Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Data Analytics Chat

Data Analytics Chat artwork

From Legacy to AI: The Playbook for Transformation

Sep 22, 202645 minAudio
ShareXLinkedInFacebookWhatsAppEmail

From Legacy to AI: The Playbook for Transformation In this episode of Data Analytics Chat, host Ben Parker interviews Echeyde Cubillo, a veteran technologist with experience from eBay to multiple startups, now leading AI transformations in enterprises. They explore why large organizations struggle with AI adoption despite greater resources than startups. The real barrier isn't outdated technology - it's legacy thinking, hierarchies, and silos. Echeyde reveals that small, empowered teams connected directly to customers outperform massive structures. The conversation offers a practical playbook: flatten organizations, bring entrepreneurial talent into leadership, focus on builders over managers, and embrace experimentation. Essential insights for leaders navigating digital transformation in the AI era. Question Timestamps 01:47 - Before we dive in, could you give listeners a quick introduction to who you are and the transformation work you're leading today? 03:31 - Large organisations often have more resources than start-ups, so why do they frequently find transformation much harder? 11:50 - When we talk about legacy, is outdated technology really the biggest problem, or are legacy thinking, processes and incentives even harder to change? 15:24 - What does an entrepreneurial mindset actually look like inside a large organisation? 19:03 - How can leaders create greater speed, ownership and experimentation without losing the governance and controls a large organisation needs? 23:28 - When an organisation has dozens of potential AI opportunities, how should it decide where to start? 25:22 - Do companies need to modernise their entire technology and data environment first, or can they begin delivering AI value while legacy systems are still in place? 28:49 - How can organisations move away from lengthy transformation programmes and start testing ideas quickly, learning from users and proving value early? 30:44 - Transformation can challenge existing roles, budgets and ways of working. How should leaders bring employees and stakeholders with them? 34:41 - Many organisations can launch successful pilots. What prevents those pilots from becoming scaled, business-critical solutions? 36:28 - How do companies make continuous innovation part of how the organisation operates, rather than treating transformation as a one-off programme? 39:43 - To bring the playbook together, what are the most important steps leaders should remember when moving from a legacy environment into the AI era? 43:23 - If a leadership team wanted to begin this journey over the next 90 days, what is the first meaningful action you would encourage them to take? 44:30 - Finally, for anyone listening who would like to continue the conversation with you, where is the best place to connect? Connect with guest: https://www.linkedin.com/in/echeyde/ Connect with host: https://www.linkedin.com/in/ben---parker/ Thank you for listening!

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

echeydeby the very definition that a company got to grow big, the entrepreneurial spirit, if it's not founder-led, and I tend to be binary, tends to have gone out, right? And in the world of AI, the reason I made that statement is that in the world of AI, there is no substitute to coding. A team of five or ten people more often than not can do more than a hundred, two hundred, a thousand people. I think to me the, in a larger company, the number of layers from the CEO down to the builders, it should be really, two to three layers, right?

echeydeUm, you have seven, ten layers. It's crazy. It's crazy. Every time you go up a layer, you lose so much context and so much innovation can be stifled, just collapse down. The, the thing about organization is that when there is friction in an org, the org will, will auto create more teams, more bureaucracy with the intent to remove the friction, but in fact it's actually adding even more friction

benHello, everybody. I'm your host, Ben Parker, and welcome back to Data Analytics Chat. I took a short break, and now we're ready t-for some more exciting conversations across the world of data and AI. Today, we're looking at a challenge facing almost every large organization. How do you transform for the AI era when you're still operating in a legacy environment? In today's episode, we're going to build a practical playbook for moving from legacy to AI and ex-explore why adopting a more entrepreneurial mindset could be the key to making that transformation successful.

benUh, joining me today is Che Cabill, and he's got a broad background in enterprise and startup world. Do you wanna give the listeners, Che, a quick int-introduction to who you are and the work you've been leading?

echeydeYeah. First of all thanks for inviting me, Ben. I'm really grateful for this opportunity and, uh, this is a very hot topic right now, all over tech. So just to give a quick background I've been a technologist developing software my entire life, for decades now. I had the opportunity to, um, and, uh, and the, the fortune to start early during the internet phases with eBay, which was a great run where we invented technology, and also it gave me the, the financial freedom to start, experimenting about creating companies.

echeydeSome of them did well, others not so well. But at the end of the day, it's all about how you correct and you learn how to build businesses. I created two companies, one that was sold... co-founded, definitely. Uh, one that was sold to Yahoo! In the early days, I was an early competitor of YouTube and then created another SaaS company in a transactional payment space that was developing ticketing application for, you know, zoos, aquarium, gardens, and the likes for about a decade. That was a lot of fun. And since then, exited both of them in good terms, and I've been doing a lot of deep AI, helping larger orgs now recently.

echeydeSo I'm pretty familiar with the concept of transforming companies from the inside out on, on, on larger corporation. It's been quite a journey and le- let's have a chat about what that means, how to go get it done, what works, what doesn't work, what needs to change, and, um, let's go through, through your life too then.

benYeah, no, I mean, we're definitely going for s- a, like, fascinating period we're going through now. So it's, um, yeah, let's dive straight in. So I guess, I mean, why do large organizations I guess, look, our large organizations have got lots of resources, but I guess why do they frequently find transformation much harder than startups?

