What we talked about
Grant Byron spent two decades building things from scratch, first as a designer turning blank pages into products, then as a product and operations leader helping companies scale. In this episode, Grant shares how he evolved from web design into general management, why being a generalist is a superpower, and what he learned leading product at complex marketplace platforms like Delivery.com.
Show notes
Grant Byron started as a designer and ended up running general management, HR, finance, and commercial strategy, not by chasing titles, but by consistently being willing to get his hands dirty wherever the business needed someone. That generalist arc turned out to be the exact preparation needed to launch Forward, his AI consulting firm, at a moment when most businesses understand the buzzword but have no idea how to implement the reality.
What we covered
- Grant’s pivot from design to operations happened organically when agency clients responded well to him, and internal developers trusted him too, he could handle both sides of the fence. That dual credibility eventually grew into a general management role covering HR, finance, and commercial strategy, giving him a full-spectrum view of how businesses work that most specialists never develop.
- The practical value of seeing both sides: when a development team at his agency was running a week late, they didn’t see the urgency. Grant did, because he knew that missing the project delivery meant missing the billing cycle, which meant a cash flow gap. Neither side was wrong, they just each lacked the other’s context. He argues that making decisions inside an echo chamber or siloed view is avoidable when you’ve sat in both chairs.
- On AI-assisted MVP validation, Grant built an app for his wife in two to four hours in Claude instead of handing it to a developer. She tested it with colleagues, confirmed the concept didn’t hold up, and the idea was killed without spending weeks of developer time on a spec that no one had validated. His view is that product owners should absolutely use AI to prototype and test concepts, but production-ready systems still require engineers, because scale, load, security, and infrastructure don’t emerge from a quick vibe-coded prototype.
- Grant’s model for AI consulting through Forward is problem-first rather than technology-first. He describes many companies as buying AI because of FOMO, without identifying a real problem it would solve. His starting point with any client is asking whether there is a genuine problem, then whether AI is the right tool, and only then moving to implementation. Implementing AI in the wrong place, he says, won’t show ROI and makes the consultant look bad.
- AI fluency is his first deliverable for most business clients, just teaching staff what these tools are, how to interact with them, and what they should and shouldn’t be used for. He also flags that companies need AI policy before employees start using free-tier tools that feed data back into training models.
- One colleague he mentioned swears by using voice interaction with AI, claiming it makes his interactions roughly four times faster than typing. Grant sees the barrier to entry as genuinely low now, you don’t need to be a technologist to start getting value from generative AI, and the way to learn is simply to install it and start writing prompts.
About Grant
Grant Byron is the founder of Forward, an AI and digital transformation consulting firm he launched at the start of 2026. He brings more than 20 years of experience across product, operations, design, and commercial strategy, having worked across startups, scale-ups, ed tech, and marketplace platforms.
- LinkedIn: https://www.linkedin.com/in/grantbyron
- Website: https://www.africangiant.co.za
Episode 145 of the PreVetted Podcast.
Full transcript
Federico Ramallo (00:00.67) Welcome back to the pre-Vered podcast where we spotlight extraordinary people and remarkable talent reshaping our world. Today, we’re joined by Grant Byron. He’s a product and operations leader with more than 20 years of experience building teams, products, and processes that help companies scale. He has worked across startups, scale apps, digital services, ed tech,
and marketplace platforms. And he has led work spanning product engineering, operations, and commercial strategy. He has also been part of social impact initiatives and global accelerator programs, bringing a unique mix of practical execution and mission driven leadership. Grant, welcome to the show.
Grant (00:51.64) Thank you. Thanks for having me.
Federico Ramallo (00:54.69) I’m honored to have you here today. So for people meeting you for the first time, can you tell us a little bit more about what you’re doing today?
Grant (00:59.138) Good.
Grant (01:06.786) Yeah, sure. So as you touched on, my experience has been very much in product and operations. And since the new year, I decided to start a new consulting firm that focuses on kind of bringing these skills to medium sized businesses. It’s consulting firm called Forward. And yeah, the idea is to focus very much on AI implementation.
I think it’s a digital transformation. What I’m seeing a lot of is companies are hearing the buzzword and they want to get involved in AI, but they don’t know really how to do it. And I think with my skillset, I can really help organizations just adopt technology, not necessarily AI, but technology. So that’s what I’m doing right now. I’ve started a consulting firm. So yeah, quite excited about that.
Federico Ramallo (02:04.642) That sounds exciting. I mean, it’s, you have so many opportunities to grow your consulting because I think that there’s a lot of hype about AI, but there’s also a lot of misunderstanding about how to use and leverage AI. So yeah, I’m hoping we can explore that a little bit more.
Grant (02:06.574) Mmm.
Grant (02:20.515) Yep.
Grant (02:27.158) Okay, nice.
Federico Ramallo (02:29.186) So you started with design and web development, right? What drew you into that kind of work early on?
Grant (02:33.698) Yes.
Grant (02:38.284) Yeah, I think from childhood I just had a passion for design and drawing and I suppose the real thing there is creativity, like I like to create and build. so obviously when you leave school, you know, you never really know what you want to do. And I just went into
into studying that and trying to use my creativity and I think just always being able to take something from blank canvas to something usable, that’s really where it started.
Federico Ramallo (03:18.764) Right, right. it’s so beautiful to build something out of nothing. I think that’s how, as a builder, that’s the first purpose that we find, right? Building stuff. And then I think that the second purpose that I found, that second higher purpose that I found is,
Grant (03:20.331) Mm.
Grant (03:24.194) Yes.
Grant (03:33.42) Yes.
