Subscribe
Share
Share
Embed
The secret sauce to your sales success? It’s what happens before the sale. It’s the planning, the strategy, the leadership. And it’s more than demo automation. It’s the thoughtful work that connects people, processes, and performance. If you want strong revenue, high retention, and shorter sales cycles, the pre-work—centered around the human—still makes the dream work. But you already know that.
The Unexpected Lever is your partner in growing revenue by doing what great sales leaders do best. Combining vision with execution. Brought to you by Vivun, this show highlights the people and peers behind the brands who understand what it takes to build and lead high-performing sales teams. You’re not just preparing for the sale—you’re unlocking potential.
Join us as we share stories of sales leaders who make a difference, their challenges, their wins, and the human connections that drive results, one solution at a time.
Stephany Cardet [00:00:00]:
Genti AI is, it's new, it's exciting. I think it's going to help a lot of people scale their work. And I think that, you know, human centric design in AI is about, you know, transforming the ambiguity of the experience into one that users can trust.
Gerard Green [00:00:15]:
You're listening to the unexpected. Lever your partner in growing revenue by doing what you already do best, combining your technical skills with your strategic insights. This episode was taken from a LinkedIn Live series about sales engineering with our CEO Matt Darrow. We hope you enjoy.
Matt Darrow [00:00:32]:
Hey everyone, thanks for joining us today. I'm Matt Darrow, Co-founder and CEO of Vivun. I started the company after a career running global sales engineering teams at private and publicly traded companies. I'm here today with Stephany Cardet. As we know now, the Spanish background, she's our lead AI product designer at Vivun. So, Stephany, you've given us a lot of sort of good tee up about yourself around, you know, things about you people don't know. But let's get into a little bit more of the professional side. Can you tell the audience a bit about your background?
Stephany Cardet [00:01:02]:
Yeah, of course. So I live and work in beautiful New York City. I've been working in user experience and product design for 10ish years. I don't want to date myself roles, previous roles at IBM and cnn. And I'm just super passionate about, like humans, understanding their relationship to technology, especially as technology evolves, and finding ways to solve their problems and tell their stories through the intersection of technology and design.
Matt Darrow [00:01:29]:
Well, then that's clearly why you were the perfect speaker for today's LinkedIn Live, because we're going to be talking about how do you achieve human centered design in agentic AI and specifically how we can think about designing AI teammates that feel really natural to work with. And this is all sort of very, very cutting and leading edge because this is also new for everybody. So if you're out there in the audience and if you're in sales or pre sales and you're trying to keep up with AI's increasing impact on your role, you are definitely in the right place. So, Stephany, all right, let's dive in from your pov. What's so different about designing for agentic AI?
Stephany Cardet [00:02:11]:
Yeah, so what sets agentic AI apart is that it goes beyond the graphical interface, it goes beyond the GUI and really leans into conversational design. It's an experience that provides users a sense of autonomy and empowerment. And it's really intended to support those users who are interested in shaping their own experience. Through the interactions. And so as you can imagine, the experience is very dynamic, it's very variable. Right. And this is really different from non agentic AI where we have the graphical elements like your buttons, your links, your dropdowns, your forms that really drive the experience because those elements communicate what you can do in the product, in the experience. And so that approach is really helpful for helping users quickly understand what they can do within a product.
Stephany Cardet [00:02:58]:
But the experience is more rigid, it's structured and it's predefined.
Matt Darrow [00:03:03]:
Well, and to your point, not to date myself, but after 15 years being in B2B SaaS and Enterprise and enterprise applications, my goodness, have I been waiting for somebody to finally take my buttons in my SaaS UI away from myself. And I'm glad that you're pushing on this. So how about usability? How's that different? In agentic AI, usability is different in.