echeydeGood question. So look, b- by, by the very definition that a company got to grow big, the entrepreneurial spirit, if it's not founder-led, and I tend to be binary, tends to have gone out, right? And in the world of AI, the reason I made that statement is that in the world of AI, there is no substitute to coding. The technology is moving so fast that the leaders that need to implement change cannot just judge AI on the basis of reading the press You know, or changing from A to B, there is, you know, advancement every single day.

echeydeIt goes so fast nobody can keep up. You gotta bring the, the coding aspect in the technical organization with the ability to connect back to the customer. So that is why it is difficult for large companies. Most large companies have a pretty many layers of management that in general, not because of the way it's been done for decades, have disconnected from the, the code base, from the coding to some extent with the customer base as well.

echeydeAnd therefore it's extremely important to bring back the hands-on aspect to tinker, to experiment, and to develop your own opinion about how good AI is for your own business, for your own customers, as opposed to kind of like looking at it from the standpoint of, you know, conferences and reading materials and so forth. Yeah. So I cannot emphasize how when there is a deep technology changes, we go through waves, and AI is one of, in my opinion, the biggest waves I've ever seen.

echeydeYou gotta stay extremely hands-on to understand where to go for the needs of your customers as it relates to changing the product definition that you currently have. And the only way to do it, you gotta be extremely close to the way how AI is being built. It's pretty getting there because you can code much faster nowadays. AI is doing a lot of coding for you, and you can figure it out sooner for your organizations. My view is that there's no substitute being hands-on, which is a little bit a paradox for larger companies.

echeydeSo the larger companies have to understand how to focus, emphasize on actual builders that are doing the work, and not so much on the management side of the business, which is basically being disrupted. If that answers your question, Ben.

benYeah, no. See, I guess there's, there's a lot more layers, isn't there, to like, obviously enterprise level, the communication. Obviously, there's more people. You know, and I mean, everybody in business, like, they wanna get A to B, don't they? It's that squiggly bit in the middle, like the change management element is so, so much more challenging when you got... if you're dealing with a larger company, isn't it?

echeydeCorrect. Right. There's a lot, um, whatever I just said, there's obviously a lot to unpack to unpack. I just wanted to now highlight the, the, the conclusion, which is becoming hands-on. So yes. I mean, let's talk about, let's unpack this a little bit. So there's a couple of, couple of elements to it. One of them is how you're gonna change the product definition, right? There's the, the second is how you change the scale of the organization, the people, the management layers and so on and so forth. So As it relates to the product definition, you know, always lead with the product. Without a good product, there's nothing to talk about'cause the customers are gonna walk away.

echeydeSo as it relates to the product definition, the biggest shock for big company is that most of the product definitions that are in place today for the legacy incumbent... Now, software companies, they tend to be what I call form based. Classical, web forms been around for years. You go from... You have a wizard, you go from a link to another link, a form to a form to a form. So ninety-nine percent of the definition is there. The customers are used to those behaviors.

echeydeSo what happens is that in the world of AI, forms are kind of gone, right? The user interface is conversational. You ask something to a piece of software that people are calling AI, and the system goes back, and you engage into a multi-turn conversation. So for the most part, if you look at, Claude, ChatGPT, very simple, almost as when Google came out of the gate, super simple input field, which is you ask something and it responds, right? So that kind of user interface puts the old user interface out of the door, right?

echeydeAnd that's the opportunity for disruption for startups. For incumbent, they're kind of like, wow, being caught flat-footed, which is, oof, I've got all of that definition of the product that I've been building for decades, but the user experience has to fundamentally change, right? So there is two aspect to it. You can be very timid, a little assistant, and you say you've done AI. In general, the customer base is not gonna care that much'cause we are not leveraging so much AI as it, as it needs to be done. You have to have real boldness to be able to replace the definition of the product, something that is much more conversational.

echeydeIt can be a brand-new upgrade to a new user experience. You could stand up a new product that competes with your own legacy product, let it grow. It's your own company, it's your own customer base, and find a way to convert from old product to new product. Many ways to get that done. But at the end of the day, on the product user experience, you gotta make radical changes, right? And that's very difficult to do for a large company because they're afraid that they're gonna alienate their own customer base, and they're gonna take a slow roll, right?

echeydeWith a slow roll, the problem, we invite other competitors To kind of fill the gap while the company, the bigger company is getting slow. So you absolutely, those big companies need to be a lot bolder. They have to be willing to take bigger risk, right? That's number one. Now, you start there. Once you start there on the product and you get to that realization, look, the user experience has to be much more native AI. You need to redo a lot of the user experience heavier for the technical star, technical stack underneath.

echeydeThen it begets the second question, well, how are you gonna deal with the organization? Now, the organization is built around the past. Roles are now changing. Individual contributors can do multiple roles in one person and go extremely fast. So what happens to the designer? What happens to QA? What happens to the product manager? So the companies really need to think that the old world of, you know, slowness, planning, roadmap, and so forth no longer works that well. Things are moving very fast.