Federico Ramallo (03:42.51) What’s purpose of building something if nobody’s using it? So being able to go from building and then getting people to use it so you can have an impact on their lives.
Grant (03:47.48) Hmm.
Grant (03:55.074) Yeah. I think that’s always been, yeah, I mean, with teams I’ve worked with, you know, not every product goes out and is successful because we used to run an agency and.
Federico Ramallo (03:56.653) becomes so powerful.
Grant (04:10.604) you know, you do a lot of like MVP type products and they get launched into the market. And I think one of the real soul killers for developers at the time was always when, you you spend so much time building something and then to see it hit the market and it doesn’t do as well as you thought. It can be quite soul killing.
I know what you mean. So there’s two parts to it. There’s building it and then there’s also how do you get it so that people can use it. And I think when you start seeing people use your product, there’s something special in that.
Federico Ramallo (04:48.364) Right, right. Yeah, I I evolved that purpose through many years, right? So it’s easy to say, but hard to put words into it, right? So, yeah. And then tell us a bit about how you move from being a coder, a builder, to operations and leadership.
Grant (04:55.436) Yeah.
Grant (05:02.254) Hmm.
Grant (05:17.036) Yeah, I think the real pivot point was, as I mentioned, I’d worked in a…
a software development agency and I’ve done design and things like that for quite a while and obviously with design, right, you’re always dealing with customers so you have to be communicating. I think the gift of design is when you can really listen to requirements and then translate that into like the vision of the customer, right?
I think when I was working at the agency, they kind of saw I had a very good rapport with customers and that kind of transitioned into like service delivery manager, where I started running projects, you know, and constantly dealing with customers. And because I had a good rapport internally with the development team and I understood how the teams worked, I could play like both sides of the…
of the fence, if you want to call it. So I knew how to handle the clients and I knew how to handle internally the developers. So yeah, kind of organically flowed into that. And then from there, I went into like a general management role and I really got exposure to like HR and finance and more the commercial side of businesses. And so like it’s progressed, right? And
Over my career I’ve really got a full spectrum of the whole flow. I know what it takes to work in the ops side, on the commercial side. I think it was more organic than chasing titles. It was more about I was always willing to get my hands dirty and take the reins, I think.
Federico Ramallo (07:04.908) Right, right. It’s a natural progression. I think that a lot of engineers prefer to become an individual contributor as a senior engineer. And I think that’s a valid career path. Other prefer to become more of a leadership role. And then there are people like you and I…
I emphasize with your story because it’s a similar story as mine that I got involved on the product side and on the business side and you get a larger appreciation of how building software is a part of a bigger thing, right? And being able to build a product and also understand the business sustainability of those products, right?
Grant (07:38.923) Yes.
Grant (07:51.682) Yes.
Grant (08:01.1) Yeah, yeah, I think there’s I’m seeing it more now how I think a lot of people get stuck in a side. So they’re either only in product or they’re only in ops. And I’ve been fortunate enough to kind of experience everything. I’ve worked on the marketing side. So I actually have a very good scope of the whole.
flow like every department in a sense. And I think when you have that knowledge, you I can nearly sympathize with each department. So like I know what each one wants and I don’t over promise.
when I know the one can’t deliver, right? So, you know, like a sales team always over promise because they don’t really know the effort that goes into building something where I know the effort that goes so I can control the promise in the sense when it comes to the commercial side. So yeah, so I think that’s, it’s quite a nice skill to have when you, you know, a lot of people avoid the generalist, but I actually being the generalist and understanding the whole thing, I think it’s helped me quite a lot too.
Federico Ramallo (09:09.475) Right, right, I agree with you. mean, I being…
having more appreciation for other roles like product ownership, like project management, that when I was as an engineer, I didn’t have much appreciation. And then when I start going to, know, jumping to the other side, as you say, then I started learning, you know, all the intricacies and the nuances of those roles. And, you know, I start.
Grant (09:33.3) Yes.
Federico Ramallo (09:43.277) value more those roles as well.
Grant (09:47.435) Yeah, for sure.
Federico Ramallo (09:51.46) So what do you enjoy most about building teams and structures where nothing existed before?
Grant (10:02.274) Yeah, you know, actually, when I’ve been thinking about things like that, like, think it all, if I look at my arc through my career, it actually always tends to, I always seem to attract the blank page type of…
scenario where it’s design, you start with the blank page and you create something, you build it and then you see it. And then I think there’s a component of validation, know, when you see someone then use the design or see the design or the team is performing well. There’s like kind of a, there’s something that like excites me about that, that when you see whatever you’ve created be successful. And what I’ve always enjoyed is
I enjoy people and I enjoy helping people. like, actually always with my teams, I always try to find what could like motivate the person, what was their career trajectory and…
If I could focus on those things and I helped them to get there or help them perform well, then I felt like I was performing well and there was a good feeling to that. So I always liked helping people and help them to build on themselves. So that was a great thing for me working in teams. Yeah, and I think it’s just I’ve always wanted to build. it’s a…
I don’t know, there’s something about creating something from scratch. And if I think about it, it’s always been from design, whether it comes to teams, I always take the challenge of going into a business and building a new team or building a new structure. So there’s definitely something that excites me about that blank canvas and trying to sort of solve the puzzle and create something effective.
Federico Ramallo (11:57.037) Right. think that building teams, I mean, when you build a webpage, you’re building something that could have logic, but it’s deterministic. Now here, when you’re building teams, you have people with agency that it’s a whole different game complexity.
Grant (12:17.771) Yes.
Federico Ramallo (12:26.169) But at end of it, you’re still building something, right? So I find that when you see that working properly, you see the impact that it has on the client. You see the impact that it has on the people, right? I had people on my team that were able to buy a new car or a new house, or they were able to get married, right? And sometimes they invite me to their weddings, right?
