conversational AI ordering, right?
And the play was not necessarily to replace humans at the drive-thru or anything of that nature, but it's really to augment and to alleviate some of the labor shortage that some of the QSRs face.
There is a whole lot of learning in that field, right?
It's like natural language processing and natural language understanding.
You're trying to figure out what language even is and how languages is constructed.
Thanks to AI now, you know, we're able to do much more advanced things in that field and, you know, things like disambiguation.
What did the customer really mean when they said this?
So that's one particular use case that's quite exciting.
And I think it's probably starting to become mainstream in QSRs.
To QSR, you mean quick service restaurant?
Yeah.
Yeah, thanks for that addition.
And then there are other areas as well, whether it's in quick service restaurant or manufacturing.
with vision AI in, you know, like in particular.
So yet another narrow use case in quick service restaurant.
We've also been part of, you know, partnerships and engagements that assess cameras at the drive-through just to try to understand things like, is the line really building up?
Are folks actually peeling out of the line?
Are we potentially losing customers?
So that was a really interesting use case.
And then in manufacturing in particular, We were helping another organization with Vision AI and building platforms for them and for data scientists to actually effectively have these robots that they have weld these massive structures.
So it was a really interesting use case.
But I think the common thread across these use cases was, let's start with a particular business problem, which I absolutely love.
And then let's work out the engineering details and sort of through the artifact of working with AI in this particular use case, you end up figuring out, okay, well, the development of this thing is just not enough.
We actually might need a proper data platform.
We need the infrastructure to support this a little bit better to make the AI more effective.
And you are sort of bringing us back to the software engineering side.
But just to dive a little bit deeper, so robotics and vision and to do welding, like it seems like there would be all sorts of problems to solve there.
Some of it sort of at the camera level, like a welding arc can be really bright.
And like, how do you even see?
But I imagine if you could solve that, like what were the different domain challenges that we had to solve there?
As Nevalia gets engaged with clients, we try to...
as much as possible to become an expert in the industry that they're in.
I absolutely love this part because there is a ton of learning in that.
Mission control was one massive domain.
So for example, sometimes when these robots see the seams, so there's perception domain as well, and then there's a welding domain and robotics domain.
But sometimes when these robots start the welding process itself, the parts that they're trying to weld would move.
And if they move for safety reasons, the robot would have to stop.
So in which case, mission control would be alerted and somebody would have to be a human in the loop for welding and for automation.
Somebody would have to actually like go in and try to fix, you know, things so robots can continue the weld.
So that was an interesting domain where you're trying to reduce.
that time to understanding that something wrong actually did happen.
So, and sort of like mean time to recovery, applying to manufacturing.
So that was, that was a super interesting learning as an example.
Very cool.
Well, and I, on the QSR side too, like I'm thinking, I'll admit, I go to McDonald's a fair amount and they always ask you like, what's your code?
At the one store that I go to, they're really good.
And I always feel like it's like, oh, hold on a minute.
You know, I got to get my phone out.
And I always feel bad for them.
It's like they're sitting there just like waiting for this guy to figure.
But yeah, the AI doesn't care.
Like it can wait forever for me to figure out what the code is.
So it does sound like there can be some real efficiencies in both of these applications.
Initially, we were thinking about.
speed of service as a metric, as a very important metric that the AI can help with.
So in other words, could the AI do upselling or not do upselling if it's like peak hours and the drive-through lines are getting longer?
What we've really discovered, and I think this is probably something that's been written about, is that it's not necessarily the speed of service.
Speed of service is an important metric, yes.
but it's really the engagement.
Does the customer feel like they are engaged with or are they just sort of like waiting, right?
Oh, interesting.
I think, you know, one QSR that does this very well is Chick-fil-A.
They tend to have very long lines, for example, but they have this upstream ordering process where somebody...
is actually constantly standing in the rain yes no matter what yeah or like a tablet or something but they tend to have longer speed of service than all of the other you know like qsrs yeah but you feel like oh i've talked to the guy i can't leave yeah so it's an interesting problem to solve from an ai perspective like how do you stay engaged with uh you know like with the customer and make sure that uh that it's really you know about speed of service but also customer engagement yeah I think we see that a lot too.
And a couple of my past clients have worked a lot on the customer service side of things.
And that's probably one of the early adoptions of AI are these customer service bots.
And I think that's a perfect example of that too, because even though it is answering you, depending on what you are going to customer service for, you're in a pretty emotional state.