Stephany Cardet [00:03:24]:
The sense that these sort of classic more graphical interfaces lean towards predictability and preventing errors on behalf of the user, where agentic AI shifts us more into transparency and error correction. And so imagine in your non agentic experiences, we're aiming to predictable, to prevent the errors by offering you again those buttons, those links, those, those dropdowns. But in agentic AI, we're trying to serve the user by predicting intent and offering suggestions. And then this produces sort of the, you know, everyone complains about too many clicks, so this reduces the clicks. But there are new challenges, right? And so how do we understand the AI's output? Right, but usability in agentic AI needs to focus on the transparency, explaining the AI's decisions, providing opportunities for error correction because it's going to get it wrong from time to time, as well as trying to help users feel in control despite of the system's autonomy.
Matt Darrow [00:04:22]:
Well, I like what you mentioned about a new series of challenges because you're right, as folks that have been building like us for SaaS products for a very long time, we always hear, yeah, cut the clicks, too many clicks, give me these things. And then you move to this new mode and it's about, well, how can I trust it? How do I know where that information is coming from? What can I do with this? And we see people, you know, using AI kind of struggling and they're trying to get comfortable with this new norm because it is really new. You've worked with so many of our even customers on this front. What are some of the traits and the character traits that you found? Sort of makes somebody really able to make this leap to effectively working with AI.
Stephany Cardet [00:05:00]:
AI has millions of users, but I think we're finding that the most successful of these users right now are these early adopters who are really willing to experiment with the technology. And so then as a designer, I'm really interested in taking classical usability principles and applying it to the context to AI by exploring affordances, efficiencies that make the interactions feel comfortable and trustworthy. So all types of users, regardless of their comfort level, can really navigate and benefit from this new way of like using technology and getting work done.
Matt Darrow [00:05:33]:
Well, yeah, they would be if everybody was an experimenter. Your life would be far too easy because unfortunately not everybody is. And that's probably the vast minority. And we even run into people that are still sort of a little skeptical about working with AI or the concept of this. So as you think about these designs, what are some of the barriers to usability and adoption that you have to solve to pull off this new approach, not just for those sort of few experimenters, but for everybody that might be leveraging AI?
Stephany Cardet [00:06:02]:
Yeah, so I want to say there are two pretty significant barriers around usability and adoption in AI, specifically agentic AI is that one, we need to help the user understand what the AI can do, what are its capabilities, what are its limitations, how the heck does this fit into my workflow, into my world? And the second one is establishing that trust and rapport with the AI. You want to ensure that the users feel confident and that the system is reliable, it's transparent, and it is actually working in their best interest.
Matt Darrow [00:06:36]:
Those are not small barriers. What can you do and how can I trust you? How do you remove them?
Stephany Cardet [00:06:41]:
Yeah, so for around understanding the AI and its capabilities, it's really about setting clear expectations through, you know, you want to do thoughtful onboarding, you want to do really clear, transparent communication in the conversation exchange between the user and the agent. You want to create affordances that guide the user towards successful interactions, like having the AI be proactive and offer next steps or actions. And then around establishing rapport and trust, it's really critical for that transparency and feedback. And so users are always going to need to understand why the AI is producing the certain output. And so you want to provide explanations, citations, opportunities for feedback, you know, thumbs up, thumbs down to really help the users feel that the AI is reliable. And again, not just working for them, but really working with them. This, this is a new relationship, right, that we're creating with technology. But bottom line, you want clarity, you want to create confidence and comfort for the user.
Stephany Cardet [00:07:39]:
And because different users are going to bring different styles, approaches and expectations to these interactions. What feels natural, intuitive for one might not feel that the same for another. So you want to also design for flexibility and optionality to really ensure that the AI can meet the users where they're at, to help make the experience feel really personal and supportive even I.
Matt Darrow [00:08:04]:
Think an interesting point around how every user sort of has a different style between you and I using AI and there's so many different ways that we might choose or have desires to interact with an AI agent. So from the design side, how are you making decisions to drive forward amidst this ambiguity? Because there's not 20 years of patterns that have been proven out to sort of follow.
Stephany Cardet [00:08:30]:
So that's a really good question. And so then you know, the end goal is to create this very human like experience that feels like collaborating with your best work friend. For Ava, for our agent. And then Ava is designed in such a way where she understands the nuances of communication style, she understands your intentions and how you prefer to tackle the work of closing a deal. And we do this through personalization. Ava is learning from your interaction, she's learning from your conversation with her. And so this not only allows her to handle the tasks like doing discovery or drafting a demo outline, but she's also mirroring your unique style and approach back to you. Right.