Grant (12:30.635) Yeah.
Grant (12:40.939) Hmm.
Grant (12:50.646) Yeah.
Federico Ramallo (12:56.663) I feel such a sense of proud and appreciation for because that person put their life on, you know, they trust me and my company and, you know, and I’m able to return that trust by, you know, by providing, you know, a good salary, stability, a good, you know, a good life, right, a good work life balance.
Grant (13:08.021) Yeah.
Grant (13:22.411) Yes.
Federico Ramallo (13:26.651) And all the frustrations that I felt throughout, the challenges of making the team work pays off when I see that.
Grant (13:39.245) Yeah, I think it’s fulfilling, suppose. I think that’s for me quite a good word for it. mean, there’s also something nice if you say building sites or mobile apps and creating something that people can use and it’s quite functional. There’s something to it. So yeah, I think.
I suppose builders will understand. It’s like in your heads you just want to build things. But yeah, that’s…
Federico Ramallo (14:11.983) Right.
Federico Ramallo (14:18.127) Yeah, I mean, I think it’s amazing.
Grant (14:18.444) But yeah, I do enjoy the people and I must admit, like something I’ve had to struggle with being more remote is, the last four years I did a lot more remote work and it can, it’s not as I can see you lose a lot of that like people skills and the side of it.
And dealing with people constantly through the screen is quite a challenge in its own side. So it does, I think the whole thing about remote, it kind of does kill that energy of what you do like that, of building the teams. didn’t have the same, like building remote teams wasn’t as fun as building like a team in what do you call it, in real life, in person,
Federico Ramallo (15:12.503) Right. Yeah, the I have a large team and you lose the feeling of how big it is until you, you you invite them all to a single place for a company meeting. And then you realize, wow, you know, all of these people depend on us, depend on the company.
Grant (15:24.714) Yeah.
Grant (15:29.301) Yes.
Federico Ramallo (15:40.035) then you start like sinking the weight of all that responsibility, but also the excitement of all the people that you’re helping, right?
Grant (15:46.571) Yeah.
Yes. Yeah, that’s great.
Federico Ramallo (15:53.946) Yes, yes. So you have worked at the intersection of product and operations, right? Why do you think that combination is so powerful?
Grant (16:06.662) I think we touched on earlier is nearly being able to see both sides of the fence. know, understanding how operations work and then, you know, I think because operations is quite internally focused and then product is quite externally focused. You’re thinking more about the customer. So I think to have both lenses on, does provide a lot more
I suppose you’re a little bit more aware of how things get affected if you make something happen on the one side, what’s the impact on the other side? I think like an example if I thought about it, which was more if we used an example of commercial and ops is, know, when I used to push the teams running the agency, you know, we would have like billing cycles obviously every month and.
When you used to push the team to deliver and then say the team’s late, they just see it as like, they’re a week late and they say, well, it’s a week late, it’s not like the end of the world. But from a commercial side of it, you can’t then maybe build, right? Because you haven’t completed the project. But because they don’t have that context, they don’t see that. So for them, the pressure is not.
on that, they’re more thinking about, well, it’s not that important. But if you were on the other side and you’re saying, well, we can’t build and then we need the cash flow, things like that, then it does become important, right? Then, so I think that’s just a great illustration of showing, you know, if you could see both sides, then you realize, okay, I’m putting the pressure on ops, but there’s this because of this commercial element. But when they only see the one side of the fence, they’re not aware of that, say the commercial element and then
Federico Ramallo (17:43.461) Right.
Grant (17:58.144) they’re making decisions based on only what they can see. And I think as you move up through management, you start to realize that sometimes, you know, staff don’t always have the full picture or, and I think it’s the same with like when you can see product and ops, you know, if you don’t have the full picture, then you’re actually, you’re making decisions more in like an echo chamber or siloed view. So I think the more views you have, you,
tend to make better decisions. That would be my take on it.
Federico Ramallo (18:31.536) Yes, it’s hard because you cannot also transmit the pressure on the business side to the engineering side because that will also make them fail. But if you don’t transmit some part of it at least, then they’re unaware. And it’s a delicate balance to how much and how little. There’s not a single answer.
Grant (18:40.704) Yeah.
Yeah.
Grant (18:54.527) Yes.
Grant (19:01.439) Yes, for sure.
Federico Ramallo (19:01.712) But yeah, yeah, yeah. I’ve seen happening on being on both sides. I’ve seen that happening multiple times. it’s, you know, at first, my first reaction thought I thought that people didn’t care. And then I very quickly realized it’s not that they don’t care. It’s that they don’t have the context. They don’t know.
Grant (19:11.349) Yeah.
Grant (19:25.888) Yeah.
Federico Ramallo (19:27.202) Sometimes they’re not supposed to know all the pressure, They’re supposed to know functionally, they just need to know what’s the next step because otherwise it would be overwhelming for them, right?
Grant (19:40.746) Yes, but I think now more than ever we’re learning how important context is, right? I think when we look at AI, I think that’s like a real way you can see now that when context is missing, how things can go so horribly wrong.
Federico Ramallo (20:03.594) Hahaha
Grant (20:05.354) That’s really one of the great things I’m starting to… What I think about AI is it’s teaching people to provide more context, which is something you’ve always wanted when you build product and whatever. You want proper context, you want proper scope. It’s nearly like AI forces that because you’re dealing with this dumb terminal in a sense. Yeah.
Federico Ramallo (20:27.662) Right, right. It’s a very optimistic kind of dumb terminal, right? Yeah.
Grant (20:34.119) Yes, optimistic, yes, always optimistic.
