Build What’s Next: Digital Product Perspectives

Designing Simpler Products With Smarter AI

• Method

Share episode

The most valuable features in your product might be hiding in plain sight. We sit down with design leader Andy Vitale to unpack how AI can strip away clutter, surface what matters, and move users from intent to outcome without the scavenger hunt. From dense banking apps to consumer software, we break down a pragmatic path: use agentic assistants to handle administrative tasks, boost findability with smarter search, and free up the interface to highlight real value.

We dive into personalization that actually delivers. Instead of broad segments, AI can synthesize behavior, preferences, and context in real time to shape the experience—while also making existing configuration options easier to discover. Andy shares how teams can pair analytics, NPS, and session data with AI-driven synthesis to spot drop-offs faster and focus roadmaps on the true unmet needs. We also explore the trust equation: data privacy, benchmark accuracy, and the difference between AI as research moderator, synthesizer, or simulated participant.

Looking ahead, we imagine agentic design systems that assemble the right UI for the moment, judgment-ready data visualizations that compress complexity, and workflow views that tell you what’s blocked, what’s yours, and what’s next. AI becomes a co-author for high performers, speeding concept validation upstream while tightening execution downstream—without losing the human taste that makes products resonate. We close with hopes and fears: faster solutions and better confidence on one side; sameness and loss of craft on the other. If you care about building simpler, smarter, and more humane products with AI, this conversation will sharpen your approach.

Enjoyed the episode? Subscribe, share with a teammate who needs it, and leave a quick review to help others discover the show.


Episode Resources:

Michael Lewandowski on LinkedIn: in/michael-lewandowski-66769b11

Andy Vitale on LinkedIn: in/andyvitale

Method Website: method.com

Andy Vitale Website: andyvitale.com


Josh Lucas:

You are listening to Method's Build What's Next Digital product perspectives, presented by Global Logic. At Method, we aim to bridge the gap between technology and humanity for a more seamless digital future. Join us as we uncover insights, best practices, and cutting-edge technologies with top industry leaders that can help you and your organization craft better digital products and experiences.

Michael Lewandowski:

Welcome to this episode of Build What's Next. My name is Michael Lewandowski, and I lead financial services here at Method. I'll be your host for this episode. In this episode, we're going to continue a conversation that our guest Andy and I started on stage about a month or two ago in a tech talk series that we run here at Method, where we really delved into how to enable the results that folks are seeking from artificial intelligence through the product delivery lifecycle. Something Andy and I talked about afterwards that we wanted to get a little bit more deeper into was really just where that product delivery lifecycle should start. And that's with unmet needs. Where are users, where are their users' needs and how do you actually develop solutions to be able to meet them? But before we get into all of that, I want to go ahead and pass it over to Andy to give an introduction to himself, and then we'll dive into the topic. So Andy, over to you.

Andy Vitale:

Awesome. Thanks for having me, Michael. Always happy to be here. So I'm Andy Vitali. I lead design at a company called Taxwell. It's a tax software company made up of two separate companies, Drake Software that handles tax provider software on the B2B side, and also TaxAct, which is consumer tax software. Before my time at Taxwell, I've led design at Constant Contact. So a SaaS marketing company. Before that at Rocket, I was the EVP of design. So Rocket Mortgage, Rocket Homes, Rocket Loans, Rocket Auto, the entire Rocket ecosystem. I've also led design in healthcare at 3M, in Power Sports at Polaris, and in financial services at Sun Trust, which later became truest.

Michael Lewandowski:

Yeah. And that's what I love about Andy is while my focus selfishly is often on financial services, he brings a real rich perspective that he's cultivated across many industries, but with always focus on design, always focus in on user needs. And so I think that's just a great place for us to get in. But before we do, because today is still a little bit around AI and how AI might help focus in on meeting users' needs. Any fun plays that you've been having with AI lately, personally or professionally, is there something that you've been using for a particular use case or a tool or anything that's just surprised you as you continue to experiment with the emerging technology that's out there?

Andy Vitale:

Yeah, you know, I there's not a single thing that I don't use AI for. And I want to go back 10 years in time before I answer that question. So I've been as I've been designing for AI software, AI tools for the last decade. So started that at 3M, focused on natural language processing, auto-suggested coding to improve outcomes in healthcare, eventually looking at how we can change customer service through sentiment analysis and changing scripts for bankers. And then how do we build digital marketing campaigns with generative AI and see how we can change that? So for me, like AI has been something that I play with all the time, always curious. So right now, uh really, really love what Google's been doing with VO and Nano Banana. Um, really fun to see what's going on with Sora. Uh, but really the most fun I'm having is with Notion right now. I think the AI tools in Notion are allowing me to build databases and move over my Evernote and think about ways to have this centralized brain for me so that it's not like how do I sift through documents to understand my own perspective on things. Okay.

