Shahin Nabavian — The Future of Mobility, Operations & Intelligent Infrastructure | Future Ventures Podcast Ep. 028

Shahin Nabavian, a computer scientist turned venture builder, has worked across operations, infrastructure, and transformation for the past decade. His experience includes a London-based financial startup, venture-building at Shell's maritime division, work at The Economist Group, and founding edtech startup Super Savvy Education. Now, he leads Team CMV (CentricMind Ventures), focusing on emerging technologies like infrastructure enabling companies to use LLMs effectively. Their main venture, Go-User, applies this to people services. 

The conversation is important because much of what is called' AI adoption' is actually just deployment without ROI. Shahin has observed companies of all sizes misusing chatbots, wasting tokens, and ending up with chaos. He can also explain why this happens in simple language. If your company earns $ 3m-$50 M and you're exploring AI's ROI, this episode is for you. 

Topics Covered 

• Why company culture beats company size in AI adoption. Why The Economist Group made more progress than expected, and why Shell's scale became a hindrance, not an advantage. 
• The real reason AI ROI is missing. It's not the models. It's not the agents. It's the infrastructure layer beneath them — and most companies haven't built it. 
• Every company needs a brain. What Shahin means by "AI brain," why generic RAG implementations fail, and what a librarian-style routing layer actually looks like in practice. 
• The evolving role of the founder and CEO. What AI takes off the executive plate, what stays uniquely human, and why "purple unicorns" — operators fluent in both business and tech — become the default hire. 
• Why Kodak and Blockbuster aren't technology stories. Shahin's view on Board dynamics, shareholder pressure, and the kind of CEO incumbents actually need to bring in to survive a platform shift. 

Key Insights 

1. The bottleneck isn't intelligence — it's plumbing. The models and tools are well-developed, but most organizations still need a way to understand how their business actually operates. This way, the AI can direct questions to the right source instead of randomly searching through every connected system. 
2. Engineering-led AI initiatives consistently underdeliver. Starting a transformation by focusing on tools instead of clear service goals can cause a lot of manual work, higher costs, and confusion for users. It's better to start by understanding what the business wants to achieve and then plan the system around that. 
3. Transformation is a challenge for the Board before it becomes a tech issue. Most CEOs don’t stumble because the technology is too complex; instead, they often face pressure from shareholders demanding faster results than are realistically achievable. To help overcome these hurdles, companies might consider bringing in a CEO from outside the industry, pairing them with a COO who understands the current business well, and giving them the freedom to learn from mistakes along the way. 

Links 

• Shahin Nabavian on LinkedIn: https://www.linkedin.com/in/nabavian/ 
• Future Ventures: https://ca.linkedin.com/company/future-ventures-corp 
• Future Ventures Forum: futureventures.ca/community 

About the Guest 

Shahin Nabavian is the Founder of Go-User and a co-founder of Team CMV (CentricMind Ventures), where he builds AI-native systems for fast-growing companies. He holds a PhD in computer science and has spent over a decade building and supporting ventures across financial services, energy, publishing, and education — including work with Shell and The Economist Group. His current focus is on the infrastructure layer that turns organizational data into something AI can actually reason over. 

Transcript

SPEAKER_00 0:02

Welcome to the Future Ventures podcast on Scaling with Clary. Today's guest is Shaheen Napavion from Team CMV. Shaheen works at the intersection of operations, infrastructure, and business transformation. He's helping organizations understand how complex system technology and execution come together in the real world. In today's conversation, we explore leadership, how to scale operational-intensive businesses, and what companies consistently underestimate when modernizing the infrastructure and workflows. Shaheen, welcome to the stage.

SPEAKER_02 0:34

Thanks for having me, Maxim.

SPEAKER_00 0:36

Oh, it's an absolute pleasure. Um the way that we like to organize these uh interviews is kind of a three-part tri-part day. Part one focused on uh getting to know you and and who is Shaheen. How did you come into doing what you're doing currently? Then talk about um Team CMV is the venture builder and Go User is the um is the company that you're currently focused on, and then kind of like what are you seeing across uh across the industry? So why don't we just jump straight straight in and uh get to on discover who is you?