What I was working for was going to ask questions directly tied to your financial and health benefits.
So you're in a pretty emotional state about the well-being of yourself, the well-being of your family.
And so even though the bots could answer questions, people want to feel heard.
People want to feel like they have an easy path to escalation.
And so I think that was one of the great first learnings of AI because you could mass support people.
And yet the customers almost were more upset.
Yeah.
Which doesn't mean it's a failure.
There's tweaking always.
It's not black or white world.
But yeah.
Yeah, exactly.
And one of the clients that we implemented the Fresh AI, back to my point at the beginning when I said like agile.
maturity model.
Now we should have something called AI maturity model, right?
What exactly is the speed required in this case?
Does it really reduce my cost in general?
Does it lower my risk?
All of these ingredients have to be considered.
But most importantly, I think, is the build, measure, and learn is what implemented.
Again, simple formula, right?
Like build this small piece, measure it.
understand the outcome, learn from it, repeat, continue repeating until like you get to the end point is basically the right thing to do.
And can it inspire engagement?
Because if there's no engagement from your, and I mean more of emotional engagement of feeling like you know that brand, that brand knows you, because if not, the next competitor comes out and you have no emotional tie.
So you're just kind of shopping around.
Okay.
So we've talked a little bit about, you know, what we're seeing in industry applications of AI.
the use case I'm interested in, and I made a podcast about it, is using AI for software development.
And when you think about the different clients that we're working with, what does good look like?
What does something that isn't working well look like?
What are the stories that we're hearing out there?
One of the things, and we say like engineering with AI and we focus on the software engineering bit, we all here come from the same background of understanding software engineering to be.
the software development life cycle.
So it usually doesn't start just from the engineering bit, but it actually starts from product.
And I think some of the anti-patterns that we might see is that focus on just pure engineering.
It's like tool selection just for engineers, right?
Ignoring potentially some of the other bottlenecks that are in the software development life cycle.
And I think we all know here that, or at least like we're seeing it, that AI amplifies your bottleneck.
Yes, yeah.
Right?
So if your QA practices or your QA infrastructure is not up to par, now you have a lot more code throughput, so now the bottleneck is massive.
If your CI, CD practices aren't there, then, you know, again, that's amplified.
If you're, you have...
poor product that's more like upstream curious about your perspective there as well if um you have poor product um you know practices or not the right skills around product and you know requirements and you know business formulation and understanding the the problem at hand again that's you know that's going to lead to you know like poor outcomes so i would love to see all of these different groups but you know, part of the, you know, the SDLC itself, the software development lifecycle, I'd love to see their input as part of the selection criteria for the tool.
And, you know, I'd love to hear their voice and, you know, like across the enterprises that we help in sort of tool selection.
But curious.
Yeah, I was actually going to say that that's one of my qualms with AI right now, at least in my paid job.
I see most people starting with AI centered around the software developers.
And so I've been able to play with it, but it's mostly all been in personal time.
Or if I'm using it for a client, I mean, have to take out all scrub, any sort of data.
It's more, hey, can you create this outline for me?
I'm working on an SOW and then plugging thing in.
So not necessarily working in my nine to five.
So I'm excited to get into that area because I don't know if you all grew up, probably not, but with the books Amelia Bedelia, have you guys heard of that?
So this is what I think of in this use case because in those books, it's they tell her to dust the curtains, dress the turkey.
There's a whole list for the maid of what to do.
So she literally throws dust on the curtains and she puts the turkey in a dress, taking it very literally.
And I...
I'm imagining that when it comes to we are now speeding up that software developers can build things.
But if you're not focusing on the other parts of software development lifecycle, well, I might be making you speedily put dust on my curtains or speedily dress up my turkey.
And that's not what I meant.
I am very interested and excited for there to be more attention and respect to the other.
parts of the process.
And it has to start somewhere.
And like you said, if we started in the product and in being able to make specs and come up with all of that, there'd be another slowdown.
So I think that we'll get there.
But as a program manager, I am particularly salty and excited for when it gets more to my life.
Yeah.
Yeah.
And this is exactly, this is the same concept that we're talking about is what we call the value stream, right?
So the value stream is like how to shift quality all the way to the left.
When we talk about quality assurance, that qualities shouldn't be owned by only the QAs or by the testers.
It should be owned by the developers by providing a quality code.
It should be owned by the business analyst to give me an invest model, kind of like user stories, for example.
