How Leading B2B Teams Use AI to Scale Brand Execution

Let's be honest, we don’t need more AI hype. We need real examples of how teams are using it to work smarter.

In this episode, Henrique Cruz of Rows shares how AI is becoming a true creative collaborator: streamlining workflows, accelerating iteration, and unlocking bandwidth for higher-value thinking.

After partnering with Focus Lab on a full rebrand, the Rows team didn’t shelf the guidelines — they activated them. Henrique shares how they now use AI to generate on-brand creative drafts faster, using the original system as a foundation. The result? More assets, more agility, and more space for designers to focus on what matters most: pushing the work further.

What this show covers:

• Real examples of AI-driven creative workflows inside a lean B2B team
• How brand guidelines + AI = speed without sacrifice
• The tactical impact of AI on creative ops and content velocity
• How to keep humans at the center while working smarter

Whether you’re leading a brand, scaling a creative team, or exploring AI for marketing execution, this conversation delivers practical insights, not just predictions.

Episode Resources

• Follow Henrique Cruz on LinkedIn
• Henrique's post that inspired this episode
• Thinking, Fast and Slow by Daniel Kahneman

Full Transcript

[Bill Kenney]

Hey everyone. This is Bill Kenney, CEO and co-founder of Focus Lab, a global B2B branding agency. I'm back with another episode of In Conversation, a series where I sit down with either industry peers or past Focus Lab partners to talk about a relevant topic. And what's today's relevant topic? We are talking about AI - tada! Okay, bear with me for a second though. We are talking about AI in very practical and tactical terms, from somebody who sits as head of Growth, and how he and his team use it from a creative output perspective.

We worked with them four years ago on their rebrand. They still strongly live within that visual identity, but we go further into talking about how they leverage AI and the [brand] style guide that we gave them to help their creative team create more assets at scale. These are the things you're generally not hearing.

We're not talking about blog post images that look super generic. We're talking about AI through the lens of a creative department. So we hear real world examples of how he's using that and how we used AI, again, from a creative ideation and draft form, not to take things to final pixels, not to replace the creative activity, but to support and speed up the creative activity.

It actually is an unlock, and that's a word we use about probably 70 times in this interview. So Henrique, super smart, super creative, and very forward on AI and I wanted to really talk about how these things are being used. That you don't generally hear or see on the streets. If you're in B2B marketing or you're inside a creative department, you're gonna enjoy this episode.

[Bill Kenney]

Henrique, excited to have you back on the show. Geez. I mean, when did we partner for the first time? What year was that? Was that before COVID?

[Henrique Cruz]

It wasn't before COVID, but it must have been 2021.

[Bill Kenney]

Okay, just after.

[Henrique Cruz]

During COVID. I think so, yeah. Because we were all remote. Yeah.

[Bill Kenney]

It is almost like, what does during COVID even mean anymore? So, it's been a long time. So it's been four or five years since we initially partnered on the Rows rebrand. Uh, y'all moved from dashdash to Rows, right?

[Henrique Cruz]

Mm-hmm. Yeah. That's it.

[Bill Kenney]

We had a lot of fun with you, your team. Shout out to Humberto as well was very open to being very creative on the project and landing on something that was very unique and ownable and uniquely Rows. Uh, and now y'all have continued to build and evolve the product. Which is gonna bring us to the conversation today about how you're leveraging AI within the product, not just as a functionality and feature set build, but both from how you use it within the creative team.

But before we go there, why don't you reintroduce yourself to the people that may have not watched our first recording years back, uh, who you are and what do you do?

[Henrique Cruz]

So my name is Henrique, and I lead our AI and marketing teams at Rows. We've built, I think, the world's smartest spreadsheet or smartest AI spreadsheet. And so growth and marketing is a lot about product experimentations. The project that we did with you a few years ago, a lot about data, data analysis, and then all of the kind of AI research and implementation that we do that's basically on, on our teams as well.