Federico Ramallo (20:38.316) When you say, you’re wrong, always the answer is, of course you’re right. Let me show you why I’m wrong, right? Yes, yes. The best way that I’ve been described that situation is that you should treat an agent as a new hire, a new junior hire. That you need to teach it how to do things proper.
Grant (20:41.226) Yeah.
Grant (20:48.635) Yeah.
Grant (20:58.933) Yes. Yeah.
Yep. Yeah, and it needs context, right? So again, so that’s why I think being at the intersection of both sides of the fence, I have more context. in theory, it helps me to make better decisions. I would think that’s, but that’s my view.
Federico Ramallo (21:08.814) Right.
Federico Ramallo (21:27.536) What are your thoughts on, you know, people talk about now product people can build MVP survive code, you know, because now AI allows them to do that. We also have the developers that now they can use agents to enhance what they’re building, right? And that’s
Grant (21:43.359) Yeah.
Federico Ramallo (21:52.325) What I’ve seen talking is that the merge between product and engineering is kind of getting blurry because now one can do the other and the other can do the first, right? So what are your thoughts about that?
Grant (22:08.618) Yeah, so my view on like firstly AI doing coding is I still believe you need the engineer, right? I believe the AI is more the amplifier to an engineer. Because you always I think I feel like you the AI is not good enough and you need the oversight. The code that it’s producing is not great, like at the level it’s producing. So I think
When it does produce most probably good quality, it needs to be in small iterations, right? Where it’s nearly like you very scoped and then you’re writing a thing. So what’s coming through quite strongly is more like spec driven development. So very well scoped specs and then you start coding with like say Claude and you go in iterations and you… But the developers going through it, right? I still feel you need the dev.
and then he’s just amplified by let’s say 10x which is still amazing right if you if you have a developer who can build a three month project in one month i mean that’s amazing that’s like crazy so so yeah so then there’s that part so so firstly that’s what i believe on the dev side now do i believe a product manager can do the development of the dev no i don’t think so
Federico Ramallo (23:13.822) Yes.
Grant (23:35.723) Especially, I think if people, you’re the thing of, I’m building a HTML website, sure, anyone could go build you a quick website in Claude. I mean, it’s not complex. But when we talk building complicated systems like marketplaces or like full-on systems that need to be production ready and handle like load and things like that, people…
don’t realize you can’t just, I don’t think you could just use a Claude to spin up a quick version and then you’re gonna get like 10,000 visitors a day and it’s gonna handle the load and you’re gonna get spammed and hackers, you know what I mean? So there’s like a lot more that goes to an infrastructure of like a proper, a properly built system. Yeah, but I do believe the product, so to go back to your question, I do believe the product owners.
Federico Ramallo (24:20.475) Yes.
Grant (24:34.004) can do MVPs, I think that’s a great idea. Like they can spin up quick concepts. Like the other day, my wife had an idea for like a rep, an app for a rep. And I built it in Claude like pretty quickly. You know, she played with it and then, but then she realized, okay, it’s not as useful as she thought it would be, right? But that was nice, right? I didn’t have to go to a developer.
I didn’t have to, we would have spent then weeks or like a week building it. I spent let’s say two to four hours. We had a concept, she tried it, it was done. And that’s, so think about it, days worth of work and you, you went from idea to testing the concept. Does it work? No. Gave it to some other reps. No, it doesn’t work. Okay. It’s not an idea you should chat, go with. But if I had to go through the process of
Having a developer, I would have given the developer some spec I’d written with ideas. They would have built something and like three weeks later we would have had something wasted a lot of money. So I’m very pro, I would say pro the, you know, the product person can definitely build MVPs, test out the concept with the teams. And then when the, when it’s ready and you go, okay, we validated it, you know, again, some market research or, you know, and there is a real thing here.
then take it to the development team, scope it out more thoroughly, and then build a more production-ready system. But the two aren’t actually crossing over. If anything, by them building the MVPs, you’re not wasting the time of the developer. The developer can spend time on the more serious production stuff. That would be my view, yeah.
Federico Ramallo (26:26.245) Right, right. Yeah, I agree. I think that having the engineers to be more product aware, understanding why are we building this? What’s in it for the user? I used to do that before AI because I think it would add more value to the developers, right? And now with AI agent coding, then the developer can guide
Grant (26:27.658) Hmm.
Federico Ramallo (26:55.877) the code can guide the automated tests to validate the expected behavior for the user based on the whys. Then we can leverage or, as you mentioned, enhance the productivity of developers because instead of spending hours reviewing syntax, can give the agent the guidance of what to build.
And that’s why product aware, it’s what I think it’s being able to be more specific on the specs and the requirements and the expected behavior. So then the agent can build it. And then as an engineer, you can review it and decide, okay, what are things that we need to improve to scale to, know, things like that, because agents are not really good at dealing with abstraction. So we have to provide that abstract.
what happens in the real world into what the coding environment would need, and then from there build objects, build classes, and all that stuff. But if we provide those specs, then the agents are really good at doing that, at just the syntax of coding.
Grant (28:07.615) Yeah, which is that spec driven. Yeah. Yeah. That’s spec driven. that I think more recently, like IBM have launched some of that, you know, they have those great education videos and I think they’ve released one like recently in the last month, like two months or something, but also around spec driven development. So it’s a thing. think to code nicely, you need to create these well-documented specs. And then
then you push it through. You don’t just start coding with God, think. It’s like anything, right? You need to have like a proper requirement and scope. Because if you just start building from your mind, it gets messy, right? Yeah. And I think when you start coding and it builds something and then to try and re-fix it and get it to… I think you end up, you start getting in like twisted because you’d like…
Federico Ramallo (28:50.705) You
Grant (29:06.216) trying to re-correct it all the time, you maybe spend more time correcting than if you just coded the function yourself or something.