echeydeAnd when things move fast on the tech side and the startups are out there just nipping at the side and peeling away part of your business, you have no choice but to simplify that organization, right? And it's easy to say, look, when you have thousands of people, it's complex. But in the face of complexity, you gotta go back to basics. And back to basics is tight, small, lean teams connected directly to the customers, and they can move very fast. What does it mean? It means a lot of the corporate layers that have been, you know, put into place, project management, this and this and that, you really need to find a way to streamline that, right?

echeydeSo I always tell these larger companies, and that's how, you know, I try to lead by example. Create a small team super small, extremely connected to the customer base on a daily interaction of the customer base to go get the feedback and move super fast. All of the matrix and the, like, planning, dependency, and all of that, you put it out the way. You need to move very quickly because there is so much unknown in the space. But you have to go back to what software has always been ever since the birth of software, small, tight team, no fragmentation of the ownership.

echeydeAvoid the, you know, extensive yearly roadmap planning. Those things are basically not working anymore. Tight team, fully empowered end to end, closely connected to the customer, then you can deliver value because you're getting real feedback from the customer base, and you can iterate

benquickly. So do you see it's more... Is it out-- The biggest problem, is it the outdated technology, or do you think it's more, the old way of thinking or legacy thinking? 'Cause obviously the processes have been one way, now moving to this new AI adoption way. Do you think it's just- I, I think it, it- do you think it's more the technology or do you think it's, like, the legacy thinking?

echeydeIt's mostly the legacy thinking. Like, the concept of statues, the concept of seniority, the concept of entitlement. You gotta get it out the door. Uh, gotta get out the door, right? The only thing that matter is that what can you do, right? In AI, you learn as by doing. You don't learn because you have a senior title. You probably don't know-- If you have a senior title and you read about AI, you know less than a kid coming out of college that is com- coding out of AI that can be 10 times better than you. Mm-hmm. So once you realize that, you're gonna be real that put aside the entitlement, put aside the seniority, just focus on what you can contribute, what you can deliver, right?

echeydeIt's going back to raw mindset of delivering. Two things matter: build the product and sell the product, right? You have to co- This is what I keep saying is that collapse the management layers, flatten up the organization as much as possible to get rid of the legacy thinking, right? Because legacy thinking means seniority, entitlement, posture, power, and all these things. None of that matters in this new world, right? You gotta run on the fly. You gotta cut that out. You gotta get out and get it out there. Just to make my point and to prove you, look at a company like, you know, Anthropic.

echeydeRight. Anthropic annualized run rate of probably fifty, sixty billion dollar in less than four years. It's bigger than most of those legacy companies out there. Let's get real. Sixty billion dollar in less than four years. It's a big company. Guess what? It's a flat company, right? You have CTOs from larger companies that go join Anthropic. They become MTS, member of technical staff. A sixty billion dollar revenue runway company is fundamentally flat. They are winning. That should be what you want to copy-

benYeah

echeydeto get a chance to not be disrupted. So a-and that's very, very difficult for companies to do because, you know, a lot of them have been there for a very, you know, they cannot flat down they climb the corporate ladder, but the corporate ladder in the wor-world of AI, I hate to say it, you know, you're climbing your level of incompetence. You gotta go back hands-on. You gotta go back. Super close to the customer, close to the product. And in a world where coding is so much easier, get into the code, right? Uh, it's just... That's what startups do. I see CEOs of larger company like Shopify, 150 billion dollar valuation company.

echeydeThe, the CEO is still coding to this day. He codes AI just to understand what are we talking about, so that it doesn't come from a line of managers that probably haven't done any AI other than read. So he gets in the middle of it, and he's gonna understand AI at Shopify to move forward the right way, right? That is a transformation that will work for Shopify. For others, you gotta make these drastic changes in-inside your organization. It doesn't mean people go. It means become an individual contributor, whatever it may be.

echeydeKeep the right people, but you gotta have that mindset. It's very simple. The alternative for me is that I see the... You can get disrupted. There's many companies that spend years and billions of development in AI preha- There's not much to show for, right? So it's real. You know, I think we've said only the pioneer will survive. You've gotta go back to operating a startup within a larger company, right? If you're not founder led. That's a very difficult thing to do.

benYeah.

echeydeAt least- And let's, let's unpack that then.

benYeah. So I mean, ob-look, obviously, look, startups, yeah, have got the, that mindset. So I guess how can we... I guess how can you embed that in a larger organization then, this sort of entrepreneurial mindset?

echeydeVery simple. You bring entrepreneurial talent into the house because probably a lot of it is gone and not there. Unless you are founder led. Founder led large companies, you know, CEO, like Shopify, Facebook and so forth, they understand that very well because we started from nothing to building a large company, and we understand that inevitably as you scale, you need to bring big tech folks, but they mix them, they mix them up with, uh, either we hire or we hire small companies for which we don't want the product, but we want the talent, right? And then we give them positions of power.

echeydeNot that these guys care about power, but so that they can make the right changes inside the company. So if you are a CEO, and you have an organization where you have a lot of, quote-unquote, caretaker by the nature of, you know, what legacy did, bring entrepreneurial talent and empower them, and let them be, right? Because they're gonna do things that are completely outside of the processes that you've, built over decades. That is the onl-, in my opinion- That is the only way that I can see work, right? And some companies, especially the founder-leds- Mm-hmm are really good at doing that.

echeydeYou need to replicate that into other organization where the founders are gone, to bring that entrepreneurial spirit back in as uncomfortable as it might be.