[Bill Kenney]

And you do a wonderful job. Everything that you're doing and the campaigns that y'all are coming out with are very creative and they're really good uses and stretches of the identity system that we created many years ago now.

So just kudos to you for the people watching and listening, if you don't follow this man, you should go find him on LinkedIn. It's definitely worth a follow. I feel like you all are very open source minded. It is kind of my read on you and how you share publicly the things that you're learning in the day to day, and you're giving people clear examples like, here's how I actually use AI to help me do a marketing thing right now. And I was like, oh, that's actually helpful. Thank you. So hopefully we can unpack some of that.

God, it was probably about a year ago. The post I'm thinking where you shared how you took our illustration style that we landed on together as a project, this kind of creative line work, illustrations that were meant to express this idea of magic, which is really kind of at the core of the Rows DNA, but it could be hard for a marketing team to then execute and make new illustrations all the time.

Not every product has an illustrator on staff, but I know that y'all do have very talented creatives. So then you came out and shared on LinkedIn how you use AI to take a base illustration style to iterate, to decide if you want to make those, and then you tweak them to the final solution. And I think you put something like cuts out like 98% of the time I think was the stat you put in there. Do you want to just share a little bit of like, was that a revelation? Do you still use that? Unpack that a little bit more for us.

[Henrique Cruz]

Yeah, exactly. I think taking a short step back when we did the project with you guys, I think one of the things that we landed on that we really liked was the illustration style. A lot of our brand messaging is contrasting with Google Sheets and Excel, kind of more of a corporate tone and we came up with this playful illustration style that we could illustrate Einstein if we wanted, but then a customer or someone holding a pen.

And what happened is that we have two designers in our 25% team that also do all the product design. So clearly, you know, more creative work is a bottleneck for us. And what happens is there are many instances where you wanna do more illustrations. For example, you onboard a new team member and we always wanna do an illustration of him or her.

We do a case study, we wanna do a branded illustration of the customer to send them as a thank you note for participating in the case study. And so it's always a bit of a bottleneck.

And as the models got better, we thought, okay, can we feed our illustration style or 10 or 20 examples to ChatGPT or an equivalent, and ask it to draw an illustration of a new character and cut a lot of the work of doing drafts because if you think about the product designer, if we say, now let's do robot hands holding an iPhone, it's gonna take several hours, but then the iteration is gonna take several hours more.

So if you can have it in a way that you can generate five or six or ten that are 95% there, you can have a discussion on top of that and then they can decide how to tweak it. It just expands our creativity as a team, and that's kind of what we found out and we still use it. Yeah, we actually created a Custom GPT for that where we sent it to the brand guidelines, so the large PDF doc that we worked with you guys on, and a bunch of direct instructions. And it's good to generate first drafts that then the team can continue working on to get to 100%.

[Bill Kenney]

Yeah. Fantastic. I can imagine, and I'm starting to experience this even myself, it's exponential what you can achieve with that unlock in my mind. Like you said, this draft form, this quick iteration to say we wanna do robot hands holding a book, but then you spend four hours trying to illustrate the right style of hand orientation.

And what's the book? I don't know. Let me think this through here. There’s a bunch of different books, only to realize I actually don't like that idea. I don't think it should be robot hands in a book at all.

I'm gonna go off on a whole nother idea that that time spent and lost, where if you can use AI as a tool really to help you iterate and go like, Nope, throw it out. Throw it out, throw it out, throw it out, oh there's the idea. Okay, now I can spend my time more efficiently and more effectively. That feels like a really big unlock, even for me as a creative, and I don't even have to use it that way, that much. I can only imagine what it feels like for you in your organization that has to feel like a massive unlock.

[Henrique Cruz]

And I think we're seeing this in other fields as well, right? You see product people and developers going to prototypes faster and making conversations much more productive because you can say, look, this is what I kind of have in mind.

With the design team, we can work in a similar way. A practical example, a few weeks ago, we launched as part of the AI flow, something that we call a reflection. So at the end of the analysis, it looks back at what it did and then sees, do I need to edit anything, et cetera. And to illustrate that we wanted a robot looking itself in the mirror, but in our illustration guidelines. It's a tricky thing to do, right?