Michael Lewandowski:

Now, this is the second time you've mentioned Notion, which is a tool I also use. And so now I'm gonna have to spend a lot of time after this podcast learning how you're optimizing it, because I'm sure I am not. Uh and I'm also guilty now because I was thinking about how I've been using it for my fantasy football lineups, but at least I am first place in my league right now, despite having uh no time to spend towards watching the weekend games. So you know, AI can be uh amazing and used for many different use cases, right? So uh well, good. I I just wanted to have a little fun there because I think it is interesting to see how people are continuing to adapt and what's you know helping them. But let's really pivot the conversation to where we wanted to go after our last talk, which is users and and unmet needs. I think you know what often is lost, as we talked about in our last conversation, is we want to get efficiencies and we hear AI can get efficiencies. How do we do it? Rather than starting the problem statement from saying, what are we unable to do for users that we know that they need today? And how might any number of ways of tackling that problem help to help us meet end user needs in that process? So one theme that we've seen emerge uh in the pre-AI era and now as AI takes hold as an area of opportunity or unmet need is how do you present value-added uh experiences to users in a way that they can discover them? And I'll take, for example, a dense banking app. Uh, we recently wrapped an engagement uh with a leading bank here in the United States. They had uh a highly valued experience that was so buried within their application because of all the things like transactions and money movement that need to take primacy within that experience, that it just wasn't discoverable. And so as I think about what some of the capabilities are of AI to be able to resolve basic tasks that we, as you know, say a consumer banker need to do, I want to change my address. Like that shouldn't have to be something I have to find within an application. It should be something I could just inquire and then do without any friction. And so I'm wondering, you know, how do you think about the density of the experiences, how much we've layered onto people, the number of applications and where a future might take us and where we're simplifying that and making it easier for ourselves as users or our end users as customers.

Andy Vitale:

I think we've spent a lot of time just trying to bury features or find a place to put them as we've added them. And that's created experiences that, as you described, are really difficult to find what you're looking for, especially simple tasks that should not need to be searched for. So if I want to change my phone number or any any simple task like that, leveraging AI to layer on the right like search capability is one thing that really helps. I think the other thing is with these agentic browsers that are coming out now, it's almost like having an assistant that you would have built as like a chat bot that's just part of the browser that can search for different things and point to you where they need where you need to go to see them. Uh, but realistically, I think from an AI perspective, we have to leverage like how we can keep track of documents and how we can make things more findable. And like I mentioned earlier with Notion, right? How you have this central task bar or this central like input field that can point you to everything based on what you're asking or what you're looking for. And then how do we get beyond that to as it's starting to see how you're clicking around? How does it start to anticipate what you're looking for and anticipate your needs and surface that in a way, you know, as we think about ways to leverage agentic AI and agentic browsing, it almost feels like we're bringing back clippy in a way at times, but but not, right? I I think as soon as you bring it up in a lot of conversations, like, but isn't that clippy? Didn't Microsoft have that? And it's like, yes, and it's way smarter and it's way easier to search things and find things and discover things. Plus, it's also an opportunity to surface other products that you have or other offerings that may serve a need that the customer has based on their search query.

Michael Lewandowski:

Yeah, no, I think that's great. And I love the rebirth of Clippy just as a general concept. But you know, I think it's just a theme, right? In wave one of digital, people chased efficiency. And so a lot of applications mainly became a place where uh institutions hoped DIY would create more efficiencies. And inherently, that creates just endless tasks that people have to find and then execute through a series of steps. And if you talk about what you're saying there in terms of having some sort of chat or entry point that simplifies the discoverability and maybe then can even take it through and agentically execute what you're asking it to be able to accomplish, it vastly frees up the space within any application of what else you might be able to do there. And I think that's a really interesting concept with the dedensification, the elimination of many administrative tasks and then a refocus in on what are the things that matter.

Andy Vitale:

Yeah. And as you think about what an interface like that could look like. And you know, I don't want to sit here and just like design interfaces on the fly, but the goal is to make people focus on the main content without as many distractions, because they'll be able to see what they're looking for and find those as they need them rather than them just being persistent all of the time.

Michael Lewandowski:

Yeah. Is it the death of the hamburger? Is it, you know, like the sidebar, like all these things, you know, I think we can kind of go into and envision, you know, a little bit of future casting, what might this look like? Well, that's good. I mean, I think I want to get to the UI in a little bit and how that might be different, but I want to continue to talk about user needs. And I think another area that I'm really interested in around how artificial intelligence might shape our ability or enable us to meet needs that we weren't previously able to is in the personalization or configuration or customization space. Curious your thoughts there. I think that's an area that so many people have talked about for years, but have really struggled to achieve uh the end objectives that they sought from there. And I'm wondering, you know, how do you think about what are some unmet needs that you encounter with regards to personalization? How might we take some of the emerging technologies and start to rethink how we might be able to get ahead of the game and start to meet those?