SPEAKER_02 1:18

How long have you got? Um so I'm a computer scientist. Only 45 minutes. I am a computer scientist by uh by trade. Uh that's what I studied, I got a PhD in it. My career was an engineer, and it quickly transitioned into actually more on uh in human interaction design uh and product building. Um, and the first startup that I was involved in, still going strong, uh a financial services company uh startup in in London. And it was the first time I really enjoyed actually uh building companies. Uh so I think it wasn't even just the products or building the teams, it's like how do you kind of grow um you know uh a company? That's what I got really interested in and spent probably the best part of the last 10 years in building ventures uh myself or uh supporting other companies, even like doing venture building with Shell, um supporting uh kind of mid-sized companies like the Economist Group, uh, do transformation and build product teams, all those kind of things. Um, and most recently, uh over the last few years, of course, uh we've been involved heavily in um obviously uh in nascent technologies, that's where we specialize typically, me and my kind of co-founders. Uh, and of course, AI was one of those spaces. Uh we in particular were keen to solve what was a uh still is a problem with AI memory and context, and how do you kind of build something that could be leveraged uh in a enterprise basically uh that can be accurate, uh cost effective. Uh, all the themes that everyone's discovering now has been something we've been working on for a few years. Um, and I guess we can take it from here on.

SPEAKER_00 3:12

Um I'm absolutely curious to get your perspective on the way that companies are going about adopting AI or nation technology. Uh, and if you can jax thepose Shell is a big super major, you know, and gas um with I'm assuming tens of thousands of employees versus perhaps uh a mean market company or uh a small small company kind of like what is the approach? Who moves faster? What is different? How do you drive change management? How do you drive adoption? Where is AI sitting? Is it centralized function? Is it embedded as part of the culture where they have a kind of champion ambassador to the driving the adoption and use cases?

SPEAKER_02 4:01

Um, very good question. So I think company culture is one of the biggest factors I think I came across. That's important regardless of the size. Uh the Economist Group is, for example, a publishing company, is uh a British publishing company, quite a conservative company. Yet I found it a good place to be able to kind of do transformation simply because they had a culture of saying they're not going to get in the way, right? If this is what you want to do, go with it, type of thing. And I think we were able to make quite a lot of progress uh in in lots of ways, and of course, not as much as you want. Uh, but frankly, a mid-sized company like that, uh, the culture allowed you to experiment and try things, right? Uh, and there was a real um desire by the internal team to do transformation, and this is obviously very early stages by the time Buddy and I was leaving. This is kind of uh far before the kind of the LLM movement. Uh, we were already experimenting with crude AI tech at the time, like uh forgot what they were. They're just glorified automation tools, basically, right? But we called it AI. Yeah. Um, you know, so I think people were willing to uh uh accept it. And the MD, uh the manager director, one of the businesses, actually set up a little working group, and we kind of picked out some spaces that we thought maybe we can use these tools, for example, building articles. Um, and this is again pre-the LLM era. So, in that nascent space, to your point, I think the culture of the company and uh someone championing it so that the MD top down or a CEO can top down, can drive that. That's that was that worked quite well, or at least it was progressive. Uh something like Shell is a beast, you don't have one division, it's it's fragmented. And again, it's really you know, the area that uh I had some connections and work with was in the maritime part of Shell. So deep deep vessels and all those kind of things are already steeped in technology. Um, and there because they had uh they had set up like you know, venture building arm or product building arm, that was much more experimental, uh, you know, experimenting uh with different technologies, it was easier to to incubate ideas. Um, the thing that I ended up doing is uh also building relationships. So when I left the the setup, um we did a like a joint venture uh RD, just me and my co-founder, which is very difficult to do um with a company like Shell. Uh you have to go through hoops. Not many people understand how I was marked out as a really big achievement. I was like, Yeah, so what? But like, no, believe me. It was everybody, and so it was again to your point. I think what mattered there is uh who's involved, where can they influence inside the organization because they have to move a lot of hurdles for two people off the street to kind of get kind of this kind of deal, and we ended up basically again pre-LLM era developing like uh what we picked out is a real problem with shipping, which is um deep sea vessels on like airplanes, right? The whole industry is like the 1950s or 40s, simply because an airplane can be much more you know prescriptive in terms of how they're gonna kind of travel and you know when they're gonna land, etc. Whereas a deep sea vessel may have to stop and clean and make repairs and things like that. So even though you can track it with satellites, the sh the captain's word effectively drove the ETAs, right? Um, and so observing them sending emails to ports and someone come going on holiday and not seeing an email was for me as a technologist like funny. Uh so we ended up experimenting with this very crude. Uh we had to build our own pipeline of saying, okay, we can just read their emails and extract what's a report data point and create an API on true systems, for example. So this is the kind of thing just to give you a snapshot of where a huge organization can start to be experimenting with different technology, go through the hoops and and and see where they can take it, basically. Uh, ultimately, I think it will be down to whether they can adopt these things en masse. And I think we're we're in a different era now. This is like a few years ago, even like you know, four years ago, it's completely different.