The mirror, how do you reflect and then it needs to look the other way, like when you're looking at the mirror and, and instead of going to the designer with this overall fuzzy idea, I created a few variations of it on the GPT. And then when I talked to the design team, I said, okay, this is kind of what I have in mind, let's explore on top of this.

And again. I think it was very good for both sides for me. Because as I was thinking through it, I could tweak what the idea that I had in mind and then for them to have a basis to work on. And then we landed on something close to what the model had given us, but obviously polished the way we wanted it to be.

[Bill Kenney]

Yeah. Thank you for the practical example. I think it's important to have those in these shows. I think what I was hoping the viewers and listeners would feel is this idea that AI is not necessarily for people like you and I, which are both sides of the table, creative agency and customer with creative in-house.

It's not meant to replace creative as much as it is to support, speed it up, and help it. I don't feel threatened in these examples, right? And I guess I would hope that illustrators. Maybe you wouldn't feel threatened either. I know it's a very kind of squishy conversation for most, but I'm sure that the creatives on your team probably embraced that.

Right? That helped them. Like you said, it's not just a brief that is just words. It's like, oh, and I've already kind of like did the quick dummy stress test and I kind of think we should go with this direction. Okay. Let me actually now do the detail work as the pure play creative. It's not taking their job, I guess is what I'm trying to say.

[Henrique Cruz]

Yeah, I think the people who like this the most, uh, are our designers, honestly, like this experimentation because it makes their lives better. And for another practical example, I think it was last year or two years ago an integration with Notion. And as part of the launch, we wanted to make it so that anyone who commented or reposted the announcement would have a Notion style avatar as their LinkedIn image.

[Bill Kenney]

Yes.

[Henrique Cruz]

We internally had to cap this at 200 people because there were only so many hours in the day for us to do these illustrations and these, it's an example where it doesn't have to be a hundred percent fidelity.

[Bill Kenney]

Yeah.

[Henrique Cruz]

And so with a tool like this, if it existed today, back then, we would've said to anyone who retweets at any time in the day, in the next seven days, we can make it work.

[Bill Kenney]

Yeah.

[Henrique Cruz]

It was really positive, but it was less impactful than it could have been if it was that now because of this new tooling.

[Bill Kenney]

Yeah, great example. And, and again, that doesn't take an illustrator's job because nobody would've made 4,000 custom avatars for anybody. Right? It's not even a reality. So therefore the unlock is that these things can actually happen now. It's not necessarily about, oh, but that took work from this person. It's not, it's like that wouldn't have happened.

This is like a new opportunity. I was also thinking about, for myself, we were working with a client last, I believe it was last year, and I was trying to come up with really creative ways to express how their logomark could be dreamt up into these physical forms.

Traditionally as a designer, I'm left with what mockups can I find? Is there a sign on the side of a building? Anything tangible that I can find would always be like the, the most valuable kind of visual in the deck. But in this example, I really wanted to dream, I wanted to say, what does this logomark look like if I turned it into a fishbowl and there were fish swimming around in it, what if I turned it into this giant marble stone monument that lived inside the kind of like opening of their office space when you walked in?

Because I could go into ChatGPT and Midjourney and Firefly and these other models. I could actually make that and I could make it really fast and I could throw them out. Say, no, I don't like that. Ooh, this is very interesting. And I could even put them in there as unpolished drafts. The excitement that got from the client because they could then dream, wow.

They're never gonna make a fish bowl out of it and put goldfish in it, but they could finally start to see the form other than just a flat 2D graphic on a white page. That's a real tangible, tactical example of how, even me as a stone age designer, right?

I feel like I'm on the elder side of the design team here now, but I was able to use those tools within a small window of time. Figured it out, got some really cool outputs and shipped it, and I was like, that was really fun. Going back to your point of like, who loves this the most? Maybe it's the designers.