It should be owned by the architects, by like every single member of the development lifecycle team.
And so how to shift like this to the, to as left as possible, utilizing AI should definitely help you in having the successful outcome in this case, all the way to the right.
And this is maybe a good.
segue, like we do have lots of questions, whether it's from clients or from even Navalians, from friends, from other organizations that we deal with.
It's like, is AI going to replace my job?
Is AI going to take over?
I believe that AI is not going to wipe up jobs.
Rather, it's going to change the job description, most likely.
And what I mean by that, if you rewind back to the internet tech era when it first started.
and followed by the cloud computing, when they were introduced, they did not eliminate or wipe out jobs.
Rather, it focused more on repetitive tasks, for example.
It focused on creating net new different roles descriptions.
It also helped upskills the talent and the skill set of individuals.
And I feel like AI is in the same pattern nowadays, where it's basically following the same rhythm.
It's not going to eliminate those jobs, rather.
So my point is when we talked about product and AI, there might be new roles called, let's say, AI engineer.
For example, I don't know.
I mean, AI product, right?
How to integrate the AI engineering into the product itself in this case, right?
So all in all, what we believe in is basic development skill set, like entry level to maybe just an average level developer.
might be impacted if he or she does not kind of like learn how to utilize AI to ensure that they can catch up to speed, right?
So will AI eliminate these kind of jobs?
Most likely, yes.
But maybe the question should be put in a way where will AI wipe out jobs?
No.
Will developers with AI experience be taken over roles for developers who don't have AI experience, for example?
Then yes, right, in this case.
And this is applicable for every other role across the IT organization.
It's not only applicable for software engineering, as we can tell.
So it's going to be more of like less typing codes and more focusing on system designs and architecture and understanding the business and integrating the AI into the actual product.
Maybe another role could be created called like...
guidelines or AI guidelines specialist, you know, things like that.
And we live in a very, as much as we live in a world of machines, I feel like we live in a more, a very human world.
So AI can help us with some of these highlighting, hey, you have this issue, we see this pattern, here's a risk, but it's not going to go into the boardroom and talk about what is the decision because...
we aren't always making black and white logical decisions.
Often it is we know that we are losing money here, but actually net-net, we're making money by focusing over here.
And that's a very human decision that a computer would maybe have a harder time with.
And it doesn't all come down to just money.
Like we were saying, there might be short-term gains of money, but then you've actually lost engagement and the heart of why your customers want to come to you.
And so then there's long-term effect.
So I do feel...
that there's a lot of short-sightedness with AI.
And that's okay.
I think that there's value in that.
So I feel like AI is really good at kind of speed to market.
And there's a very short-term vision there and application and value.
And I feel like we still need the humans to think through the long-term.
So an exact example of that that I was talking with my developer friend about is they needed, they wanted to accept.
essentially pixel push something so they wanted to change the font of something to be one size bigger and immediately ai did that it worked great it was beautiful but the way it did it in the code was not at a global level so you're creating this frankenstein so there's value maybe you've released this promotion and you need it to look right or do something a certain way so there's value in the speed to market Great, AI did that.
But you still need the human thinking, hmm, two years from now, if I have done everything that way, I have made this huge monster that is unmanageable.
I think you're hitting on something.
I like the democratizing aspect of AI that allows anyone to do something.
That is only good, right?
I'm not going to say like vibe coding is terrible.
I think it's great that, you know, someone can sit down that's never in software before.
They can do it.
But the point you're making right there is like, yes, and that will eventually lead to messes that if you don't understand software development, you don't know how to avoid.
Or give me a scalpel.
That doesn't make me a surgeon.
I hope not.
Yeah.
You know, and then give my doctor cloud code and they're not necessarily a software developer either.
So while there is value in that democratization, it doesn't mean that like, oh, we don't need software developers anymore.
And I think one of the best examples of that is security.
So anyone out there can be using it and they can have be creating value or doing something that they need, but they have not sat with developers.
They don't know how necessarily the thing is built.
And so they don't really understand security is a buzzword, but they don't really understand the risks that they're incurring and how and why and when.
So, yeah, it is valuable in everyone's hands, but there's also different risk levels.
It is good to have people who truly understand how the machine is built to be doing the heavy lifting.
Exactly.
You check if the AI parameterized that SQL query.
And if not, that means it's open to string injection.
So we all know security is important, but it takes some experience to know that you want to parameterize your SQL queries.
And what I just said about the human element, there's these trade-offs.