SPEAKER_00 8:37

Yeah, I mean, it's interesting to me. Um there is studies upon studies that are coming out that uh companies are not realizing uh ROI on their AI investments. And uh what I'm trying to square off is how does this make sense? Like I know that in our company, we were a small company, um, just shy of 10 people, and we must have uh tripled our productivity in terms of what we're doing. We've built mini RPs, everything is urgently automated. Um like I mean, everything like from a conversation to a proposal we can do in like 45 minutes with with brand kids with skills, like everything everything is enabled. So I just look at this and I'm trying to understand why our companies struggling to realize the ROI when I mean this is a personal example. We've built mini RPs within three months that have that just the build of the DRP is is the equivalent of one year of of licensing uh that we would have had to pay. So like, okay, I incur my cost in the first year, and then yeah, I have to maintain it, but it's custom built for me. Uh I get to use it any way that they want. And so I I just it do you have any hypothesis as to why companies are struggling to realize DROI?

SPEAKER_02 10:03

Yes, I think that it comes down to a few aspects. One is what you're expecting the workforce to do is adopt different ways of working, and whilst you may have a certain percentage, and uh what I name, no, I can't name this company. I will name admit a uh a European company making let's say over 100 million in terms of ARR, you know, 300, 400 people, tech company. Looking at them from an inside, um they are very progressive in trying to adopt to your point. Um, and I think it does take them every company, you know, maybe it's a head count figure that goes every 50 employees or 100 employees, it starts to kind of become more and more challenging. Where you're gonna say, okay, here is here's a bunch of tools. What probably it hasn't happened is a very simple way for people to understand, oh, I can, you know, people are just deploying a chatbot and saying, hey, use it. And the main thing they've been trying to do is trying to make that chatbot safe in terms of the data that's in there. They're struggling with things like rag and memory and you know how much they could put in there. But when when I speak to them about you know building this company brain, which is where their users started to go towards, I realize, you know, at the moment, the with they they said, look, we rolled these things out, and there isn't much adoption because people are so busy doing their day job, and it's a fast-moving company, right? That they haven't had time to go, okay, how do I reinvent how this works? Basically, yeah, but how do I re-manage it? So I think that particular company is starting to actually say, okay, we need someone to come in and own a new way of working or the process and start to get people to adopt it versus relying on individuals. That was one of the key observations that I had. And this is a progressive company. So then you go to the companies where the senior leadership is still not bought in, they read the headlines, um, they they kind of bring in someone that says, Well, we need to do AI. And perhaps they don't understand also the costs that are going to be associated, because even you know, the savvy people who are using this stuff are paying through the you know, nose by not setting it up correctly. So I think there is a those are kind of a couple of reasons where people are not seeing the return on investment and maybe the adoption curve. Um, I want to say there's resistance, but there hasn't been that much of a demand right now. But Claude is changing things, so that's a different topic. And I think um energetic stuff that are changing this year.

SPEAKER_00 12:52

Is this your favorite uh LLM company? Um, or do you use many of them at the same time?

SPEAKER_02 12:58

I use many of them at the same time, yeah. I'm constantly switching between things. Um, and I think Claude is great at certain things, and I I still I started with Chat GPT, so there's a lot of history and context there. I switch back between, you know, I think I'm using that a little bit on more on some of the thinking stuff, whereas Claude has been more on the practical stuff. I use Gemini um for a little bit more of the deep research on perplexity. I haven't done much of that. Um so between these three and notebook LM, you know, LM are kind of the things I'm using at the moment.

SPEAKER_00 13:34

Yeah, I know it makes sense. I I I echo your um preferences for me, Chat GPT is like if we want if we want speed, perplexity, if we want accuracy uh in terms of citations, cloud does almost everything. We have a number of 20x subscriptions because between cloud design, cloud codes, co-work, we're burning through tokens. And this is to your point of like figuring out how to um use the tokens effectively because you burn through them quite a bit. So kind of like you know, keeping in mind the context window, compacting conversations, like like yeah, yes, memory is important, like context is important, but like you need to have just the right amount before you can uh uh create uh context noise and uh actually deter um uh deteriorate the response. Yeah, yeah, absolutely. It's very big. So somewhat related question. Um, and and you talked about leadership and and kind of their the their buy-in in terms of uh um nation technology. What's your perspective on how the role of a founder or CEO or senior leader will evolve in the future? With the um, well, I mean, AI is what 75 years old, uh, but it's only been popular in the last four. Um so like with the emergence of AI, um, what what would be the the role of the founder, the role of the CEO?