benYeah. I mean, so, I mean, we see that a lot with, like, the FAANG companies, like, I mean, a lot, a lot of the talent now are going to the other enterprise companies and startups because they're looking- Yeah for that. They want ownership, they want more scope. Obviously making more decisions. They're getting in getting in deep with the business, and we're definitely seeing so many more people want that.

echeydeYeah

benobviously, obviously- Yeah salaries across AI has gone, yeah, crazy now, isn't it? Just'cause pe- businesses need this talent. Like you said, they need to embed it in the business'cause, like, these people have been there, done it already, haven't they? Like, it's sort of-

echeydeYeah, yeah

benthey-

echeydeAnd, and, and do... That's right, Ben. And these people in general, they don't care about titles, which is why I always advocate for very flat company and avoid the, you, you know, the side effects of seniority that, you know, can stifle. Forget about the title, just think about the outcome, right? You bring somebody entrepreneurial, that person will call, that person will talk to customers, that person will have a sense of ownership. They will grow the business, and you can measure how they grow the business. You gotta pay them handsomely. There's no doubt about it, right? And it's not about- The, the corporate mindset is about what is the outcome?

echeydeIf your outcome is that you put a product out there, you measure the adoption, you measure the engagement metrics, the repeat, and you as an individual contributor/leader who's very hands-on and moving the needle clearly on those metrics with a tiny team, you're probably doing more than other teams that have 10 times the resources by all means you gotta pay them handsomely. Because we are doing the work of, we are being 10 times, 20, 100 times more efficient. That- that's the world in which we live. And think about it, w- with AI, the AI coding tools are unbelievably good, and they're getting better and better.

echeydeYou can... That's why I don't subscribe to, you know, management that would say,"Oh, we have a legacy technology." That's-- That argument doesn't hold water, right? If you really want, nowadays you can rewrite the code base at a fraction of the cost of what it would have been two years ago. So you gotta be bold, you gotta say,"I'm gonna rewrite the code base, it won't take that long." And after that, the issues may be more on the migration of the data in enterprise and completes more reality on the ground, but you gotta be willing to rewrite large section of your technology stack quickly.

echeydeAnd it's not a, For some people it is their baby, and it, it may be hard, but you gotta do it to survive.

benYeah. So I guess how can leaders create greater speed, ownership, experimentation without losing the governance and controls that a larger organization needs?

echeydeThe, the, at, at the tech level, a lot of the controls can be automated, right? Like, the pipeline to deploy the code nowadays are pretty sophisticated. You can run a lot of tests, you can do a lot of automation on security, like all of the teams that do a ton of like manual review for security and the likes. The, the in truth, you can actually automate a ton of it, especially with the new AI tools. So anything that is technical is almost always a candidate for automation.

echeydeYou gotta be willing to go from somebody who's had a job who's very manual on security or testing and just automate the hell out of it, right? There could be some resistance, and we use those people to bring the, the knowledge, but you g- you, you gotta automate, and those jobs are gonna change more into helping do the automation and so forth. So on the technical side The PII, the security and all that definitely can be automated and, there's a lot of headway being done. On the legal side, it's a little bit more tricky, uh, it's a bit more manual, right?

echeydeBut I would argue if the legal team is really working closely with the kind of security risk people, there is often an opportunity to automate again and again, right? So it's getting that mindset to like,"Okay, my job was manual." Don't fear to lose your job, but make your job automated so that you can do something at a higher level, right? Tough, because there's gonna be some resistance, but that's-- that can be done. That can be done, right. Mm-hmm. The AI tools nowadays are extremely powerful, right?

echeydeOn the dev side, we all know that this is where most of the su-success is. But on the legal side, anything, anytime there is something that is fairly manual, I would just like Do double click on it big time and see where, how much of that can be automated. And you'll be surprised, a lot of it can be automated.

benYeah. So do you think it's more businesses just having that resistance, they just need to let go and just start going more, investing more into AI, automating more of the work basically? 'Cause obviously there's, obviously- Yeah there's so much resistance from firms like, es- especially like if you're in the regulatory space, obviously there's a lot more to deal with, isn't there? And things like that. Obviously, each industry will have a different impact.

echeydeThat's correct. You gotta take the day one mindset, let go. The job that you worked so hard for decades to get to corporate ladder or whatever, day one, rethink it over. And that's difficult. That's what I keep saying is just going back to that entrepreneur. Yeah, do it for you before somebody else is gonna make a decision for you, which is not as comfortable. Do it yourself. Yeah. You're not gonna put yourself out of a job. You're gonna find something even bigger and better to do, but replace yourself with automation. Yeah, no, I definitely think like more roles, obviously it will evolve, and it's, I guess, like I said, there's, you still need humans behind the roles'cause obviously- Yeah AI can do automate,

benit can do great work, but obviously you still need like the thinking behind this, don't you? You need the engine telling what to do and-

echeydeYeah.

benYeah, yeah actually steering it, don't you? You need to, someone still needs to be driving the car sort of thing.