Federico Ramallo (29:14.947) Right, right, becomes a, you’re working about becoming a 10x engineer with, you know, AI enhancement. But in order to do that, you need to have clarity of what needs to be built. So you can use the tokens efficiently, you don’t go back and forth with the agent, you can just provide the specs as clearly as possible, right?
Grant (29:15.39) But yeah.
Grant (29:23.368) Yes.
Grant (29:32.255) Yeah.
Federico Ramallo (29:37.488) And that requires a level of understanding and deeper understanding of the product and the users.
Grant (29:37.641) Yeah.
Federico Ramallo (29:46.47) the engineers before didn’t need to because we work with the requirements document and that’s it, right? And then now we built, right? And in that process, we discover the specs, right? Yeah, and then on the product side, what I’ve seen happening is that, as you mentioned, now if we want to build a prototype and validate it with users, we can just vibe code it very quickly, test it, see the results.
Grant (29:52.924) Yes. Yeah.
Grant (30:11.476) Yeah.
Federico Ramallo (30:13.393) without having to go through the development cycle, which could take weeks or maybe months. And you can iterate that faster. And then you can show that to the developers and say, well, this is what we want. I had a conversation with the agent, and now this is kind of what we want. So we go.
Grant (30:27.55) Yes.
Grant (30:34.098) Yeah.
Federico Ramallo (30:36.753) we can go instead of using a product requirement document, we can just show something that it’s more interactive. I’ve done something like that with storytelling. So I talk with the product owners to instead of, or in addition to giving us the requirements, let’s build wireframes, mockups, and let’s tell a story.
Grant (30:46.92) Yeah.
Federico Ramallo (31:04.377) So instead of being a boring meeting where we would review a long document, everybody wanted to participate because it was fun, right? It was, let’s hear that story, right? And even though people that would not build it, they wanted to be there, right? So, and in that process, we provide more clarity on what actually needed to be built, right?
Grant (31:04.723) Yes.
Grant (31:14.482) Yes.
Grant (31:20.647) Yes.
Grant (31:29.159) Yeah, I mean that’s like the perfect definition of like a use case in a way, right? It’s a… I’ve always liked, you know, talking about the journey the person is going to have. so what will Johnny do? Johnny will come, Johnny will click on you, and he will, you know, because Johnny wants to do… I think actually doing it like that, the story, it does help you to kind of walk through the steps quite well.
So yeah, I like the idea of the story. And I think also sometimes, I know it’s a lot of work, but like, as a knowing your persona, sometimes can also be super helpful. You know, who is that real target audience? Yeah.
Federico Ramallo (32:15.312) Right, right. And the other thing we’ve done is put designers to work with the product owners to help them have more clarity on the requirements. Because we used to have requirements that were not as mature or precise. And instead of going back and forth with the development team, then the designers could just work.
Grant (32:26.789) Okay, yes.
Grant (32:30.675) Hmm.
Federico Ramallo (32:41.552) back and forth with the product people. It’s faster to just change a mock-up than to change an application, So the more they would go back and forth, the better. So by the time we got the requirements, we had a better story, we had more clear requirements, we have not only the happy path, but also the negative path, things like that. And it actually allowed us to build faster.
Grant (32:50.129) Yes, way faster, yeah.
Yes.
Grant (33:06.025) Hmm.
Grant (33:11.709) Yes. Okay, nice.
Federico Ramallo (33:12.358) Yeah. But there was this transition moment where I had to convince them for the buying of working with the designers because that meant three weeks pause of new requirements, right? While they were building on better requirements, right? And it took a while to do that convincing process, right? But after they saw the results, they appreciated it, right? Everybody appreciated it.
Grant (33:31.697) Yes.
Grant (33:41.233) Okay, yeah. Yeah, I think, you know, it’s, I suppose, we all a bit, I don’t know, like humans can be a bit lazy. they tend to, I think there’s so much value in doing the work upfront.
and really being able to scope and design like and really think the process through. So by the time it really does get to the build, it’s actually way easier. But if you if you slack on that prep upfront, then the build just always will take so much longer. You miss the marks. So, yeah, there is value in there’s like a lot of value in doing the work upfront and really being prepared. I suppose it’s like building a house, right, with the architectural plans.
Federico Ramallo (34:19.856) Right.
Grant (34:30.921) You’re gonna build all over the show, but if you have a very clear view of what you building and what you want to like construct then it’s always far better
Federico Ramallo (34:42.834) Right. Right. Yeah, I agree. Completely agree. I think in this case, it was a little bit of of pressure because he had a lot of things going on under backlog. It was a little bit of they didn’t know what they didn’t know. So it was about nobody asked the questions. And by the time they would bring up the new feature to a spring planning meeting, you had, you know, a bunch of people. So
Grant (35:10.227) Hmm.
Federico Ramallo (35:12.848) You can ask a few questions, but you’re always limited by the time, right? So being able to convince them to do prepping before by asking those questions. The first iteration was with me. I sit down with one of the product owners and we kind of go back and forth, right? So was much more efficient. When they saw value into that, then I convinced that product owner to work with the designer.
Grant (35:29.096) Yes.
Grant (35:32.669) Yeah.
Federico Ramallo (35:42.717) that was better fitted on the skills to produce the mock-ups, right? I’m really bad at that. But a lot of what the designers were doing was not only design, but also introspection and asking those questions, right? So it became…
Grant (35:46.043) Yes. Yes.