SPEAKER_02 15:12

The role of the founder CEO. Well, I think depends on the stages of the company, right? So I think you know, zero to one stage. The the great thing is now you can do a lot more than you know you could before, whether it is to get the idea and concept and prototype and take it as far, or do the marketing parts, or you know, everything that you kind of want it always where like I just need someone to help me out with you know this, this piece and that piece, all the blockers. So, and then there's the productivity side of like, of course, if you're managing so many different things, um, you know, set up a bit of a core work, or you know, some agency work which manage your emails, all of those things really, really help. So, in that respect, it's super helpful in in that. Um, I think as the companies obviously are growing, that your role starts to change, and so your the tools, what they do for you can be adapted as well. So, again, if it's you wanting to be involved in the product side as a founder or CEO, you know, there's ways to stay in touch without being over the team, uh, understanding what's going on, type of thing. Um, hiring is just one of the key areas that we're also looking at, you know, building out supporting internal hiring is something that people spend a lot of time on, especially when you're scaling. Uh, in fact, the it's funny when you kind of speak to or hear most scale-ups or later stage startups or more mature startups, you're like, What do you what do you do most of the time? They're like, We're sitting there hiring people, and that's become my job. Um, so there's a lot of things around that space. Um, like people are building out chief staff type uh automation workflows for things. So, how does it change? I think, in principle, like anything else, it should allow you to focus on areas where uh perhaps you can add the most value without getting stuck into the administrative and operational bits that you may want to just not do anymore. Um, the more senior, of course, you were and you had money, you may have had an assistant, but now I guess everyone could get such support. So that's yeah, on a little bit of a snapshot, I guess.

SPEAKER_00 17:34

Yeah, it makes sense. I mean, our hypothesis is that we're going to see um AI play a vital role in decision making because it's uh well it's it's just compute, right? So it it takes away away the emotion out of decisions, and emotion is important, so that's why um the role of the CEO in from our perspective is one that they they need to inspire, they need to lay the vision, they need to empathize, they need to you know uh drive the culture. Uh that's not going to be the domain of AI, but in terms of decision making, driving clarity, AI is amazing in this. I mean, you talked about REC. Um you can pull information from a lot of different places to assemble an answer. That's uh very precise. Um I in that vein, I'm would love to get your perspective around uh based on your experience. Um you you talked about some of the use cases and automation, kind of like where it drives ROI, but like what are the biggest operational blind spots that you consistently see inside growing companies and what are the best use cases that companies should go and adopt first, be it hard or hiring or anything else?

SPEAKER_02 18:54

Um interesting. So I guess I guess one of the biggest things that is challenging is the nature of data and where they live and all those kind of things. So if you want to leverage AI, that's is something that you need to work towards from an infrastructure perspective. If the tools is kind of more on the um also a learning care for the workforce, we talked about this, but I guess uh the some of the biggest problems have been the fact that you can't really get so far, people haven't managed to get the terabytes or petabytes of data that is in the organization in a way that they can leverage it. So the models are there, right? The agency tooling is there. Uh, people have to obviously work on the use cases for their companies. You know, some might want to improve the engineering, you know, um, throughput, for example. Um, so the real challenge is how do you get all of this into a structure that you can leverage the right? Again, we talked about this, which is the context, uh making sure that you don't the costs do not go sky high. I'm sure if you're on LinkedIn now, the current trend over the last month is oh, we just spent 150k uh in in tokens, we we could have paid a 5k analyst, you know, to do the job, right? So the the challenge is is, and I would say one of the key things is the infrastructure bits. Uh we have looked at this a lot, and hence why we said, oh, every company needs a brain. But even then, for me for us, when we're pitching it to people, they're still not quite getting the challenge. So I would say that is for me where the biggest kind of blocker would be. Because once that is in place, you can get a group of people who are highly creative, understand a business, start developing things, and others go, oh, that's amazing. I could use that, for example.

SPEAKER_00 21:02

So, from a context perspective, um, at least what we're seeing is a bit of a dichotomy. On one hand, we we see um companies that they would have somebody that's like deeply passionate. Like um, recently we were working with a company, we had a business person, not even a developer, um, really passionate, had learned how to do Python coding, had adopted LLMs, and it was spreading this these scripts that essentially were doing a lot of interesting work. Um, but that's not common. Uh uh more often than not, we're seeing companies that uh they're just kind of in analysis paralysis, they don't know where to start. So maybe this is a good opportunity for you to jump in and and and talk about uh team CMV, uh kind of like what you're doing there, um, as well. As a co-user, just to contextualize the conversation.