echeydeYeah, of course. Look, I think leaders need to tell folks in a sincere way right now, not in a hypocrite way, that, look, you're replacing yourself, but you're also extending yourself, expanding yourself. Yeah. Like there's always more work that can ever be done because if you are creative, there's always new things to create, new opportunities to go after. So it's important that the, the corporate environment be such that, replace yourself, I'm gonna fire you. No, it's, there needs to be a contract with employees in such a way that they are comfortable to kind of like change the nature of a job to be more automated.

echeydeSo look, employees at the end of the day are gonna, vote with their feet, and they will know if the leaders are, are, are honest, sincere or not. And that's how the world works of, for, you know, of, of all times. I mean, a good company needs to keep that contract, that it doesn't mean that people go. It means that you're gonna do bigger things. It's easy to say, but I think that's just where the world needs to go, hmm. Mm-hmm. 'Cause you expand your c- you expand your capabilities, which means you should be able to grow the business more than what you have done because you have, the brains are being used for something more important.

benYeah.

echeydeMore ideas, more products, more creativity, da, da, da.

benOkay. So w- I guess obviously there's, there's endless opportunity now, isn't there? Um, so I guess where-- when an organization has so many different AI opportunities and, like, every business unit is gonna be yelling out to have their business case, how should it decide where to start?

echeydeGood question. I usually start, you know, when I don't have a good answer or any other business, start with the customer. Start with the customer. If it impacts, you know, that the product is changing, obviously you go you talk to customers, right? Customers don't always know what else needs to be built. They don't really see the future as much. But start with the customers, engage conversation, and just tell them,"This is what I think I can do. What do you think?" Get through a process of validation from zero to one to ten. That's one aspect, the second aspect is if it's outside of the product, we already, covered that topic.

echeydeYou should look at ways to automate the business processes, right? So it's either the product or the internal business processes. On the product, for me, I talk to customers all the time. I exchange with them on a daily basis. I have twenty million ideas. You know, the one I-- I will present a few ideas. A, a lot of these customers will sh-shoot down some of them, but they're gonna harden up. They're gonna, catch on on a few good ones, and then I start to kind of,"Okay, I got a bite on this idea. Let me continue the conversation." And as I unveil it, then I can ask engineers to work with me, and we just go deeper, right?

echeydeBut I personally always start with the customer, always. And I think there is no substitute. If you don't start with the customer base, you cannot create, inward-based organization that may do tech for tech. Tech for the sake of doing tech, which happens a lot, by the way. There's a ton of waste in that respect there. Don't hesitate to pick up the phone and talk to customers. Keep it simple, especially when you are in such a big, uh, transformation phase. They know the frustration of the product. They know, you know, what doesn't work, and then just work with them.

benSo then should companies need to modernize their entire technology and data environment first, or can they begin delivering AI value while legacy systems are still in place?

echeydeI would start on the latter, right? Because i- if you change the whole thing, you're probably not gonna hit, right? You could be off base and take a few core mission critical workflow. That's what I mean, mission critical, not small workflows. If you-- that are not used that much. If you take a small workflow, you're being too timid, time is gonna go by, you're gonna be behind the, behind the eight ball. So take a core mission critical workflow and get it right. Right. The thing about AI is that it's actually very easy to put a prototype and, uh, out there, and you think you're doing great.

echeydeOnce you put it in front of customers, they're probably gonna say,"It doesn't work. The quality is not great. The hallucination is horrible." You know, we've seen that from the chatbots ten years ago to things are getting better now. Mm-hmm. Very low adoption. So you can see you put something out there, it doesn't mean that... Put it out there, work with your customer base, get it right. It may take six months, nine months. Once you get it right, it will inform a lot of the tech stack that needs to be done. Like I said earlier, don't design a tech stack because you have a title or you are reading a paper.

echeydeIt doesn't work that way. Get it from customers, find the issues, iterate quickly, make it right as fast as you can, and take those learnings and then over to the rest of your workflows quickly. And start, like, expanding the transformation on the rest of the tech stack, and then you chop down the architecture, and you start to move the other architecture out, right? And that's how I would do things. As a matter of fact- a lot of the legacy architecture there that are in place, I mean, the amount of rewrite that needs to be done, not at just code base, right?

echeydeAt the architecture level is just very significant. To give you an example, once you have a workflow well done, then most companies, especially in enterprise, they're gonna realize what the legacy mindset was in relational databases. Yeah, you do need relational database for some core transactional workflow and so forth, but most of those new product definition at the user interface level are very unstructured, which means that the rise of unstructured databases is upon us. And that's a fundamental architectural change on the entire organization.

echeydeSo the mix and match of relational SQL to a structure is a big shift happening in the industry with AI. And that changes a lot of their test stack dramatically, not to mention a lot of the roles underneath. So... But like I said, just get a workflow right, learn it, make it good, and then quickly go after other mission critical workflows as quickly as you can- Mm where, while you're essentially doing, um, a tech swap. And the tech swap is faster today than before because you, you can write a ton of code with AI.

echeydeThe harder part, not to underestimate, is plan the migration of the data, right? Your data may be dirty. You got a problem. What do you need to do, you know, to clean up that data, which what needs to be done to be able to do a good migration? That's where it's gonna be a little bit harder to, uh, squeeze the time it takes to go and get it done. Right. But that's how I would approach a, a massive technical transformation within an org, processes and management layers aside. That, that technical side, you, you can go quick but be careful on the migration of the data.