Stephany Cardet [00:09:10]:
And then she's also designed to have an awareness of the tools and resources that are available to her, so she can discern when to reference perhaps a document you've given her, or when to go out and search the web or perhaps do both. And so what this is going to do is it's going to create a very natural, more fluid exchange between the user and the agent. Unlike a lot of other agents out there where you have to give it very explicit instructions for every action that you want to achieve. The other thing I'll mention is that Ava and I mentioned it's a very human like because Ava has a superpower, she knows what she doesn't know and she's highly specialized on decades of really awesome sales engineering knowledge. And so then she's able to look at the state of the deal, she's able to reference her expertise and she knows what needs to be done and when, as well as what questions or what actions to move the deal forward. So again, this proactivity is really something we lean on, this goal driven approach that really ensures that Ava is feels like a trusted coworker and not just a wizard system, if that makes sense.
Matt Darrow [00:10:20]:
Taylor Bukowski she was on talking about how do you work with AI as a teammate and specifically what you're describing with Ava, and she was pretty fired up in a whole bunch of different ways. So could you share some more of your secrets? Like how do you create this experience for different types of users? They have different interaction styles, they all have different levels of sort of comfort with the technology. Keep taking us two or three clicks down.
Stephany Cardet [00:10:44]:
So you've probably heard this before, but there's this very timeless advice in UX design and it's that don't make me think right. Design experiences that minimize the cognitive load and agentic AI really challenges this notion because the experience is well driven by the user and the user intent. There is no predefined flow. So users do need to think. And so then we aim to sort of help reduce that burden, bring that burden down, but again, making her proactive. She's offering the next step, she's offering the next thing. She knows what to do next. And so then she's taking the lead.
Stephany Cardet [00:11:19]:
Ava's taking the lead in this experience and she's messaging users with the next possible steps in the deal. We do this with an affordance we call quick prompts that you'll see in the Ava app if you're a user. She's also identifying gaps in a prospect call. She's suggesting ways how to address those gaps, or perhaps she's prepared a demo outline because she knows the demo is coming up. So this means users have to think less compared to other agentic experiences where they have to specifically prompt the agent, interpret the response, and then come up with sort of the next steps themselves. And so the goal with Ava is that she's always providing really clear, actionable steps that are often already in motion, because again, she has that knowledge and she knows what she doesn't, she knows what she doesn't know. And so this particular approach is going to benefit all of our users. But it's especially, I think, helpful to those who are new to AI and are getting accustomed to this sort of open world experience where they don't have the graphical elements like the buttons and the filters to guide them.
Stephany Cardet [00:12:21]:
This, as you mentioned earlier, this makes up the majority of our audience. At the same time, I want to point out, we do have a subset of users that are comfortable with experimenting and pushing the boundaries of what's possible within the Agent 2 experience. And these users are really invaluable to helping us uncover new use cases and really thinking about expanding the potential of the experience. TLDR here is like, by balancing the proactivity with the thoughtful affordances to remove friction, we can really create an experience that works for a wide range of users, regardless of their own personal comfort level.
Matt Darrow [00:12:56]:
My own sort of way that I interpreted what you were saying is if you're relying on the user to have to prompt AI to get what they need, that this is going to be sort of dead on arrival in terms of the experience.
Stephany Cardet [00:13:08]:
Super painful. Yeah, yeah.
Matt Darrow [00:13:10]:
Well, this. And even relating it back to one of the things that you said at the beginning, Stephany, around the whole point of this is to humanize the AI, to make it feel like I'm working with my best team member. And that's very, very different than how I've always interacted with software. And even if I were to humanize this further, if I hire somebody, they're not just going to sort of stand there in silence and wait for me to like, declaratively give them something, Right. My best team member is going to come to me and say, like, hey, Matt, like, I know you have XYZ happening, so I already did the work. So here it is.