I loved those tools in that moment. I had never been so kind of creative, quote unquote, as I was able to be right then.

[Henrique Cruz]

Yeah. I think another similar example that we had was, when we do these bigger launches, we like to do these mock or real billboard campaigns, right where we play versus Google Sheets and versus Excel. And we wanted to do a campaign with a mock billboard in front of Apple, in front of Google, really in front where you could see the headquarters.

There's nothing on any website with mock designs exactly on top of, you know, the Google Plex. But with AI we could design it and we can ask, you know, do a mock billboard in front of the Google Plex. And then we took that into Figma and we built our own billboard in the space, right? But the foreground was there and you could see it was Apple and then we added the logo, et cetera.

So it just really, again, accelerated what we could do. Otherwise we would not have done it. because there's no database with this. We would not spend hours kind of doing the Google plugs or the Apple, uh, headquarters just fit the billboard in.

[Bill Kenney]

Yes, yes, yes. Yes. You know, it's the challenging pushback of like, no, this will not steal creativity. It might actually just amplify it to a degree that we've not experienced yet and we’re talking about those things now.

[Henrique Cruz]

I think, you know, realistically these things, it's always a bit of a double-edged sword, and if all you do is freelance work for illustrations, some of these people will be impacted. The ones who don't migrate to AI tools of course will be impacted. I think what we're seeing is that as the company, we are doing more design work in aggregate than we were before.

We are doing more assets, more videos, more illustrations, so I think the world will have much more design and then the people who can adapt and use these tools I think can just leap ahead. I think we talked about this a few years ago, there's still many campaigns that we want to do that we end up not doing just because we don't have, because the bottleneck is always improving the product or something else that we're working on and I think with these new creative tools, we're getting more stuff out for sure.

[Bill Kenney]

What do you wish that you could do with AI within the marketing department now that it’s not quite ready for? Where is it still falling behind?

[Henrique Cruz]

I think a big part is video. We were chatting a bit before because we do a lot of product demos of announcements, a lot of splash screens for the users to see a lot of email newsletters. And this is always like rich content that the design team spends a lot of time on, but it's static, right?

And to translate an image of a feature into a video, it's a lot of work with motion design and we spend a lot of hours on it, and I think we're not very far off from being able to bridge that. Taking a screenshot or a splash screen and animating parts of it. We were actually playing with this this morning, like combining the new Google video models with the new prompt models and creating something animated.

Again, we know that social now lives on video and therefore B2B brands have to migrate to more and more video content. And traditionally this is really hard to produce, really hard to edit, et cetera, et cetera. So I think this is one area where we'll still see a lot of innovation in the next couple of years.

[Bill Kenney]

Yeah, maybe it levels the playing field a little bit. Again, I don't want this to be seen as through the designer's eyes, this whole conversation. I really want it to be seen through the B2B company's eyes of what they can achieve and what unlocks for them, and I can only imagine that if a smaller org, I'm not even saying, you know, I'm saying like a true startup, bootstrap, team of five needs to come out and do some of these things.

Kind of these interesting not static images. They probably just can't, both from a capacity cost. Um, who do we find? Are we even there yet as a business? But if those things become a lot easier, then it kind of levels the playing field. They can punch above their weight class probably from a creative standpoint.

[Henrique Cruz]

Yeah. And what you see, I think, and this is a bit what's missing, is you see a lot of these new video models announcements. It's like, I can make an elephant riding a bicycle in a football stadium. But brands don't want that, right? Brands want something in my identity, doing something, illustrating something and, and I think we will eventually move into the models being able to do this.

Because whenever we see these cinematic announcements, it never really fits what we wanna do. It is really cool, but not for a B2B company most of the time.

[Bill Kenney]

Yeah, it's like it's creative in the artistic expression, right? But it has to be on brand. It has to match the identity, you know? So we were working with the company this year and we're working through a mascot. So again, to do quick iteration, it's not to have AI make the mascot, but it is to say, here's the frame of the mascot that we're heading towards.