benYeah. I, I mean, I guess our stakeholders are obviously gonna want to see results as soon as possible. So how can organizations move away from lengthy transformation programs and start testing, I guess, their ideas quickly, learning from users and, I guess, providing value as early as possible?

echeydeI said just all of the planning artifacts that have been built over decades that worked for a different era, like, like I said earlier, you kind of need to Move it out the door and say,"Look, the roadmap, the yearly roadmap planning, that stuff is over," right? The pace at which AI is changing technoly- technology-wise so fast that what's gonna be in front of us in one year, we don't really know. So really compress your roadmap planning to be like in months, not even quarters, right? In months, at the most quarters. So compress your roadmap planning to be smaller.

echeydeThe time it takes to get stuff done is already compressed based on AI coding. So just change that and go in front of those customers, like I said, as fast as possible. In a world where development cost is going down, the bottleneck is the process. And the process bottleneck would be lengthy roadmap planning, Lengthy coordination of between the teams. And the way you resolve, resolve that is that don't do roadmap for a long period of time. If you are talking to customers, you know exactly what needs to be done.

echeydeExecute quickly to outcome within a quarter or less. And on the, on the organization matrix, as I said earlier, self-contained team where there is no fragmentation of the ownership, directly custom, uh, connected to the customer base and deliver. Even if you need to go all of those teams, if they belong to different orgs, put them together in a true, you know, mission team, if you will, with one ownership, even though they have different bosses all over the place, deliver that value to the customer as quickly as possible.

benOkay. And then I guess, obviously, transformations, they always challenge firms obviously budgets, n- new ways of thinking. So how can leaders bring employees and stakeholders with them?

echeydeOn the tech side Engineers, PMs will be happy because things are gonna go faster. As I said earlier, the, the problem is that when things change, m-m-- human nature is to have fear. Fear that, you know, whatever they worked, their entitlement and seniority and all that may just go out the door. So it's important to let them know that, look, I mean, at least your, your comp stays more or less the same, but your role may change. You may go from manager to individual contributor. You-- your job is gonna change quite a bit.

echeydeWhat you did manually has to be automated. So it goes back to the point I was making earlier don't create an environment of fear where people think they're gonna be replaced and let go. Yeah. Engage them. Be-- maintain that contract of employer to, to, to employee. Create an environment of trust like,"Look, even if your job is different or you're becoming an IC because there is gonna be a collapse of the management layers as just what's happening and inevitably it will happen, it doesn't mean that, just retrain yourself, learn how, how to do, learn how to manage less, learn how to be an individual contributor, learn how to focus on outcomes," just put them into a position of, of not having to fear.

echeydeAnd that, that's difficult, so you need to, you, you need to kind of like walk the talk, right? You cannot be saying something and doing something six months later because the trust is gonna be gone. So maintain and enhance and strengthen that, that contract with the employee where they are good, they are talented, but doesn't mean that, they need to go. Just keep them. They are talented. The, the definition of the job is changing. And frankly speaking, for the most part, there is gonna be a lot of, um, manager to individual contributor transition.

echeydeRight. Now, for those that cannot make that transition, it's... I just don't know what to say. That's just the world, where the world is going,

benyeah, and I do agree. I think you touched on an important point. I think fear is the big thing, isn't it? I think there's so much talk. Yeah,

echeydeyeah.

benAnd s- I think it's a lot of noise that's just put in the marketplace where, in fairness, a lot of people don't know what's going on with AI, do they? It's still an unknown journey, isn't it, for many? It's- Yeah,

echeydeyeah,

benyeah. A- a- and look- you know, it's gonna improve, but it's, it's still unknown, like, really. Like, a lot of businesses obviously they know what, how they can do it, but who knows what's gonna be the future. Like, it's still that it's a new era.

echeydeYeah. I think the future is gonna be there is a Cambrian explosion of startups. It's so much cheaper nowadays to start a company. I mean, the product development can be done much faster. You line up your customers and then you go all in on on, on go-to-market. So if there is a Cambrian explosion out there, I think there's gonna be tons of jobs in tech. There's gonna be- Yeah a lot of smaller company popping up and, and just going, you know, a- after what the big companies have built, which is what I keep telling the big companies that, look, every software company out there- In a big market is up for grabs.

echeydeYou gotta really adapt quickly or you're just gonna be replaced. And you gotta take that entrepreneurial spirit. So embrace the fear, embrace the unknown. Don't focus on fear. Don't think the unknown is bad. Take it as an, take it as an opportunity. The what, what's unknown is good for discovery, is good to create, is good for opportunity. So you need to take a mindset, which is,"Wow, look at all of the things I can do," versus like, my job is not the same. What's gonna happen to me?" Fall into a mindset.

echeydeLook forward, don't look back. Just look forward about what you can do, right? Even if your job changes completely, just take that mindset, which is like,"Wow, I can create something like I've never been able to do it. What is it that I'm good at that I can just go into that new world?" Day one mindset.

benYep. Anything is possible. Uh, so I mean, you also, you m- you touched on another point about obviously like, uh, like anyone could do pilots nowadays. Um, so I guess what's preventing the, these pilots from being scaled to like business critical solutions?