Stephany Cardet [00:13:42]:
Exactly. That's exactly right. Yes, yes, yes.
Matt Darrow [00:13:44]:
These are, I would say, not so subtle differences that are going to become more and more clear because we're about to be oversaturated with AI agents. So how do you see the design of Ava sort of making her unique?
Stephany Cardet [00:13:58]:
I think what makes Ava unique is the things you've sort of already touched on is that, you know, it's the combination of the humanization, the personalization, the proactive engagement. She's not waiting for you to reach out to her. She's taking the lead, she's taking the initiative. She's reaching out to you with the suggestions, with the work in. In flight. And so then again, the balance of the productivity, the specialization, making her deeply human, I think it's really going to make her stand out. I mean, she does stand out where, you know, agents are either too generic or too rigid, or you always constantly have to be supplying really explicit instructions.
Matt Darrow [00:14:37]:
Let's take this maybe three steps down the road in the future, especially as the boundaries blur even more. Before humans and our AI teammates, why do we need a UI at all?
Stephany Cardet [00:14:48]:
Well, whenever there's an interaction, there's going to be an interface, Matt. Even. Even its most simplistic form, you can think of us like, as humans, like our bodies, our body language, our facial expressions. That's a. That's an interface. It's a form to. It's a form to communicate. And so for Ava, or gentic AI.
Stephany Cardet [00:15:06]:
The interface is just. It's just that, it's just a familiar bridge. It's a mechanism in which to communicate with Ava. Right. It's just that mechanism in between user and agent, if that makes sense.
Matt Darrow [00:15:18]:
Maybe. Let me try to take a crack at an analogy because again, if I contrast to my, you know, years and years and decades of using like a SaaS app, like the interface sort of is the app. Right. I'm sort of operating on the database through it, but let's see if this sticks. Right. You're saying that sort of the UI in this agentic world, this is just a shell and it needs to feel familiar, but the agent and the capabilities in the intelligence that, that exists behind the UI and purely is just coming to life through the ui.
Stephany Cardet [00:15:46]:
Yeah, totally spot on. The interface is just. It's just that it's an interface, it's a shell. It is not the intelligence driving the experience. You can think of it as like messaging a Slack coworker. The person receiving your message, they're not in Slack, they're also using it on there, on the other side of it.
Matt Darrow [00:16:04]:
I sure hope they're not.
Stephany Cardet [00:16:05]:
Yeah. So let's hope that. So again, the UI is just the medium, the mechanism to facilitate the exchange, to facilitate the conversation.
Matt Darrow [00:16:15]:
So in theory though, we could keep the same UI and supplant a different agent with different expertise and abilities. No?
Stephany Cardet [00:16:26]:
Yeah. I mean, you can think of Ava as like the brain. The UI is the body. You could take the brain out of the body and replace it with a totally different brain. The body, the interface can remain consistent, but that intelligence could be totally, totally different and adapt to an entirely different set of needs.
Matt Darrow [00:16:44]:
That's awesome. I think what's so powerful about the approach here around why the UI UX is, is important. But then there's the knowledge and the work representation, how these two things sort of come together. So Stephany, if we wrap up and if the audience only had one thing to remember from the conversation, what are you going to leave them with?
Stephany Cardet [00:17:05]:
The one thing they should remember? I'd say that, you know, Genti Ki is. It's new, it's exciting. I think it's going to help a lot of people scale their work. And I think that, you know, human centered design in AI is about, you know, transforming the ambiguity of the experience into one that users can trust. And I think as I work on Ava, the girl is really simple. When you interact with her, it should feel seamless, effortless, comforting, because it's your own unique style and preferences are understood. And you know, at the end of the day, and this is something we all stay in the design world, at the end of the day, the best design is like, invisible. It's the technology so powerful that you barely notice it.
Stephany Cardet [00:17:47]:
It just works. And it feels completely natural to use.
Gerard Green [00:17:50]:
For additional resources, check out vivun.com and be sure to check out V5, our five minute soapbox series on YouTube. If there's a V5 you'd like us to talk about longer, let us know by messaging me gerard green on LinkedIn.