I'd like some different orientations of if it turns its head, what might that look like? If it frowns, what might that look like? If it smiles, if it jumps, what might that look like? Only to iterate, right?

I did this within a five, ten minute window just to be like, oh, right legs would do this. No, I actually think AI got that terribly wrong. It probably would do this. Okay, I don't need that one. Oh, I like that frowny expression. That's something I can pull off of. But where it was really failing was the more iterations you do, it gets off brand.

The line weights get really thick and weird. All of a sudden, some of the detail, whether it was meant to be more flat or less flat, is starting to kind of disappear and it kind of hallucinates a little bit on the brand front and you have to bring it back.

And then it's funny, like it apologizes, right? Oh, you're right, this stroke weight did get thicker. Okay, I'll fix that. I can control that better then you get even better outcomes.

[Henrique Cruz]

Exactly. And I think you'll see a lot of these models, for example, they give you png files and then it's really hard to edit them, right? And it's not exactly the type of format that you want, but. I don't think we're far away from having a Figma plugin that does some of these things. Right. Like a variation of an image into a video or into a GIF added. I think those things will come up.

[Bill Kenney]

Yeah, and probably come fast, I would say. It seems to me like y'all are on the front lines of AI and its adoption because you're actually doing it and building it yourself too within your product. You're not kind of waiting for it to arrive at the doorstep and go like, oh, this is interesting. You're actually one of the innovators.

What do you think is next? Like, where does AI go from here?

[Henrique Cruz]

I think that the current models that exist, there's still so much that we can do with them that we don't do yet, because if we don't know how to put them together well, or that they're not fast enough.

Let me give you a practical example. Like we're, we're all very focused on the next model's gonna be much, much smarter, but it might be that a big unlock is speed, like the same models that are, that we have today being much faster rather than much smarter.

Why? If you think about a lot of B2B examples, like Rows or CRM, um, in Rows, if you want to format a cell, for example, as a percentage. You could ask the AI, but it's really only productive if the AI can do it in half a second or one second or two seconds, because otherwise you use the user interface for that.

Or if you're a CRM and you want to fill a form, the AI can only help if it's really fast, much faster than you in filling a form. And I think what we're seeing is the models get very smart, but there's still, you know, the latency might be 30 seconds, 45 seconds, and until it's two or three seconds, there's a lot of potential that's not unlocked because you cannot migrate people's habits into the AI until the AI is much more productive at the small stuff as well.

So we're actually, we're gonna release this soon, uh, about how we think about the architecture of AI and do you know this book Thinking Fast and Slow? He introduced this system one, system two, thinking, right, of the intuitive tasks that are immediate and the slow tasks that we plan. And I think the evolution for AI products is a bit like that.

Like system one tasks are almost immediate. Like, delete this column, move this column, it has to happen right away. But then if you ask, create a cash flow model for my company based on this notion, doc, it's fine if it takes ten, 15 minutes.

[Bill Kenney]

Okay. Understood.

[Henrique Cruz]

It's one of the things that we're bullish on, like having really fast models for very productive, simple things and then letting it play out and think, uh, when it needs to.

[Bill Kenney]

Fantastic point. I hadn't considered the, if it can't do it faster than me clicking into the interface, selecting the row and saying change to, then why would I use AI to do it?

[Henrique Cruz]

If you don't establish the habit of using it for simple things, it's gonna be much harder for you to get around using it all the time, and if you don't use it all the time, we have less data to train. We have less feedback on what works, and the loop doesn't work as fast until you end up just not fulfilling the potential.

[Bill Kenney]

Very smart, which is why y'all continue to not surprise me on how you're continuing to innovate on AI and how you're thinking about it. I love this idea. I can't help but think that this approach is the start at the simplest point and expand from there. It is a bit of a land and expand strategy on how you're saying if we can get the user into a habit.

If they use it for this thing, they'll go, oh, I can probably continue to do that, even if they're doing that subconsciously. I think that's a really smart way to think about it instead of maybe the other way, which is trying to backdoor it. We only use it for the deep stuff and the other stuff is still just simple to click around.