Andy Vitale:

I think personalization, what's interesting most to me there is we've we've always tried to figure out what data we have that we can start to uncover similarities in different user groups and people that come to our sites and people that use our products. But now for the first time, we're able to like access that data and synthesize it on the fly. So the experience that I have when I go somewhere could be very different than the experience that you have when you go on a website because it knows maybe my shopping behavior, maybe my interests, maybe my location, right? Things that we technically know now, but it takes a little bit longer to surface and a little bit longer to tie them together. So as we can compute faster, I think it's able to surface more relevant things to my specific use case as opposed to just eight, like the 80-20 rule, the Peter principle.

Michael Lewandowski:

Yeah, no, I love it. And almost combining the prior concept with this one, I think about your notion, right? I think about configurability. I think about how many people probably don't know that given applications or tools even have the ability to be personalized to their needs. And if you had the ability to inquire and say, hey, how might I actually shape this? So this isn't here and this is there, is that an option? The answer could be no, but maybe it is actually yes. And we haven't even been able to discover that that was an option until this point in time. And so personalization might just be again about a certain degree of discoverability of the configurability that's already built into applications and then the ability then to set those things uh in the ways that are suitable for you, that then also help companies achieve their business objectives because now you're spending more time in that environment really in a way that that suits you and your needs, let alone as it knows more about you and saves some of your data and persist some of that as well, too, right?

Andy Vitale:

Yeah, I I think the thing that we touched upon too are the unmet needs and discovering, uncovering unmet needs, that's a lot of times, you know, we had research teams that did that. Like, how do they understand as you watch someone use a product, as you watch someone go about their day, like what are they really trying to solve as a problem? Like what's the underlying root cause? And instead of surfacing like a feature and trying to see if it fits, it's really like, no, this is the unmet need that we have to find and it's been unarticulated. So I think we're getting more understanding of how people use products and we're able to synthesize that on a fly, on the fly, and we're able to get feedback through NPS data, through any type of input mechanism. And then we can actually start to see like where we drop the ball, where as an experience there were holes in it, where the drop-off was, where in the past we had to uncover more and more data, and we weren't able to identify if the flow is 14 steps, is it the top of the funnel, the bottom of the funnel, where the fallout happens? And now we're able to really piece that together faster. So it's able to just give us more power.

Michael Lewandowski:

I love you taking it back to the observability and the behavior aspect of it and just the craft of understanding user behavior itself. I think I'm always surprised by how many times I get into research and you find that users don't even know that something exists. Again, this nature of we didn't know it was there and we actually love it. We just had no, no, no idea of where it was within the app experience or that it was an available feature. And so, you know, knowing that you have abilities then to do research, to look at users' behavior and know where they're clicking through and where they're not clicking through and where they're abandoning experiences, you could maybe drill in and do further research, understand, you know, what would it take to get users to be able to understand, you know, that that exists for them and that value is there, waiting for them today. And almost this notion of like, rather than develop something new, how do we actually optimize the things that we already have as value? And can we increase engagement via that and and and leverage some of artificial intelligence capabilities to do so rather than just go off this path of continuing to create more and more and more, which we've already talked about, has its own challenges, right?

Andy Vitale:

Yeah, I think we're looking at the future. It feels a lot like addition by subtraction. How are we actually cleaning up what we're offering people and making sure that our core offerings are actually right? There's a lot of times that we look at customers and we say, you know what? They use our main features. Let's add something else. Let's add another thing. And then you realize over time that there's a smaller company that does what you used to do really well a little bit better. And that starts to splinter off some of your traffic and some of your retention. And that ends up being a problem. So now we really have time to focus on like, how do we get the core things right? Because we can operate fast and learn fast, but we could also build the next feature just as fast. So it's an interesting time where we really focus on what's most important.

Michael Lewandowski:

I love the notion of the return to the core and the focus on the core, as well as the cleanup of the auxiliary. And I think there are like segments that you were almost describing as we go through the conversation here. There is administrative, and maybe AI can help us make that not part of the experience faster for discoverability, faster for resolution, but no longer cluttering of a UI, right? There's maybe the undiscovered aspects that are value-added already that we can instantly increase engagement by just helping bring those to the surface from a discoverability perspective or otherwise. And then maybe there's like just ignore it and then go back to your core and really focus in on why users ultimately want to leverage your products. I think I have so many conversations with folks where they do want to continue to do more and more and more, but they're not doing any of it well enough to capture the engagement of their end customers. And so, you know, I think that's a really interesting, you know, uh notion that you're raising there as part of where this could go and the cleanup and focus on core exercise.

Andy Vitale:

And I think part of that comes from we've looked at how we can leverage AI for efficiency. But now we really have to look at how we can leverage AI as an accelerator to what we have now and the people that are on our team and the offerings that we have. So if every I I'm not great at math. I will do this math analogy again. But if every person on your team is doing, you know, one and a quarter times the work they used to do because of AI, then you're able to get more done. It's not about how I can like eliminate things and get down to bare bones. It's how can I accelerate faster?