Grant (36:00.563) Yeah, because you have to understand how the user interacts. I mean, that was what I had to always do, was what we call the UX, right?
I mean as a designer in my time we always used to do wireframes first, right? So no design. How is the user going to interact with the interface? Like where are they clicking? Where is the eye tracking? Above the fold things. What’s the first thing they’re going to see? And the eye always traps to the right first. So that’s why the ads always used to be on the right of the arrow.
A designer, I mean, in my time, the designers weren’t just…
there to make a pretty picture. I suppose that was always the definition of a UX UI designer is they’re way more through the process. They’re how the person’s interacting with the screen. I think when it came then to product, product would be more about defining the features and that the designers would be thinking like how’s the user interacting with that screen or yes. So the two have to nearly like work together in my view.
Federico Ramallo (37:11.228) Bye.
Grant (37:17.035) They have to work together.
Federico Ramallo (37:20.262) Yes, yes, I agree. mean, it does provide a lot of clarity on their thinking process.
They have the knowledge of the users. They understand what the user needs and wants. having the UX designers that can help them put that into first wireframes, then mockups, and then they can go back and validate it with the users. Now we have some AI that can assist on some of that process, but I feel that the UX designers still need it to guide that process.
Grant (37:37.693) Yes.
Grant (38:02.087) Yeah, I agree. Otherwise you might all start looking the same, you know, because AI is pretty much going to design the same interface. And it always designs pretty much the fastest route, right? So it’s not necessarily thinking for your maybe for your use case.
Federico Ramallo (38:02.822) Yeah, I didn’t.
Federico Ramallo (38:08.262) Right.
Federico Ramallo (38:12.465) Yes.
Federico Ramallo (38:23.59) Right, that’s why I think that the UX should still be there because otherwise you get a bland design that is not appealing, right? Yeah. And then…
Grant (38:33.138) Yeah. Yeah. So again, it comes down to the, there’s always like in my view, there’s the human and then there’s the AI that’s creating like speed, right? So whether it’s a UX person, but they still utilize the AI to amplify whatever they’re doing. I think where AI really replaces things is when it’s like more automation and you know, like
things like that, then you don’t need the human necessarily in there. Maybe there’s just someone, checkpoints. But I think when it comes to creative stuff, I still think it’s nice to, it’s more about amplifying the human than replacing the human in that context. But obviously AI will replace humans in different contexts. So when it comes in like more simple automation stuff, I think, sure.
Federico Ramallo (39:19.29) I agree. Yes.
Federico Ramallo (39:31.984) Right.
So I wanted to ask you, going back to forward consulting, can you tell us a little bit more about how you help businesses move from AI strategy to execution?
Grant (39:49.097) So obviously it’s still quite new but what I am seeing is right I think there’s a lot of businesses that aren’t always tech businesses but they are obviously using SaaS products and things like that and these businesses tend to just take things you know because someone said it’s good or something like that and you start ending up with this
this whole system which not necessarily the best fit for your organization. Because I think most people forget to approach a thing by like the problem.
what is the problem you’re really trying to solve, not just by saying, oh, there’s this piece of tech, I need to use it in my organization, and then they go and they implement it. You really should be trying to solve a problem, right? So is there a real problem? And then by understanding the problem and what the requirement is, then you can say, well, what is the best tool to solve that like issue? So I think a lot of companies don’t do that, right? Like maybe the tech companies are like great at it, but there’s a
of like companies that don’t do it like normal businesses operating in normal in other sectors right they just don’t they just get it wrong or like they’re wrongfully advised so I think there’s a great opportunity there and then the second point is so I see myself as utilizing my current skills but then when you look at it through the sort of the AI lens I see it still as the same it’s they still need to understand how they can implement
technology into their business. So I would still come in, do you actually have a problem? There are organizations and they have one stuff. So the first concept is come in, I understand the problem.
Grant (41:45.608) One of the things we’ve been looking at is also to just work on providing AI fluency. Just help people to come up to Steam with what it is. Because a lot of people don’t know what is it and what should I be doing with it in my business. Just even teaching them what the products are, how do you interact with something like prompting and things like that. There’s a gap in for that type of thing.
And then the second part is implementation. How do you leverage AI in your business? obviously AI is the hot topic now. So that is something I’m obviously keen to do, but…
I’m just open to coming in and doing digital transformation in businesses. Anywhere you can come and help the guys get set up, build proper processes that can help the business to scale. Because if you don’t have the correct processes in place, it becomes much harder to scale when you start adding numbers. Things start to get wobbly.
Federico Ramallo (42:55.473) Right. Yeah, operation multiplies.
Grant (43:01.68) I would say as in the consulting, it’s about coming in, helping out with operational scale as well as how do I use AI in my business? You know, what like even to help them at a basic form is like a really help, right? Because no one knows right now. What do I do with it? It’s still so new. So.
Federico Ramallo (43:23.933) Right.
Grant (43:24.592) So coming in and showing them some automations, helping them get some automations into their business, that’s what it needs. But again, I’m not like someone who’s just going to come in. I don’t want to just come in and say, hey, you need this.
It’s more about first understanding, like what is the challenge? What is the requirement? then, okay, is AI ready? Like, do you need AI in your business? Can you leverage it or not? And then obviously, take it from there.
Federico Ramallo (43:56.69) Right, right. I think that there’s a lot of hype about AI and lot of misconceptions. And as you said, businesses doesn’t know how to leverage it, right? Or they don’t get it because it’s so different, right? But I think that, right, right.
Grant (43:58.248) Mm.
Yeah.
Grant (44:09.564) Yeah.