SPEAKER_02 22:03

Yeah, yeah. So Team CV, I mean, Central Mind Ventures is the CME part of it. So we from about 2020, we started to do some venture development for ourselves because previously had kind of worked as I said with other organizations. And we had created, so we had spun out an ed tech, it was called Super Savvy Education. We had done the joint venture with Shell, for example. And we our core aim is always to look at the current technology that's coming out and find use cases a little bit ahead of time, which is 100% risk, not even 99% risk, because the reality is A, you have to work with a technology that's moving fast, and the market is often not ready. But if you can get it right first, that is probably the goal. So we this is what we've been doing, and uh the since uh about 2025, we focused a lot more on uh AI memory and experimented and content management, but then we came up with the notion that every company needs their own brain, and the brain being you got a whole bunch of unstructured data in the organization, and you need a way for it to be mapped and understood. And this is not RAG because RAG doesn't work, so how do you solve it? So this is what Go User 25 was. We started some design partners in different fields, different use cases. Some people had very unstructured data with phone calls because they were serving customers. Uh, others, as I said, had you know terabytes or did they have petabytes of data? I can't remember. And so what they're struggling with is saying, okay, you know, how do we uh in a cost-efficient way, you know, manage and structure this? So very extreme different you know use cases. The first one was just a startup, you know, who has unstructured data, and another one a company was much more mature. Um, and that's that's what we're where Go User uh has been focused on. Now, more recently, over the last say the turn of the year, once we saw the agency uh workflow, uh agency kind of um tools coming out, niche down because actually we found it quite difficult, if I'm honest, in trying to get companies to understand because we would have to go in and really do the forward deploy, not just strategy. I mean, not just the engineering, but the strategy because it's like the companies are not there yet. Perhaps what we're talking about is in about six months' time, and I noticed why combinates are actually for this year's this current batch. You will see AI brain for companies as one of the things they're looking out for, right? So it's it's just coming out as a as a theme. Memory, AI memory was there as a theme, but brain and how you you know, you know, structure it. So this is where we kind of started to kind of focus on, as I said, but more recently we're starting to verticalize it because we said, okay, this is very broad, we have very broad use cases, we're we're a small team. And what we're now positioning where the brain fits in is building like AI native systems that could run themselves as a collaborator in an organization versus a tool. Um, and at the moment we're experimenting in the space of um kind of the um the people services within an organization. Um, so everything that handles you know hiring all the way out to onboarding and beyond, uh, also workforce planning, etc. Uh, because actually we think that having an autonomous uh kind of a system that you don't need an ATS system because the current model is you have a human operator and a tool, and we're getting to the point that actually you can have uh a system effort with one or two people managing the full department, basically. Uh, and this is a value for a lot of companies who just simply cannot afford to have an expansive um team, for example. So you always get the founder operators or small business owners doing this, uh, all the way to scale-ups who may be growing, but you really have under pressure one or two people trying to do the work of 10 or 20 people. Um, so our evolution is still our core still brain, because the whole point about the motive this for us is every company is unique, they have a footprint, they have a way of doing things, um, and you need to learn that in order for everything else to become valuable or useful uh to it. So that's the gist of kind of what we're doing.

SPEAKER_00 26:43

So and I couldn't agree more with you. I mean, we um when we uh thought about AI brand for companies, that was I don't know, 18, 24 months ago, and we adopted a tool called Tanker. So Tanker was pulling conversation from WhatsApp, from teams, from different conversations, text messages, um information sitting on on drives and things like that. Um, but big companies or enterprise companies typically they were approaching this by building data lakes. And so um, how would you recommend companies think about AI brain and kind of creating these institutional memories, institutional knowledge, and architecting retrieval method generation systems, kind of like what should be their approach? Because to your point, precision comes from having access to the right information, the right information be labeled correctly. Um, so um uh help help the viewers and listeners um approach this from the right perspective.