Michael Lewandowski:

Yeah. No, that's great. I'm gonna throw one more use case around unmet needs at you, and then maybe I'll open it up to see what others you've had cross your mind. But you know, particularly in financial services, but I think again, everywhere is the data. There's disparate data about ourselves in so many different locations. And in banking, it was around aggregation, one of the greatest failed projects in my mind of like, hey, you have five different bank accounts in your average life, you know, with different institutions, different types of financial relationships. Let's enter our passwords, pull it in, try and aggregate and accomplish financial health, you know, financial wellness as a result of that. Well, all of those links broke, it became password heavy, the value add and insights really weren't there. But I really feel like this is the moment where we could think about like, if we have access to our data and there's the ability to interpret and produce, you know, representations of that data, to interact and ask questions of it, to do some of the synthesis that you talked about earlier. Is this the moment where like we can start to take the disparate financial lives that we have and create better synthesis, better insights, better actionability behind all of that?

Andy Vitale:

It I think for sure. Um, we tried to do that years ago in healthcare too. There's a lot of longitudinal data and there's different healthcare systems. And how do you pull together your own like health records? So, this is how do you pull together your own financial life cycle and just be able to know how to measure, how to monitor your finances so that you're living your best. I don't know, I don't want to say living your best financial life, but you're as prepared as you can be. And I I think that's a thing that AI does really well. So, good example at my company, it's open enrollment time. And as I think about what is the right insurance to choose for me, I'm going into like some of my data from my HSA and seeing the doctor's visits that I had and what they cost. And then I'm outlying the two health plans and saying, based on one scenario or the other, what the price points are, like which one is better for me and why, because this is on average how much I spent. I would have never like it, it would have taken me days to figure that out, or I wouldn't have even tried to figure that out in the past. Yeah. And now I feel like I'm making more confident decisions. And I think that's what AI has done to help consumers the most.

Michael Lewandowski:

Yeah. No, I love it. And I I love the healthcare analogy. Um, I think it's very relatable in that coming out of our last conversation, we talked about as skill sets, we're leaving a specialization era. We've been for generations in an area where I'm a product person, you're a designer, that person over there is an engineer. We only focus in on those things. And similarly, our financial institutions, our doctors have done the same thing. They've stopped caring about us as a whole being, as our community bank may have done a couple hundred years ago, or as our general practitioner, as the only person that we had access to a few hundred years ago, may have actually thought about our whole health. And we went down this pathway where only people are thinking about parts of the whole. But I think what we're suggesting here is that AI has this potential unification factor to interact with data that's currently disparate with specializations and really unify it in a sense of recommendations that are ultimately better for us financially or physically or otherwise, right?

Andy Vitale:

Yeah. And I think it's also humanized some of the providers. I think now that I have more knowledge or I feel more confident and I feel more intelligent in questions that I have, asking, you know, our finance team at at work or asking the, you know, my doctor, even, right? Like I can now talk to my doctor as a person, as opposed to someone that everything they say is is right. I have to listen to it. So I think that that's been really important for AI to help people, like I said before, feel more confident, but also catch up in knowledge and not feel like they they can't answer the the question. But I think where you were going with this is no matter as we build products, I think the the playing field is more level. It's anyone can have an idea now, and anyone can bring that idea to some sort of fidelity, to some sort of life to communicate that to other people. Where in the past, it might have been someone in sales that had an idea that scribbled something down that brought it to someone else that may have dismissed it because they didn't understand it. But now it's like it it just arms you with more things to talk about and more proof points and more um, you know, just yeah.

Michael Lewandowski:

I mean, the barriers to entry, the ability to take something to a position of where you can evaluate the validity of it, you know, all of that's removed, the subject matter expertise that used to be uh almost a fence that you couldn't get over if you didn't understand even your personal health or whatever it might be. You know, all of those things are are starting to be reduced and it it changes the needs that we have as human beings and the ways that we can then interact and it creates new opportunities, right? Yeah. I think the last thing um before I open it up around other use cases that you've thought about is um something that's really been on my mind lately is there's always been a fun thing to test and research around comparative data.

Andy Vitale:

Right.

Michael Lewandowski:

I think one of like the hottest, coldest reactions you ever get is like, do you want to have a sense of how you compare, compare next to your peer group financially or health-wise or otherwise? Um and in a place where we're moving where you can tokenize data and and make it abstract from individuals, you know, personal, but then represent it in aggregate, you know, as a comparative lens. I'm curious, you just like your own opinion as a researcher, as a designer, do you see us being able to finally tackle some of those use cases and the personalization and the aggregation and the synthesis and the insights realm where people can actually explore and learn from what other people and their peer group are doing financially, physically, otherwise?

Andy Vitale:

I think the data still is a little bit of a barrier there because of data privacy and figuring out, you know, but but I think the appetite is there more. How do I benchmark? How do I compare versus people? Even if we strip out like the core information, how do I know, you know, as a designer, how efficient my team is compared to five other teams at five different places? Or if I'm, you know, a healthcare system, how do I know how I'm performing versus something similar in size and a different geographic location? And just how does that keep me, you know, making sure that I'm performing optimally versus just my own self year over year? Sure. And and that's one as we start to uncover that, I think that that will also have the potential to gamify the system, which we know will drive more engagement and drive more usage. So people are naturally curious at how they're performing or you know, what other others in a similar vein as them are are doing. So it's really an interesting opportunity. I just that hasn't been an area that I've thought much about. I've used more like competitive analysis. If if I'm buying a new computer and I want to pick out a monitor, like help me understand which ones without having to click through 15 links are the right monitor or the right thing that I'm looking for.