Grant (44:15.526) Yeah. And it’s not for every business, Sorry. Yeah, it’s not. So that’s why I’ve always, well, do you really have a problem? Is it something that your business does need? You know, so not just get FOMO and say I need it. Yeah. So in other words, because if you do implement it in a place where it’s not needed,
It’s going to be a failure, right? As a consultant, I would have failed and you won’t really get the ROI on it. And I think when you’re running a consulting business, you need to prove ROI, right? Return on investment. You need to show that and…
if you implement it in the incorrect way or the incorrect place, you’re not going to be able to show it. So for me it’s about putting it where it is needed and if it is needed then it will show our eye on it think.
Federico Ramallo (45:15.783) Right, right, yeah, I completely agree. It’s important to tell where to leverage AI, but also where it doesn’t make sense, right? Because that’s more powerful, actually, And there’s also the additional dimension of consuming AI and building AI, right? I mean…
Grant (45:26.854) Yeah.
Yes.
Federico Ramallo (45:44.328) Most of this business, they just need to consume it. They don’t need to fine tune models or do anything like that, right? It’s more about setting the right context, setting the right rules, and then the agents can provide a lot of value, right?
Grant (45:54.053) Yeah.
Grant (46:03.516) Yeah. I mean, even companies need help around policy, right? Because, staff are just going to be using stuff, agents, right? So even then, companies need help, like, what is the correct policy even to implement? Because you can’t ignore it, you know, and say, well, that’s fine to use. And then they’re using free versions and sharing all your data to the learning models. Okay. So
Yeah, so I mean, I think there’s lots of scope in the consulting space around it. Each person needs to find their space if that’s what they want to do. Yeah.
Federico Ramallo (46:47.187) And what are your thoughts about or recommendations on how to leverage AI on software development teams?
Grant (47:01.391) I haven’t, so to be honest, I mean, was only, I’d finished with a software development team mostly when all the Claude stuff was really starting to come in. I didn’t, but I mean, what we, one of the things I did see, right, is I found that sometimes it can make the dev teams a bit more, if I wanted to say the word like stupid in the sense of we had,
we would give like certain like support issues to a developer and he constantly leverage AI, but he was actually smart enough to go through it himself. And I think…
he would end up with an answer that was really not the right answer. And I feel like if he had spent the time himself on a smaller task, he would have got a much faster result than like trying to go through all the phases with AI, if that makes sense. So I did see instances of that. I think, I think, yeah, the…
There was that challenge, right? I felt that dev teams sometimes they try to rely on it too much and then they nearly underperformed because…
It’s nearly like they forgot that they were actually the resource that was meant to be doing it. And then I had someone else that I worked with, like our CTO, where he was more like, you know, he would use it more to to scan the system, find the things that he needed to find. And I think that can be quite useful, right? Especially when you have really large legacy systems and you need to find how everything’s connected and it might be a bit more like spaghetti code than
Grant (48:50.953) and I feel like the AI can be quite useful then to help you navigate those types of things. But yeah, I suppose, I mean, that’s really all I can say around that. I I’m not an expert on implementing the AI with development teams.
Federico Ramallo (49:13.054) Right, right. I think it’s very valuable that you are able to say what you don’t know and what you know, because I think that makes you a much more reliable consultant. And I know you have the capabilities to research and figure it out if you need to, right? Yeah.
Grant (49:14.565) Yeah.
Grant (49:24.539) Yeah.
Grant (49:28.742) Yes.
Grant (49:37.017) Yeah, and leverage. yeah. yeah. I mean, I mean, that’s how it is, right? You can’t know any everything. Yeah.
Federico Ramallo (49:43.349) Right, right. What I was talking with Suati Swaboda, she’s working on Shopify, and she told me that she’s spending more time reading code than writing code, right? Because she has all these agents and Shopify fast forward to how to use, they’re leveraging more on AI.
Grant (50:00.923) Yes, that’s it.
Federico Ramallo (50:12.98) in the software development process. So she told me that, and I think that I agree with that, but the over-reliance, as you mentioned, I think that’s an important factor. mean, if we rely too much on the agents, on designing the solution, then that becomes an issue, right? It’s…
Grant (50:33.942) Mm. Yeah.
Federico Ramallo (50:37.766) I tried, I’ve been trying cloud code recently. I was using ChachiPT before and I started using cloud code. It’s amazing what you can do, but you have to be very specific and the specificity becomes the issue, right? Or the lack of specificity becomes an issue, right? But you could give, I mean, a Ruby script, I I prefer Ruby myself, but I would…
Grant (50:51.824) Yeah.
Federico Ramallo (51:06.322) I could build a Ruby script that could take me a few hours, but I wouldn’t because I have other important things to do, right? It’s not critical, right? But now I can say to Cloud, build me a Ruby script. These are the requirements. I’m not necessarily liking the styling code, but it works. So it’s good enough for my use case.
Grant (51:15.579) Yes.
Grant (51:29.242) Yes. Yeah.
Federico Ramallo (51:33.213) I think what I’m trying to say in other words is craftsmanship starts by us doing something with our hands, experimenting with it, making mistakes, right? If we over rely on the agents that they build it, I mean, I could, but.
Grant (51:43.856) Yeah.
Federico Ramallo (51:52.349) I don’t really want to maintain that script that the agent coded, right? Because the styling is not the level of craftsmanship that I would appreciate, right? But it’s good enough, right?
Grant (51:56.517) Yes.
Grant (52:07.27) It’s actually interesting. you actually make me think now like if I had to compare it and you and I remember
When you run agencies, you’re always trying to leverage the slightly cheaper resource so you can make better margins. So the trick to a good agency was always to have a good mix of senior developers and junior developers. But what you make me think now is, imagine…
all of the code was generated by juniors. In essence that’s what you’re getting right? You’re getting like junior level code. It works, it does the thing, but did it follow the best practices and methodologies?