So very rarely are we building something where the client says, we have all the money in the world you have all the time in the world you can gold plate this you can do it absolutely perfectly and also now the world moves so fast that the definition of perfect, by the time you were done with perfect, it might be behind.
Exactly.
So there are all of these trade-offs where it is we know we have to be secure, but we have to balance that with the cost.
We have to balance that with the timeline.
And you're making all these decisions.
And so that really comes into place when you have people that are just using this tool and not necessarily understanding that.
There is this blind trust.
blind acceptance that this was gold plated.
And I'm not saying that it is a bad thing of the industry that everything's not gold plated.
It's a reality.
Well, we are not a time, but you guys have invested quite a bit with us today and I appreciate that.
We have a few minutes left.
I have a few questions that I usually ask folks at the end of every podcast and I haven't decided.
Should I make sure you all answer each three or not?
So let's start with one and see who jumps in and see how it goes.
what is something that you've had to change your mind on?
Now, there's a rule.
You can't say AI.
You can't say, I used to think AI was going to fail, but it didn't.
I'll tell you what I'm sort of like grappling with, right?
We were always of the mindset of small atomic commits and, you know, like, and work in small chunks.
And like, I still think that is true.
But also AI can move so much faster in larger chunks.
So I'm curious what's going to, like, would I be changing my mind sometime soon about that particular philosophy of like smaller commits equals better?
I can see where it can be better from, you know, like a CI, CD perspective and controlling, you know, like if you need to roll back something.
in production but uh but yeah that that's something that i've been you know that i've been grappling with you know like uh quite a lot lately so um yeah so for me basically i mean uh hany knows my wife nerman right and um she she has started like along with her like friend the non-for-profit organization back in 2001.
uh they started off with like um two individuals, the two co-founders like Nerman and her, Aitad Al-Shilabi.
And now after like, you know, these many years, they have about like 100, 110 kind of like employees working directly for them.
And she's across not only the state of Illinois, but across all the states.
And so the thing I'm bigger on.
What does it do?
So they focus on like domestic violence and woman awareness, welfare, family, welfare and things like that, like helping different communities in general.
But bottom line is I remember when she first started and for the first like at least seven, eight years, if not more, you know, it's a non-for-profit organization, right?
And she wasn't generating any money out of like the agency.
So no income.
And we were living off of one income, which is me.
Just starting out.
Just starting out, right?
Exactly.
And then even when she used to make some income here and there out of the organization, she would reinvest back into the non-for-profit, right?
Typical.
And one time I said like, hey, you know what?
Maybe I had a bad day that day, but it was like, for how long are you going to be doing this for free, right?
I mean, it doesn't get to me yet.
So why don't you just like, you know, find something that is...
I mean, you have a master's degree and you have like good education and good experience.
Find something that is that can generate some money and help us out.
Right.
So she told me she told me a statement that I will never forget up to the day.
And she's like, Ibrahim, you don't understand the satisfaction of helping the needy people that gets into myself in this case.
And it's very rewarding.
Right.
I said, OK, you know, that's fine.
Now, it didn't hit to me until like a couple of years later after that statement.
I was working on one of the non-for-profit organization projects in Amman, Jordan, helping the Iraqi kids with the malnutrition through the UNICEF, United Nations.
So we had lots of Iraqi families coming in, like widowed women, like orphan kids, and we were just trying to help them out using technology of how to report malnutrition and issues and things like that.
And that's when it got to me over like those five days that I work so closely with those folks to help them out.
I mean, we all like we all care about others and stuff like this, but it doesn't get to you until like you really go through it and feel feel it.
And so that's when it got to me.
And I'm like, OK, you know what?
Even if she works forever as a non-for-profit without generating any income, I'm very OK with that after getting that that strong feeling of satisfaction in this case.
So that kind of like.
It didn't like, I mean, I don't want to like, you know, be looked at or come across as like, I don't care about others.
Right.
But it just like shifted my mindset in general.
It's like, Hey, you know, each family has to sacrifice something or one income may be in favor to help others or the needy people that should be acceptable.
Maybe that was kind of was maybe almost unacceptable, like for a long time, but even if it's a long time or permanent.
It should be acceptable to me by then.
Yeah.
I feel like I kind of have a similar one where just kind of understanding value.
And so I feel like in my career, it was just moving forward at the fastest rate that I could and value being from my career being money.
And I think you can only do that for so long and then you get burnt out.