Michael Lewandowski:

Well, you know, I think you came back to two of the most important elements of AI that we're gonna continue to run up against. And maybe both are honestly trust, just different versions. One is trust in the data and my data being able to be accessed by other people and what I'm entering in there is now part of a public domain. And is that encrypted or not? You know, that's kind of trust factor number one. And then trust factor number two is is the comparison you're giving me real? Are you really comparing someone me to someone who else is like me? Are you comparing me to a human being at all, or is this fictitious data that's entered the system and it's now giving me false act you know recommendations or biases or things that you know maybe are influenced by other sources that I don't necessarily trust? And so I think that the the multifaceted aspect of trust is such a critical aspect of AI, of data that always is the barrier for I think people when they're trying to achieve what is an unmet need around comparability in a safe way, right?

Andy Vitale:

And I think as as you mentioned, trust, one of the things that we're starting to explore are different research tools. So, you know, tool one, most of the tools out there, they do some AI synthesis, but they're more reliant upon humans to do the research. Tool two actually does research, it will moderate research sessions, it will do the synthesis, and there's a level of do I trust this in general? Tool three doesn't talk to actual humans, it has AI representing the people. So you're actually doing research with the AI as the participant, not the moderator. So, in terms of which you're more likely to trust, I would say, you know, AI as the moderator to me is more trustworthy than AI as the participant, just because I'm not a hundred percent sure that it knows all the nuances and different things that we uncover. But I think that if you look at all three opportunities, you've got AI playing the synthesis part, which is efficiency. You've got AI doing the moderation part, which puts a lot of people at ease that would have had to moderate, but maybe a little awkward for the people that are interacting with it. And then there's the place that also feels most comfortable. It's like, let's just, you know, trust the AI data that comes in. But it's also the most skeptical because, you know, it's not a hundred you don't know how accurate it actually is until you use it.

Michael Lewandowski:

Yeah. No, I think that's um it it really fascinating. And it would be great to run parallel studies on similar topics. I think it it enters also into an interesting realm of whether or not some of those tools, I think there's always been a dilemma out there around whether people understand their unmet needs. Right. And when you're doing research, are you just listening to what people describe as their lives, or are you able to read in between those lines and understand truly where the unmet need is? Or is it our job as innovators to come up with products that just meet people's unmet needs that they don't even know themselves don't have yet? I'd be curious, one kind of just general thoughts on that, as well as in the sense of the tooling that you're creating. Whereas could some of those scenarios where you're actually interacting with synthetic AI-generated users maybe even uncover unmet needs and ways that interacting with users who don't know that they have these unmet needs yet, is is has always kind of been a challenge point.

Andy Vitale:

Yeah, I I mean that's a good point because people often don't know what they don't know. So it really requires digging into what they're not saying, where this may pick up on the things that they're not saying because it has a collection of things that people said. So that's that's interesting. As as I think through synthetic data or synthetic users, the the part that is also interesting to me is as I watch people write a lot of things leveraging AI, you have to, and and I teach online in a master's program at Kent State. So my students are using AI in their prototyping class, and I know they're using it for some papers. And that makes me think are we starting to get a narrative that's that AI is injecting into its own writing of how it wants itself to be presented or, you know, any topic, right? If I put history of design or design's ability to add value, if 50 designers did that paper, would it be different than if AI did that paper? Because AI might have a different lens that it wants to put out there in the world. I don't think we're there yet. I think that's like the skepticism that you hear from people. But it's interesting.

Michael Lewandowski:

You know, what it does um touch upon is something that we're seeing. And we won't depart fully into this part of the conversation, but all the data right now of those who are using AI in the product delivery lifecycle right now suggests that high performers are getting gains out of it and that low performers are not. Right. And so I think when you talk about your students at the university, I'm not going to make assumptions here, but if you just generally ask something, write me a paper, it's going to produce something that's pretty uniform and probably, you know, I don't know, seed level, right? If you as a high performer are figuring out your own writing style and you're stuck on, I'd really like to go deeper into this, and you're asking five questions of AI or around one or two paragraphs, and then you're fine-tuning your results of that, you're going to get something that's distinctly different than that maybe 80% of the population that's performing that singular, you know, C plus product. Right. And so you have a very uniform outcome, and then you have a very, you know, exceptional, but very out of the standard norm, so to speak, set of outcomes that are going to come as well, too. And there's a really interesting dynamic aside from the skepticism that comes out of that, just because there are going to be many uniform and then many just outlying type of results as we get into that kind of environment, right?