Uh, y'all are killing it. I love what you're doing over there. I love celebrating you all. I love that we partnered with you all. Um, kudos to you.

[Henrique Cruz]

Yeah, I have to say that maybe we're four years into the rebrand. One of the things that people still praise the most is the brand, like how clean the product is, how the brand image and the illustrations are refreshing for a B2B product, like a spreadsheet. We get this all the time.

Also, we get a lot, how much did the domain cost? Those questions, we also get them a lot. But on the visual identity and the tone of voice, et cetera, we get a lot of praise. So that's a big kudos to you guys. We wouldn't have done it without you.

[Bill Kenney]

Yeah, I love to hear that. I would throw it back at you, though not everybody's brave enough to embrace where y'all landed, right? It does take some faith and some courage to say, oh, we can be different. I mean even the color palette, I think would give a lot of people in B2B, kind of anxiety of like, what do you mean this is gonna be our color palette? Shouldn't we do something more standard? More technical? Uh, you all really embraced this idea of like, you need to be able to own an identity and therefore it has to be unique.

[Henrique Cruz]

Thank you. Good.

[Bill Kenney]

Alright, so as we come to a close here, let's now kind of play devil's advocate. What things have not changed in your business, a B2B business as a result of AI and you believe they will not and should not change? What things should just be the way they are. And AI is not really here to make those more efficient. Not everything needs to be more efficient.

[Henrique Cruz]

I think at the end of the day, the fundamentals are still the same, right? Your job is to build a product that people need, that people want, that people are willing to pay money for, is to build a brand identity that people associate with you that they feel comfortable with, that they enjoy using. And that hasn't changed, and I don't think that will change because of AI.

There's this saying that we overestimate how much things change in a year, and underestimate how much they change in ten. I actually don't think this is true. I think we overestimate how much they change in one year, and we also overestimate how much they change in ten years.

Because when you look at things like, I always think about, for example, when you look at the penetration of e-commerce in commerce in the world or in the US it's kind of 11, 12, 13, 14%. We are 25 years into the internet and it's still only 13 or 14%, which means most commerce is still exactly the same.

And so I think we're gonna see the same thing here. We overestimate how much things change in five or ten years. It will take longer. At the end of the day, customers want to find a product that solves the pain for them. They want to have a brand that they can sometimes laugh at or that they can associate with, you know, doing this job well that looks professional or looks cheeky depending on what you're trying to sell.

And I don't think that has changed. I think we're gonna be more creative, we can do more things. But the really important things, I don't think will change.

[Bill Kenney]

You're still building a brand, which is serving a customer, showing up on time and consistently every day and doing right for the customer.AI doesn't make that any different, right? That need is still the same. Well said. Good finish. Good finish. Way to pull us back down to reality. Uh, I know the world can get into an AI frenzy and you know, I think there are the right times and places for that.

But that was a very good, realistic, kind of humble, but the world is still the same. Everybody can take, take a breath, take a beat. Innovate, but you know, let’s keep a realistic head on our shoulders. All right, sir, I'm excited to see what's coming out.

I know you've got some new benchmark testing that y'all are looking to do, which now I can connect the dots, right? If you're talking about latency and speed, I'm definitely interested. I suspect that you all are gonna score well on that. Um, and just looking forward to following along on the journey. Like I said, it's been such a pleasure to watch y'all grow and, for me, really lead the pack in a lot of ways.

[Henrique Cruz]

Thank you. And back at you. It's always a pleasure to follow you and the team, and we really hope for the best that you guys keep, building and helping brands be successful.

[Bill Kenney]

We appreciate it, that's our goal, right? Whatever we can do to support people like you. Um, whether we have AI tools to do it or don't have AI tools to do it, right? You can't prompt your way to a brand system. The people that go through that journey understand that, right?

[Henrique Cruz]

Yeah.

[Bill Kenney]

All right, until we meet again.

[Henrique Cruz]

Thank you.