echeydeYou mean what's preventing do startups

bento scale? In terms of what's preventing like o- like I guess enterprises uh, to scale, like,'cause obviously e- like you said, like a- anyone can build a POC now, but I guess- Yeah the big challenge is going to enterprise level. So what's, where's the, I guess, the bottleneck there?

echeydeWell, the good news about the big companies that they have a ton of customers, right? So they can... I- it's just what's preventing is e- e- exactly what I've been saying for the last, you know, it's just like the, the, the legacy mindset. Just stop doing something new. Just connect the builder to the customer directly. Like all of those layers that companies create from engineering to customers, ton of layers. That's kind of all, right? Because you are literally inventing, you're creating a startup within a company. Let the builders talk to customers. Don't put everything in the middle, right?

echeydeBecause but every time you put something in between things at the human level, the coordination tax is high. Not only that, but the translation drops, right? So the messaging has to be from the builder direct to the customer as big companies are not used to that, right? Because you, they built layers and layers of customer success, marketing, product management, you name it, to not put an engineer in front of the client, but the builder directly in front of the client. You gotta, you build, when you build, you're building for a customer.

echeydeYou gotta be able to talk to them on a daily basis. But now, if you are an engineer in big tech to talk to a client, well, you need to have 20 different approval. Product managers, customer success, marketing, da, da, da. By the time you go through all those layers, you know, they're demotivated, the messaging is gone, the translation is not there. It to-- it takes 20 million different years, nothing happens. You gotta go direct to customers from the builder.

benAnd obviously, look, how do you ins-ingrain this, like, builder mindset, I guess, into companies then? 'Cause they won't... Do you gonna need constant innovation, or do you think that's gonna naturally progress now because of AI, AI working? Do you think it's... Or do you think companies still need to operate better in this

echeyderegard? It's definitely. Like I said, look, you, look, you gotta... It, it, it doesn't happen easily. You gotta bring entrepreneurial talent at the top that pushes down. If you wait for it to be bottom up, good luck, right? You have too many organic resistance, organic, you know, entitlement built into the system. It will die off, right? Okay. And, and these entrepreneurial builders, they will walk out the door. They become very frustrated. That's just reality. I mean, we cannot be i-idealistic about these things. You can be idealistic about product.

echeydeYou should not be idealistic about human nature. So when it comes to that, bring somebody at the top who's entrepreneurial and just cut through all of that mindset and bring that entrepreneurial style. You know, one of the reason I like Elon Musk, you know, I'm not gonna talk politics. I have differences on his point of view, is that that's what he does. If he goes to a company- Say Twitter or the likes, I don't want to bring the wrong analogy. He will see all of these things because he's a born entrepreneur. He knows all of the scar tissue, and he just will do whatever it takes to go into a builder mindset directly connected to the customer.

echeydeSo if you are a large company struggling, bring entrepreneurial talent towards the top as close as possible to the CEO, if the CEO is not a founder, and just bring that experience, that mindset into the larger corporation. The last thing you want is leaders that are more like, you know, No, I've never done, built a company, started a company," very caretaker, which is not gonna get it. This is not a bottom-up exercise, this is a top-down exercise from bringing entrepreneurial talent into big tech. That's just the reality, right?

echeydeThat's just the way I see it in many places. As much as I can be idealistic on the

benproduct- The platform

echeydehuman nature doesn't really change. You gotta bring that, that style into a larger tech to transform.

benYeah. No, I mean, it's gonna be, it's gonna be interesting how businesses do, do the change because it's, there's a lot going on, isn't it? There's a lot to deal with- Yeah. Yeah, yeah especially for these enterprise ones. Like, there's a massive change.

echeydeAnd then like I said earlier, I think the, the, the good news is that I think these bigger companies, they really need to do acquihires or big AI companies bringing in part-part the founders in as much in charge as possible and just, like, reset the whole thing. We are going through a massive transformation. Those who don't adapt quickly will be... I mean, the incumbents will be very gone quickly. Maybe around, it's just gonna be one-fifth of the size of which they are then. And that's why, this is a super, super... You know, VCs are VCs, but they really understand every software company is up for grab.

echeydeThey really understand a startup will rewrite and go after the customer base of the incumbent very quickly, right? That's why they are just funding them to the tune of massive amount of money just for the sake of say, accelerate it. So you're gonna see a change of the guard into many markets unless the incumbents, are smart enough to be more entrepreneurial, aka bring that entrepreneurial mindset from the outside in and just shove it down.

benOkay. So then just to bring the playbook together then, so what are the most important steps leaders should remember when moving from a legacy environment into the AI era?

echeydeI'd say number one, focus on builders, right? Builders, people that actually do the work. Number two, flatten up the org as much as you can to reduce the white space between the builder to the customer, right? Don't let the management layers get in the middle of that or all of the other artifacts. Builders connected to customers, that will go a very long way, right? And after that, which is what I've said all along if you don't have that in the house, bring the entrepreneurial talent as quickly as you can do. Right. I may be oversimplify, but that's those components, and the fourth one is really forget the orgs where you have big armies.

echeydeDoesn't work. Tight teams, very tight teams, and definition of tight teams mean builders connected to the customers. A team of five or ten people more often than not can do more than a hundred, two hundred, a thousand people. So if you are smart enough to have tight teams in many different places, then you're good because you can create a lot of good products. You should always have the fountain of ideas whereby in the past you were limited by capacity, now you're just limited by your own backlog of ideas that would make sense for the customer.