SPEAKER_02 27:48

Um, one of the key things that we took away when we started this and looked at scores in the market, whether it was your Mem Zeros and um uh quite a few other tools that or solutions out in the market, um they what we realized is that actually what matters most is uh an understanding. There needs to be kind of an understanding of the business and the data, right? So not even going down to the labeling of data, but what is this business? What is operation, what is the context of this data, what are the things that need to happen. Um, and when you create that, right? Um uh when you kind of create the mapping and understanding of what this organization, how it works out, et cetera, then the data makes sense, right? Because what a lot of uh engineers would do uh with RAG is just pump files in. There you go, just get in there, and you just kind of take chunks here without understanding any you know notion of what's you know what this is, what's it for, etc. So the notion of a brain. So to your point, when we talked about to the very big organization, we weren't interested in being involved in the data, right? For them, for our we said, look, we're gonna have to be a layer that effectively is almost like a librarian that's gonna map what's what was happening, etc. So that when we get it, then we could say, because they're already they're using for brain anyway. So if you aren't using that, you can have a different version of the product where it comes with the model, for example, or any model you choose. Um, but they are also, you know, every uh SaaS tool that it had will have an MCP, right? And you know, what does it do? What does Claude do? Is he gets a question and goes, okay, let me go to uh Atlassian, let me go to here, let me like it's going all over the place. Like so the cost, the tool can use it, etc. So I was like, for those big organizations, what you need is mapping in this way that kind of can guide the brain. It's like, hey, don't waste your time if someone's asking, I don't know, what is the holiday policy for this or how many days we got off? Like, just go here, right? So that's quite a different approach. Um, yeah, you can a lot of people could build internally. And what we also realized is, and the reason we've gone verticalized into a solution is a lot of big organizations will want open source, not because they don't trust you. Um, just historically, when I've looked at it, and I've been on the other side, it's only because if you go out of business or you get taken over, right? They need something, right? Uh, as there. So, you know, there's a lot of developer tools, a lot of open source stuff. Um, and so what what the advice I would give those organizations is bring in some of these specialists, you need to have a notion, don't just use these systems blindly, and it will work at really much better, right? It will help the models. Uh, you can have and you can even have your own small model, um, which is you know advisable as well in terms of costs. Um, yeah, I don't know if that answers your question, but yeah, that's kind of what no, it it it it it does.

SPEAKER_00 31:13

I mean, this is very much consistent with our experience. We would uh bring a small model on-prem for because for some calculations, for some computations, it's it's way better, more efficient, depending on kind of what we're architecting. Uh, sometimes intentionally we would disconnect MCP connectors because uh it it consumes it consumes tokens, right? And and so and and it can create context noise. Um, so uh like you joke about like spending 150,000 uh dollars in tokens, but you could have hired five thousand dollars. I don't know. It it's true. Like if you don't control, if you don't architect that you can consume a lot of tokens very quickly.

SPEAKER_02 31:54

Yeah, yeah. And I think that's it. I think what you said is if you don't architect, and perhaps the one thing I would add is I think we're perhaps we are often um coming from the to these uh kind of problems from an engineering mindset, and we're not coming to this from a service design mindset, yeah. Right. Um, and the only reason I can say that is because I lived both worlds, right? I started as an engineer and then we went to kind of more design thinking. Yeah, and so when I am in these meetings, like you said, and they're saying we're doing this, we're doing this, and you you know they you mentioned you unplugging stuff. That's what uh these organizations are doing. There's someone sitting there and going, no, no, no, no, no, let me just physically like you know, unplug something so it doesn't look in here and goes there, and all these manual workarounds. And uh when we sit through this, I said, listen, guys, why don't we just work out five critical use cases because your teams aren't thinking about it? Let's work it out, design what that service is, right? They want to achieve, and then let's go into the architecture piece, right? I think that is probably one of the things that I'm not seeing as enough, if I'm honest.

SPEAKER_00 33:12

Why?

SPEAKER_02 33:13

I'm always curious as to why, what's the root cause because engine because engineering leads this uh often, it's kind of an engineering-led or a technology-led um initiative. Um, and so even so you see the trends even with designers now knowing what the hell do we do in an organization, right? Because it's like, whoa, look, look at all these outputs. Let's just build first. And now designers are you know catching up and standing around saying, well, maybe you should do this. Yeah, I think there will be a correction. Um product managers are trying to understand. So suddenly everyone's doing everyone else's jobs. So product managers are suddenly prototyping, creating code. Designers are saying, Oh, you guys don't create the front now? Like, you know, now you're responsible to cut. So we're in a world, I think you know it will change. Uh, I don't foresee this kind of noise um happening. Uh, either roles will be completely redrawn. And so actually, here's different types of roles. Two types of roles in the company, uh, product and UX and everything merges into one thing, right? One one person has to do all of it. And which is, you know, given your gray hairs and my gray hairs, uh, I can say this, which is we've gone through these cycles before. We've gone through back in the day where uh the front-end engineer was the designer, right? If you recall, like you know, there was never a designer, it was like, oh, you do front-end. Oh, so you have to do all of it, basically, right?