Andy Vitale:

Right. And that's the way I see AI being the most powerful, like as your co-author, as your thinking and thought partner, so that you can start to build an idea, you get stuck somewhere, AI helps, but then you look at that and you tweak it and you fine-tune it. And same thing that we do with prompts, right? How do you add more specificity to get greater output? Because that's the magic in what we're doing today is the level of specificity in creates a much better output than, like you said, the broad term.

Michael Lewandowski:

It's interesting. Yeah. I hadn't even really full of thought about until we got into the conversation here. Well, you know, I think we're at an interesting point. I certainly want to give you the option to continue going down the path of unmet needs and if there's any top of mind to you. Uh but if not, I'll also give you the option to take us in the direction. Direction of front end. I'd like to do a little future casting and just have a little fun around what does this front end look like of the future when we start to maybe clean up or or take some of the density out before we do, you know, yeah, yeah.

Andy Vitale:

Let's let's go front end. I think that that will start to take us in an interesting place. I think Google years ago had the most basic, like this is the interface, it's a search bar. Like, what do you do? Yeah. And I feel like Chat GPT has picked up on that and some of the other AI tools of like, this is just a text field that you enter in and then the interface shows up. But I think we started to see over time, you know, little bottom right corner, chat bots, right? Like interfaces in general haven't changed much over the years. And in fact, a lot of products have gotten very homogenous just because best practices and you know what people got used to and understanding behaviors. But I think we're getting to a point now where we can surface information on the fly. So, how do we build design systems that are agentic? How do they actually put together what composes the interface based on what you need, what you know, what your preferences are? I think that's where we're going. I don't think we're there yet. I think there's a little bit of discomfort in trying to go there. But as I think through, you know, something like our software, as you're going through, you know, tax filing with with um, you know, comet coming out a couple of weeks ago or months ago, and you know, now Atlas just coming out, it's like, I'm playing with it to see how much it interacts with what we're doing. And then we've we've got to, as people that build products, understand what agentic systems that aren't ours, let alone the ones we're trying to build, how they impact the experience that we're creating and how we can view them to be supplemental to what we're doing so that they can help us surface the right information or give people the right instructions. You know, it's anything like even uh something like a drop box when I sync on my phone, how how does it know? Like it just syncs everything. But maybe I've got 15 photos that I just, you know, took of my wife and I at a hockey game, and I don't need all 15 of them because my head is turning one way, my eyes are closed. Like just pick the best one and upload it, right? Yeah. I shouldn't have to do that. It should be able to analyze it and do it for me.

Michael Lewandowski:

And at some point in time, you almost hope that it does from a data storage perspective and all the energy challenges that are already coming. Uh I think about that at my son's soccer games when they're being recorded. It's like, okay, who's watching these hours of footage that are sitting somewhere on a cloud, right? Right. Um But no, I think you mentioned so many great things there. So I think the search in the chat, obviously, is going to be a core part of the experience for the near term, right? It's just so integral to how Google and others have set up the ways in which we do, as does the progressive disclosure from there, of how you get further and further into the information itself and have opportunities to explore it. I think a couple of others I'll throw out at you that I want to get your insight on are data visualizations. Like I think if we clean up experiences as we did before, as we think about where you need to spend more time in a task, evaluating your physical health, your financial health, your performance last year versus this year in taxes. Maybe it's even in the software delivery lifecycle of how you're performing. Like, I feel like data visualizations are going to really become important as a mechanism for creating instant feedback loops and the ability for folks to really comprehend where there are decision points in their life or performance, you know, insights that they need to have. I'd just love to unpack that a little bit with you and see what your thoughts are there.

Andy Vitale:

You know, as as we have this conversation, I realize that we're talking about challenges that we've had for many years, right? Like the companies that visualize data really well are easier to understand and easier for people to make decisions. So now that puts that power in kind of anyone's hands. So think about internally, an internal team's got to use something like Tableau. How do you, whatever you use to look at your analytics and visualize that? Now it's it's visualizing that more powerfully, but in the past, we would go to someone else and say, hey, I don't understand this like business intelligence part of it, like these BI tools. How do I build certain views and do certain things? Where like with AI, you just can you can with a prompt, you can ask it a question and it shows you a different rendering of a different visualization that's easier to understand. I think also one of the things that we do is we've used Notebook LM to look through all of our NPS data. And instead of sifting through like lines of a spreadsheet, we're creating like a six or eight minute podcast that we can listen to that doesn't, you know, take up time from your day and visual, like cognitive load of looking at spreadsheets. And it just presents information in a way that we've never been able to access that quickly before.

unknown:

Yeah.

Michael Lewandowski:

I love where you went with uh this is a historical challenge. And it's something that we've always lived with and dealt with. And you think about the times where um you'd have to go back to a data team to have something reformatted in a in a way that you can then present or interpret in the way that you want. And now you could probably take a picture of it and ask Gemini to instantly reformat and factor that for you in whatever way or graphical format that you wanted to, or translate it even into a summary. Right. If you don't know how to interpret the graph, you could ask a question and probably do so. I think another one, just to build on that with you, is and this goes back to our comment conversation around on aggregation, is statusing and workflow. Like something that we've always seen internal teams struggle with is what is actionable in my world? What am I waiting on my left hand to do versus what can I pass along to my right? Where are we in the overall process as an end user? Am I waiting on something or is there actionability on me? And so whether it's like the pizza tracker or the Uber, you know, like there's all this opportunity to take the data and aggregate it and present more of that workflow that can then make sure people understand what's prioritized, what's available, what's next steps, things like that as well.