And then I just really changed my mind and tried to decentralize that view and started looking at my life more as a portfolio.
I remember being in a conversation and I was saying that I was really not doing well at work.
I felt like I just was underperforming.
And I mean, it wasn't in the outcomes, but I just felt like my heart wasn't in it.
I feel like consulting can often ask.
200% of you at this time is when I was traveling for consulting.
So you're also living on the road.
I mean, you're just living and breathing.
And I drank that Kool-Aid for a long time.
And I remember having a hard time when I wanted to focus on something else.
And someone said to me, because I said, I thought I was failing.
And they said, do you think you're failing at life?
And I said, no, absolutely not.
I mean, the relationships I have, I can show up for my parents.
I think of last year being able to go for four months and help my mom through her big cancer surgery.
It was funny that I really did not and do not feel like I'm failing at life or ever have been.
But I was putting so much of the failure on just monetary value and getting the promotion.
So I don't know if that's so much of a change my mind or is more of expand the view.
And it really helped decentralize that compartment thinking that I had of almost like there's five Alinas.
There's the family Alina and there's the friend Alina and there's the passion project Alina and there's the career Alina.
Took the pressure off looking.
I am one person and I am a portfolio.
Yeah.
I think we all get there in our life.
I think it is maybe a younger view to have at first, but you can't operate your whole life like that because you can't give 500% for long.
Yeah.
Something you're most proud of?
My relationships.
I love my people hard.
They love me.
That's what I care about most.
My kiddos.
Yeah.
I was going to say the same thing in this case, like having...
Like my two boys, making sure that we invested in them in education and get them the best position possible.
I'm so proud of that.
Something that's giving you joy right now.
I often have workaholics on this show and they kind of look at me like, what?
I'd say for me personally, just because things have been, you know, so incredibly busy with Navalya as we've grown with AI, I can actually do so much within a 30 minute span.
And that has been absolutely fantastic and such a huge enabler.
Whereas before I needed, you know, like a good block of time, you know, three, four hours to actually do something or tinker with an idea or, you know, try to be creative and exercise a little bit like creative brain power.
Right now I can sort of like, you know, drop upon that pretty quickly in between meetings, like in 30 minutes or like one hour.
And I can really try to see, you know, where I want to go next with, you know, like with this idea.
I thought you were going to say tennis.
Well, yeah, I mean, I want to sort of respect the engineering with AI.
No, this is what's giving you joy.
Tennis gives me a ton of joy.
For our listeners, we are at a conference this week and we're staying at this really great hotel.
And apparently it has a pretty good tennis, not at the hotel, but.
Nearby.
So there's Indian Wells, which is a massive ATP 1000 tournament that happens, you know, at the beginning of the year.
So I've been on there every morning and hitting out some of the pros there, which has been fantastic.
I think just being in person, I really miss being in person.
I do love the flexibility.
Don't get me wrong.
I've worked from home.
But I really, I really miss being in person.
And I've just been thinking about the...
the off-the-cuff conversations that have happened, the opportunities, like, hey, do you want to record a podcast?
It just is flowing and the creative energy, it's just made me really happy this week.
Yeah.
Yeah, I mean, there's no doubt that lots of things can give you, like, even simple things can give you joy.
There is no doubt, right?
I mean, I feel like, you know, again, helping the needy people is definitely, like, honestly gives me a lot of joy.
My point is like even like running the operation for Navalya in a way that even like that things could be stressful in lots of cases.
Don't get me wrong.
Interviewing like, you know, new potential candidates and talking to them, learning more about them, tell them more about Navalya.
Working like, you know, with the Navalians on issues, working even with the clients.
And when you see like you have fixed problems for your clients and like, you know, help them basically deliver successfully.
It's just priceless, right?
So these kind of things within the work itself, the daily work, five, six days a week, sometimes seven days a week, right?
Honestly, despite all the hard work, it's definitely rewarding from an enjoyment perspective or joy perspective.
Yeah.
It's fun.
Not so much the TDM, the status reports, but just...
The getting in the room, throwing spaghetti at the wall and really just trying to, I call it solving the problems of the world.
Like you just stay up and you're just talking about everything.
And that's what it's freeing us up to do.
And it's really rewarding and connecting.
Exactly.
Exactly right.
Okay.
Thank you so much, you three.
I appreciate having you on the show.
Like you said, this was last minute and I was glad we were able to get it done.
So thanks everybody.
Thank you.
Thank you.
Appreciate you.