SPEAKER_00 34:43

Um, well, uh, and and the tools have enabled much of this. I mean, that the the the typical process before that exists in terms of like how you develop, in terms of you have a business analyst, getting the business requirements, the create uh like creating all the documentation, passing on to a technicalist, building a technical. Like to me, this is has all been compressed. Uh, I mean, even even in my own personal, I'm a CPA by background, I'm a business guy. Yes, I've always been on the technology side, but like like this year, probably 10 to 15 percent of my time is spent in Visual Studio just coding. Um, and so I it for context, uh, there's a term that's when it was with Deloitte, uh, that has really stuck with me, is uh purple unicorns. So it and and the the way that Deloitte refers to this is the is the combination of a person that's both business and technical. So I find that this will be continuously an increasing requirement of having people that have both technology predilection if they're not if they don't have the technology to depth, at least they have technology predilection and deep business acumen. Because you to your point, you have to start with the end goal in mind. What are you trying to solve for? And then you can you can make the technology do anything and everything, but you need to understand what you're trying to solve for.

SPEAKER_02 36:06

Yeah, yeah, 100% agree. Uh, I was one of the best engineers I ever worked with. Always had the attitude of like a when I was in a meeting, and then someone like a designer or product manager would sit in the meeting and ask if something is possible. And I know exactly what that person's gonna say. It's like everything is possible, like it depends how much time we have, what do we have to do, what our constraints are, but you can build anything in principle. Yeah, so you're right. I think that it the art now is that there's something that's going to be timeless, regardless of technology, which is how good are you identifying a real problem that you can execute towards really well, and then have the ability to just build such a good relationship with that customer that even if there's going to be other competition that you know, those things I think are just gonna be around forever. Um, you know, this whole thing about I think we hear a lot now about distribution being the moat or or at least as critical. Yeah, but I think the one thing that people don't then talk about is the other side, which is once you have the customers, what is absolutely critical is back to anything that becomes um you know not uniform, you know, because uh when competition you know can catch up so quickly, in the end, you customer experience like is gonna still be the one, right? Because what's the difference? Why would I deal with you? I love what I'm working with you guys. Why would I suddenly swap someone else? Right, yeah. So again, I think we see these principles are never gonna change, but the tooling and the processes um maybe the percent personnel might be changing.

SPEAKER_00 37:50

Completely agree. Um, from an ROI perspective, like for context, I'm based out of Calgary in Canada, and Calgary is like really the oil and gas capital of Canada. Like our border in many ways is like Texas. Uh, it's the richest province in Canada, uh, very rich on natural resources. What are the highest ROI cases that you have seen in uh oil and gas?

SPEAKER_02 38:19

Oh I can't. Well, I've got to dig back now into the shell years. Um I don't know. I I'll be honest, I don't think I have a factual answer. The only things we're talking about from a technology perspective, or at least the type of technology, not like drilling technology. Um I guess the the key thing that, and I haven't, I should, I've kept up with it, is improving things like just the efficiency of getting to a port, as you know, is a huge problem in that space, right? Which is you know, the the the traffic, the mismanagement. And a lot of it is around just inefficiencies of operations that sit within a port. So for your listeners who don't know this, uh, you have ships burning oil to just get to a destination port, let's say you're going from Texas to Walter, let's say, and then they get there and they go, No, no, no, you can't come, you gotta wait here for four days, five days, and you burn all that you know, oil. Because we have an issue, not an issue, but you know, there's so many different parties in a port that don't communicate data, etc. And this was one of the biggest things that they're trying to source. If you improve data sharing, improve use AI for managing all this, you can get people coming in and out and reduce carbon emissions and of course costs for a lot of these things. I think that that would create a lot of ROI for the for the industry. Uh, I'm not yet sure because the industry is so fragmented, you have so many shipping companies, so many within the ports themselves, so many different parties. I don't know if it's solved yet, but I think that would be a huge use case to solve.

SPEAKER_00 40:04

Yeah, uh, no, no, for sure. I mean, this is very true. A couple of years ago, we implemented a system your uh European company called Ocean Smart to deal with this. It's kind of like when do you dock? Like you have a 24-hour window to load products, or you have to be in and out, like kind of like managing the the schedule logistics around this because what you described is a hundred percent spot on. Yeah, absolutely. It's it's complex and it's costly. I um like for a small company it this may not be an issue for for big companies, like the uh it's seven figures that quickly adds up.

SPEAKER_02 40:42

Um just shifting gears a little bit.

SPEAKER_00 40:51

Um what separates a company that successfully modernized their infrastructure from those from from the ones that just talk about transformation, but never actually uh go to execution.