Andy Vitale:

I I think as a consumer of information in terms of my workflow, my my day, my day-to-day, right? Many things come in, many things go out. But how do I know exactly like what's up next for me, what my next best action is, what came back to me. And now that it's my turn to look at it again, what has been done in a summarized way so I don't have to dig through like 40 lines of an email thread to see where we're caught up, right? Or just, you know, if I'm working in a document with someone, yeah, I can click a button and see all the changes turned on, but just tell me, like on these four slides that were yours, like these changes were done or these comments were left so that I I can zero in on the things that I need to, as opposed to get distracted by everything else that's not really of concern to me.

Michael Lewandowski:

And it's great. You know, I've worked with companies that right now have active agents that can scan emails and attachments and answer basic questions via integrations into Outlook or Teams and things like that. Right. And and do a lot of what you're describing there of like not having to go through the laborious effort of of search and synthesis and summary and then response, but you know, have have a means by which they can already do that for you. And it's actually an interesting foray to one that I thought I would I would toss at you as well, too, is with all these insights, I think there's a fun world of alerts that we'll get into. And so I'm curious, is kind of in your background and experience, it's like everyone has their own tolerance for insights, alerts, you know, notifications, things along those lines. I I think as we get into an AI area and we have agents, you know, perhaps performing tasks on our behalf, like how do we help users potentially become the aircraft controllers of their own lives or their own professions and and kind of manage, you know, what that workflow looks like, where it could become probably more chaotic before it gets organized.

Andy Vitale:

Right. I I think to me, what people do is you fall into three buckets, at least I do. There are things that I absolutely don't need to do, don't need to see. It can go on. There are things that are like, all right, kind of get started on this, flag me so that I can like put eyes on it and then let it go. And then there are things that I'm like, I absolutely have to pay attention to that. Let me handle that completely. So I think giving that level of like these are the the guardrails, these are the um, I don't know the word that I want to use, but essentially I'm setting up like my own workflow. And then from there, like there are certain things that I need to be alerted about on my even my phone. Like I'm the person that has badges for like four apps and nothing else. And everything else, when I want to go there, I will uncover what I need to do there. Yeah. So it would be nice to get a summary at the end of the day of like, hey, you didn't look at these 15 things, but they're really, really important. Right. Right. So go ahead and do that. But ideally, we we want to get to a place where we're spending our time on the most important things and the things that are not important at all, that that could just fly. And even if it's just in the beginning, give me an approval cue just to look at and and hits end or whatever the task happens to be.

Michael Lewandowski:

Well, I think that's an important thing. And I think it's a good transition point to the final area that I wanted to walk through with you today, which is more around how to identify needs and design solutions that will help meet them, right? And I think it'll be a mixture of of kind of common best practices that have been around for a while, as well as maybe some new things. But I think what I'm hearing in what you're saying and also in how I've been thinking about it is we need to design solutions to be ready for configurability, for personalization, for the flagging of not important and absolutely, you know, immediate action required, right? So that as we navigate experiences, we have the ability to kind of create that configured experience for ourselves and then allow it to cultivate to kind of optimize ultimately our interaction with whatever the end uh experience really needs to be and what the objectives are.

Andy Vitale:

Yeah. And I I think we're really at an interesting point in time because there's a lot of work that happens upstream, and there's a lot of work that happens downstream. And I think that design historically has played kind of in the middle. It can flex to go more upstream for discovery. Product plays a lot in upstream places, and then design plays in downstream with execution with engineering. And I think as we think about AI tools and where an organization sees design playing, I think sometimes organizations see the value in moving it more upstream, more discovery, more concepting, more like let's validate these ideas that we think of quickly and then move it into the process. And then there are other organizations that are like, let's move this more downstream. Now, these designers, now that Figma has this MCP server, can they start to push out code, right? And and both places, depending on the maturity of the organization, are viable places to be. But ideally, I think the value in, especially with AI, is leveraging our ability to learn quickly, to concept quickly, to capture what ideas are important and more importantly, which ones aren't, so that we can throw out all of the bad ideas as quickly as possible to invest all of our effort on the good ideas. And to me, that's where this shift starts to really add value.

Michael Lewandowski:

Yeah, I couldn't agree more. And I think it it's just a perfect transition into the final topic, and and maybe we'll just touch on it because I think you did a beautiful summary there. But what doesn't change? Like what's always an imperative in the call it product delivery lifecycle and the research and design process that you know, we just always need to maintain some essence of what starts to change as as AI gets introduced. I think I think you touched on a lot of it there, but I want to give it a little bit more air. Yeah. Um, because I think it's an important topic around what's just general best practice that we should never abandon versus where do we feel like the areas are to really lean in and take advantage of some of the changes that are coming.