echeydeSo tight teams, and go down flatter organization. I think to me the, in a larger company, the number of layers from the CEO down to the builders, it should be really, you know, two to three layers, right? Um, you have seven, ten layers. It's crazy. It's crazy. Every time you go up a layer, you lose so much context and so much innovation can be stifled, just collapse down.

benYeah, no, it's, it's fascinating because previously, wasn't it? You'd have so large, huge teams. Now it does seem like- Yeah a lot of the companies now are building these relatively small f- like, businesses that are close to the- Yeah um, sort of leader f- f- or founder really of startups.

echeydeExactly. Um,

benso yeah, no.

echeydeWell, look, look, B- Ben, look at NVIDIA. Yeah, the CEO-

benYeah,

echeydeexactly at NVIDIA, he's got, he's got fifty heroes you don't re- you, you don't need to manage that much nowadays. 50, 100 direct, great. Go have a, have at it, right? You don't need to do one-on-ones all the time. A lot of the management playbooks built over the years, just think differently. Just think about the outcome, think about what people can do, enable them. Your job should be really as a leader to remove the, um, the friction. It's easier said than done, but to remove the friction, you gotta simplify the org. The, the thing about organization is that when there is friction in an org, the org will, will auto create more teams, more bureaucracy with the intent to remove the friction, but in fact it's actually adding even more friction.

echeydeSo they create layers upon layers because there's friction and bureaucracy creeps in. You've got to shake it out. Keep it simple.

benYeah, I mean, people want more ownership nowadays. Like it's just obviously we do-- I do hiring. So m- a key thing for a lot of people moving is m- they want more ownership, like more scope. They want to make more impact. Yeah. And if you- Yeah can obviously do-- we've got the technology now to do the heavy lifting for people's jobs. Like if you're a coder- That's right look, you can you've got someone do that. You have got the autonomy now to go and do bigger things, and that's obviously what people want. They want more control on their work, don't they?

echeydeThat's right. But what people want is impact, and impact in the front organization doesn't tie back to your title. If you are moving the needle on a product, and you're able to create a$10 million product line very quickly, then you should be paid handsomely regardless of your title. And that pa- speaks specific person who's a builder, couldn't care less about the corporate, stuff. So focus on builders that are entrepreneurial, hi- high agency, they want to get things done they create good products, reward them. Simple.

benSo then if, I guess, if a leadership team wanted to begin this journey over the next ninety days then, what would you recommend is the first meaningful action to take?

echeydeTake a mission-critical workflow and rewrite it completely in AI, right? Tight team, small, everything that you know about how you organize and all that, cut it out. Find builders that can deliver, connect them to the customers, put that tight team, get out of the way, let them do their job, and learn from what they are doing, and then from what you learn of what they are doing, then you try to translate that over to the rest of the business as you go. So it's almost like create a pilot team, give them some kind of free rein, don't try to have human nature to control and all of that, and tell them how to do things.

echeydeLet them figure how to go and get it done. And you have to let go. Once they go and get it done, they go in front of the customers and they see the success, you're gonna learn along, a lot along the way how you can take that mode of operation and just translate and replicate. But it's not rocket science. You have to let it go. Somebody talented will figure how to go and get it done. And- Let go of the control. You gotta let go of the control.

benBrilliant. Cool. Well, thanks for your time today, Jay. What I can do is I'll tag your LinkedIn link in the post of the podcast so people can reach out to you. But really thanks for your time today. It's been a pleasure having you on the podcast.

echeydeThank you, Ben. Thanks for the opportunity. it was fun, and I'm looking forward to more.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. What It Takes to Become a CIO Today

    Nov 10, 202638 minAudio

    In this episode of Data Analytics Chat, we welcome Saadia Khilji, Global Chief Information Officer at Coverys, a medical professional liability insurer that has served the industry for more than 40 years. Based in...

    Credit the people on this episode

  2. Are Companies Wasting Money on Data & AI?

    Nov 3, 202648 minAudio

    In this episode of Data Analytics Chat, we welcome Lior Barak, Author & Founder of Impact Operations. After 15+ years in data organisations, Lior now helps data teams and their leadership treat data and AI as a capital...

    Credit the people on this episode

  3. Your AI Is Live. But Is It Actually Working?

    Oct 27, 202639 minAudio

    In this episode of Data Analytics Chat, we welcome Ryan Stevens, Senior Director of Applied Science at Ramp, the financial technology company helping businesses save time and money across cards, bill pay, expense...

    Credit the people on this episode

  4. Nestlé’s Chief Data & AI Officer on Scaling AI

    Oct 20, 202649 minAudio

    In this episode of Data Analytics Chat, we welcome Jessica Matheron, Chief Data & AI Officer at Nestlé. Jessica brings a rare mix of experience to the role, from a French engineering and mathematics background and a...

    Credit the people on this episode

  5. AI Is Moving Fast. Is Your Data Ready?

    Oct 13, 202646 minAudio

    In this episode of Data Analytics Chat, we welcome Erica Howard, Head of Consumer Data Ownership at TD. Erica has spent 18 years at the bank, starting in the call centre before working on TD's early data governance...

    Credit the people on this episode