SPEAKER_02 41:06

Um if I'm honest, I mean I don't have data, concrete data. I think your McKinsey's of the world come up with these types of reports for like, oh, you know, if you don't these transformations, look at look look at the lag arts versus but yeah, I think there the only thing I can tie it to is perhaps the companies that are innovative or more innovative or thinking about innovation, invariably will invest in transformation and will invest in infrastructure changes, which are hugely expensive if you're an incumbent company and why it's always very difficult for those guys to play against typically against like startups traditionally like coming in with a new tech stack, of course, with AI is different. Um, the only thing that I would say is that um the question for me may not be just infrastructure, is there is their mindset and a strategy for innovation? And I think that leads to the correlation, in my opinion. And as we know, then the reference point is easier. Companies who innovate versus the ones who stand still, you know, you class everyone talks about your Kodaks and your blockbusters and everything. But that those are examples of companies who didn't invest in infrastructure, for example, right? Uh, in the case of let's say blockbusters, you know, and I think they went late into it. Uh, they could have bought Netflix and all those things. These are stories everyone knows about, but I think it goes to your point. It's it's not a it's not a technology uh question. Again, your technology will say we can do anything as long as we know them, as long as you give us the budget, the remit and the timelines, we can do it.

SPEAKER_00 42:51

Um I I wouldn't mind just double-clicking on those two companies and getting your perspective because I mean, in the case of Kodak, Kodak invented the digital camera. In the case of Blockbuster, they actually were working on a competing product to Netflix, and that's why they they refused to buy or decided to pass on. So it's not like they were blind, um, they clearly had the foresight to invest in innovation and and RD. What do you think prevented these companies from actually seizing the moment you talked about distribution, we talked about velocities, the modes, kind of like if you if you parlay this into the current era, kind of like what should companies do or how should they think about innovation, RD and position themselves for the future?

SPEAKER_02 43:41

Be forgiving of the CEO, I would say. A shareholder challenge for a lot of the boards. So for most companies, especially if they're publicly traded or very big, as you know, the board really is there to make sure that shareholders get their return. Um and transformation is slow. Like you know, building infrastructure traditionally is slow. There's a lot of risk to it. What's the game? Like we swapped this in, you know. Uh the I kind of we did one thing at one of the companies that I was supporting where we went from you know uh women into AWS. This is back in the day, like where everyone was just like hosted. Um not everyone was on cloud, right? So I was like, let's let's do it now. To justify that to a CEO, like they're like, yeah, and so what do we make more money? Do we you know it goes down to the that lens, right? Do we make more money? Do we save money? You know, what what lens it is? And in this case, actually, we save quite a lot in the long term. So that's how we got it through. So to answer your question, I think people are resistant to it because they don't want to be the one that broke a 50-year-old, 100-year-old company on their watch. He went off and did it.

SPEAKER_00 45:17

Mark Zuckerberg can afford to spend, I don't know, 20 billion on uh the virtual world metaverse, yeah, yeah.

SPEAKER_02 45:26

You know, they and and still be there. Um, not every CEO, not every board has that. So when I said you have to be forgiving, you have to bring in a CEO who has the mindset, maybe the experience of taking a company, right? So if you are a uh a company in a traditional industry, the advice I would give is try to hire a CEO who comes from a different industry or from a technology industry that help you move forward. Have a capable COO who knows how the current business works, and together, you know, you know, let them drive it basically. Um, that's that's my two cents.

SPEAKER_00 46:06

Amazing.

SPEAKER_02 46:07

Um, Shaheen, any any parting thoughts of wisdom for uh founders and companies that they're looking to scale and um uh yeah, um just I guess keep on experimenting, but I guess it's um experimenting, you know. Um with the word escapes me, right? Do the experimentation with a view to try to roll it out, not just for the sake of it, right? Um, and I think probably getting your hiring right for this is very important. Um, getting the right people in the company will help you uh with that. Um, because in the end, I think what we probably to summarize what we all we talked about is we think that the technology is not going to be the problem per se, but you need the right people, the right strategy, the right kind of culture to be able to do it. So I think that's maybe the last takeaway from this.

SPEAKER_00 47:08

That's awesome. Love the conversation. Where should uh our viewers and listeners follow you? How should they get in touch with you?

SPEAKER_02 47:17

I am absolutely awful at uh talking about the work, but I will maybe send you some links that we can share.

SPEAKER_00 47:24

Okay, we'll drop them in the show notes. Uh Shaquin, it was an absolute pleasure having you on the podcast.

SPEAKER_02 47:31

Thank you for having me, Maxim, and great to speak to you.

SPEAKER_00 47:34

My pleasure.