Andy Vitale:

Sure. I I think the the one thing that doesn't change is we have to remain people-centric, human-centric, right? And and I think that could be argued at some point also, because as we get to agentic systems, you're not maybe building a product for a person, you're building a product for an agent on behalf of that person. But my argument against that would be at the end of the day, the end consumer is still a person. So, you know, figuring out what that person's needs are are always going to be important, right? Leveraging our process to mitigate risk for the organization, to make sure that we're not building something that's not going to resonate in the market that will never go away. I think being able to understand what people resonate with. I think creating that emotional connection with people is what differentiates a lot of brands. So making sure the experience, whether it's digital, physical, I think we're going to see a lot of like service design really coming, hitting front and center of how my digital and physical encounters with someone are consistent across the board. Those are things that I think you just they they can't go away, at least not in the foreseeable future.

Michael Lewandowski:

No, I mean, I think you touched on another passion point of mine, which is embedded finance. And I think the the future is embedded finance. It's just going to blur further and further into our everyday lives as the technology enables itself and we'll be interacting with institutions throughout the course of our day through agents and otherwise. And I think it's funny you talked about agent to agent to ultimately human at some point in time. And if you think about embedded finance, that's kind of been the model, but with developers instead. It's like, hey, we need to develop things that then developers can interpret and then present via their own technologies to then eventually get to end users. And it's just a translation layer ultimately. But if you're not thinking about the end user, you're never going to actually have utilization of the services or offerings that you're providing. But if you're not thinking about the developer, you're never even going to have it offered in the first place. So I think it's going to be the same thing with agent to agent to consumer down the line, is you need to understand how machine readability then will eventually enable human enablement at the end of the day, right?

Andy Vitale:

Exactly. And that's kind of where AI plays, is it will build, it should be able to start to build these interfaces on the fly for whoever is consuming it, whether it be an agent, a developer, a person on the other side of it. It will start to take the same core functionality that's needed, but just present it in the way that that person consumes it or that thing consumes it best.

Michael Lewandowski:

Yeah. And I think I talked to you right before we started the podcast today that we're starting an engagement here where we're talking about making a design system readable for a low-co, no-code AI-empowered tool to produce prototypes. And it's like, all that's great, but there's got to be an end user at the end of the day that's going to use whatever's put out there, right? And if you're not navigating the translational layers between the systems to ultimately achieve the end user goal, you're going to have leakage around the adoption or usage along the lines, right? Yeah. Well, perfect. We've covered a lot of ground, Andy, today. And I really want to thank you. But before we wrap, I want to kind of just get to, you know, a hopes and fears exercise with you. Sure. Like as you think about user needs, as you think about AI impacting the craft that we both, you know, actively participate in on a daily basis. What's your greatest hope? What's your greatest fear as you think about what lies ahead in the next year to three years?

Andy Vitale:

I think the thing that I'm most hopeful for is getting into these more personalized, more the ability to connect with the end user experiences. I think the more we start to capture and understand how they want to use the products, I think the ability to generate multiple concepts, I think we're going to hone in on the right solution and a lot faster and a lot less expensive. Um, with that comes the fear of does everything start to look the same again? Like there's this big debate around craft versus, you know, just putting out anything and not focusing on craft as much. And, you know, we call it sometimes taste as well, but but really it's that visceral reaction that people have when they see, when they interact with something. And I think that while we can train things to follow a certain set of guide rails or design principles, at the end of the day, there's something that comes with the experience of having built something and created something to know when something's just not right. And that's not always tangible and it's not always programmable. And yes, you can say, you know, there's I I can tell, you know, based on what you're building, where people's eyes are gonna focus and where it's gonna go next. Yes. You can tell that they'll read something and what it will look like, but you won't be able to tell how well it resonates with them until you test it. And even then, it may be just because something's usable doesn't mean it's going to convert. It doesn't mean it's gonna land with people. So you're gonna need that. And I'm afraid that we may get too far away from that over time where that doesn't hold the same value. But then I think it becomes cyclical. And I think we come back to that and people realize that was what was missing.

Michael Lewandowski:

Yeah. Well, thank you, Andy. Uh I couldn't agree more. I think at the end of the day, the hope has to be that there are unmet needs that the technology continues to unlock for all of us. Uh, and uh, of course, you know, you you gotta hope that there continues to be experimentation that allows us to kind of make make sure we're continually striving for those borders and not just arriving in the same place uh as is sometimes the tendency. So uh again, thank you for joining, Andy. Uh thank you all for listening. Stay tuned for the next episode of Build What's Next.

Josh Lucas:

Awesome. Thanks for having me. Thank you for joining us on Build What's Next, digital product perspectives. If you would like to know more about how Method can partner with you and your organization, you can find more information at method.com. Also, don't forget to follow us on social and be sure to check out our monthly tech talks. You can find those on our website, and finally, make sure to subscribe to the podcast so you don't miss out on any future episodes. We'll see you next time.