Federico Ramallo (52:50.74) Right.
Grant (53:01.208) of coding, most probably not, because it tried to get there at its fastest route, which is what it’s programmed to do. So what you end up with is a system that’s built, let’s say, by juniors. Now your senior level person comes, and as you say, you don’t want to maintain that.
that’s when it starts to become a problem, right? And can that scale? Normally it won’t scale so well. And I think, you know, when everyone uses the samples and goes, yeah, but Claude, can code so well and blah, blah. It’s always in the context of, I think, smaller developed things and not large scale systems that need to like really handle load and.
I mean, that’s just like, I suppose, like an educated guess. But yeah, it makes me think that when you say like that, now you didn’t even want to maintain what it wrote. It was kind of like a senior not wanting to work on the junior’s code, right? And that’s like a real thing. And then what you end up with is legacy code that becomes like spaghetti code and then no one actually wants to work on it. So then the only one who’s going to work on it is the AI, I don’t know.
Federico Ramallo (54:11.636) Yeah, I I would go and say, reduce this or make it better in this particular way. The other, and this is more on my personal experience, the other way that I’ve been able to leverage more cloud code is,
Grant (54:11.662) Yeah.
Grant (54:19.299) Yes.
Federico Ramallo (54:30.642) So I’m context switching, right? They have multiple meetings throughout the day. So I don’t have the four uninterrupted hours to just code, right? So my productivity goes down because of that, right? So, but I can do 30 minutes of coding with cloud, telling what to do. And then, know, cloud keeps the context and works as a junior developer. So I can, I become better at fixing the juniors code.
Grant (54:38.98) Yes.
Federico Ramallo (55:00.724) and getting the quality of the code that I want and Cloud Code can keep the context and in the same day I can spend 30 minutes here, 30 minutes there, 30 minutes there and then I’m going to be more productive on building a feature than without Cloud Code.
Grant (55:01.178) Yes.
Yeah.
Yeah.
Grant (55:14.981) Yeah.
Grant (55:18.934) Yeah. Yeah, I suppose, I suppose again, it’s right. It’s just understanding what you’re working with. So if you understand the limitations or what it can do and what it, you know, you can maximize that that capabilities. I suppose where it starts to fall apart is a lot of people don’t understand what they’re using the tool, right? They don’t understand like
all its capabilities and all its limitations and people start to believe that already like AI is at like this at really advanced level where it’s not right. So yeah I suppose it’s just again about understanding and knowing what you’re working with. But I mean we’re still very early right.
It’s kind of, it reminds me of like when the internet came about, right? And everyone started building websites. And I mean, you remember what the first websites looked like. They were terrible. And then you look at the websites today. So, I mean, we’re only at the real beginning. And I don’t think anyone really knows where we’re to end up with the AI stuff. But I’m sure it’s going to be incredible, whatever we end up doing.
Federico Ramallo (56:39.571) Yeah, when I talk about models cannot do abstractions, I usually add yet, because that happens this year, maybe in next year that’s going to be solved, right? So it’s going faster, yeah. So Grant, we’re running out of time, but I truly appreciate you being here today. We had a great conversation with you a lot. Any final remarks before we wrap it up?
Grant (56:44.228) Yeah.
Yes. Yeah.
Grant (56:53.061) Yeah.
Grant (56:57.539) Yeah.
Grant (57:13.561) Yeah, I don’t actually… I’m trying to think. I mean, what…
Federico Ramallo (57:21.909) Any advice for people trying to implement AI in the business?
Grant (57:28.773) Yeah, think if you… I think firstly, every business should just focus on fluency. Just become AI fluent, understand what it’s about, understand the basics.
understand how to interact with you know, with Claude or chat.gbt. And I think something I’m always told by another colleague is I think the key is just to play with it, right? It’s to engage with it. So you.
You’re never going to understand it unless you’re playing with it. So go install Claude. Start writing prompts. Start with prompts. Go on a paid version for a month. Use Cowork. And I think…
The first part is the generative AI, the prompting, stuff like that. And then I think, you know, if you really start to get into it, there’s, think there’s a lot of like valuing, even playing with the genetic stuff, right? A lot of people are, you move from just talking with the systems and then you’re moving to having something do something for you, you know, while you’re not even there or, so I think the…
The key point there is, I think what most people would say is be curious, go out, play with it, learn and try and understand it a bit. And I think anyone can do it, right? It’s not now at the power of the techies or anything like that. I think a salesperson or anyone could just go and play with it and start to leverage it in their everyday life. That’s it.
Federico Ramallo (59:13.823) Amazing, amazing. And I completely agree the technical requirement or technical knowledge to start playing with generative AI has been going down. I you don’t need to be a mechanical engineer to drive a car. You don’t need to be a technologist to play with gen AI. So, yeah, great insight.
Grant (59:18.029) Yeah.
Grant (59:29.763) Nah.
Grant (59:36.185) Yes, I think the barrier to entry is very low, mean, anyone can go and install it and start typing. And even if you don’t want to type, you could do the voice.
Federico Ramallo (59:51.507) Yeah, I heard people work better with voice than typing, and they’re doing amazing things with it. Grant, thank you for joining us today. go ahead.
Grant (59:57.562) Yeah.
Grant (01:00:01.069) Yeah, I have a colleague who swears by it. He just swears by the voice. says it makes it about four times faster to interact with it.
Federico Ramallo (01:00:13.372) interesting. Grant, thank you very much for joining us today.
Grant (01:00:17.562) Thank you. Cheers.