Masters of Automation

View Original

How to drive better customer experience (CX) in Enterprises with AI Agents w/Tatyana Mamut

Your browser doesn't support HTML5 audio

017 - How to drive better customer experience (CX) in Enterprises with AI Agents w/Tatyana Mamut Tatyana Mamut, Co-founder & CEO, Wayfound prev. Salesforce, AWS, Pendo and more

Summary

Tatyana Mamut, CEO of Wayfound, discusses her transition into entrepreneurship and the vision behind her AI agent platform. She explains that her goal is to create a platform that allows companies to work with AI agents as intelligence, not just pieces of software. The platform aims to empower humans to collaborate with AI agents in order to create better customer experiences and transform workplaces. Mamut emphasizes the importance of focusing on customer value rather than efficiency and encourages companies to build transformative customer experiences with generative AI. She also offers a one-month free trial of the Wayfound platform to listeners.

Takeaways

  • Tatyana Mamut started her own startup, Wayfound, to create a platform that allows companies to work with AI agents as intelligence, not just pieces of software to drive efficiency gains but accelerate customer experiences and net new product experiences.

  • The focus should be on creating transformative customer experiences with generative AI rather than just optimizing internal workflows for efficiency.

  • Companies that prioritize customer value and create better customer experiences will outperform those that solely focus on efficiency.

  • The future of UI is conversational, with chat interfaces providing a more intuitive and customer-centric experience.

  • Mamut offers a one-month free trial of the Wayfound platform to listeners! Click on the link here to get started with Wayfound -

See this form in the original post

Alp Uguray (00:01.629)

Hi everyone, welcome to Masters of Automation. Today I have the pleasure of hosting Tatyana again. Tatyana, welcome

Tatyana Mamut (00:09.954)

Thank you so much. Glad to be here.

Alp Uguray (00:12.677)

Glad to have you as well. Since the last time we spoke, there's been a big change, right? Like the industry moved, you also transitioned into a new company. Now you're starting your own startup called Wayfound. I love to kick things off like directly with what made you move, make this move and what made you start

Tatyana Mamut (00:37.696)

start off by saying that one of my life goals was to never be a CEO. And the reason that was one of my life goals is because I don't particularly like to restrict or constrict myself. And, you know, I feel that CEOs are a lot of times, you know, too far, too much in the public eye to everybody's looking at them in an organization.

And I like attention only when like I want the attention. And so I really, I very much told everybody in my life that I never wanted to be a CEO.

However, sometimes the universe has other plans for us. I was, so I had been an advisor to an amazing venture studio called 24 and Up, founded by one of the co -founders of Pendo, Eric Bodic. And there was a project happening around essentially AI chatbots and how to like build on like chat GPT. And

As the developer and I were talking, I said, well, have you considered kind of pivoting this into AI agents? And he's like, I was thinking about that too. Let's really, but I didn't really have anybody to work with on it. Let's start working on it together. So we kind of pivoted the product and the platform into

AI agent, what everybody was talking about is orchestration then, but then we realized it's really not orchestration. It's more like managing intelligence. And one of the critical pieces that we understood was that people needed a new platform to really trust the intelligence that was emerging.

Tatyana Mamut (02:30.464)

So we started building out a platform. I went off and as an advisor to the studio, wrote a deck, an investor deck, sort of like a vision deck about where this was going in order to help the company, hire a CEO and raise money with this investor deck. And then the founder of the studio said,

this is kind of really like your deck, Tatyana. I don't know if anyone can deliver it the way that you can. And, and he essentially talked me into becoming the CEO. he is very persuasive, but I also really took the time to say, okay, what else can I be doing with my life?

And I really feel that I have a unique confluence of experiences around anthropology, understanding how cultural change actually happens around organizational design and building new organizations and around building SAS platforms that

it would be very hard for me to find honestly, someone better than me to do this, because my goal was to find someone better than me to do this. But when I actually got real with myself, I had to admit that probably this is something that I should be doing. And it's really my calling, right? It's my passion to bring this type of platform that creates trust and intelligence that creates an opportunity for leaders.

to really create workplaces where homo sapiens can work with AI sapiens in really productive ways. That is my calling. And so I stepped into it a few months

Alp Uguray (04:22.035)

That's really interesting. So one follow up question I have to that is, and now you've been a CEO, what are some of the things part of that role that surprised you that are different than how you expect it to be?

Tatyana Mamut (04:38.048)

It's much better than I expected. Let me just bring up the fact that it's much better than I expected because a little background on me. I've worked for a lot of CEOs. I've also been on boards where we've had issues with the CEO and there has been a CEO transition. Let's just put it that way. And so

I never thought it was a fun job, frankly. I always was like, thank God I'm not in that role. That seems really crappy. I was like, who could not pay me enough money to do what these people are doing?

But again, I wasn't passionate. didn't have the passion and the drive to bring those products into existence and into being in the way that I have the passion and drive to do this. And I find it really fulfilling to be the person who, you know, I mean, I have an amazing team, amazing team.

And they're all doing their own things and they have amazing ideas. But it's just so fulfilling to be able to bring people together to get to a similar goal, right? As a team and as a community and then bring investors and customers along and tell that story, right? At the end of the day, I love telling the story of the future that we're building.

And that to me as the CEO role is probably the biggest one for a startup CEO. And I love it.

Alp Uguray (06:16.896)

And all of the CEOs that I spoke to mentioned that because they have the conviction and vision in the product, as well as where the direction that the company can go to, that makes them convince the others to believe in the same vision. when you were starting from the venture studio to more of an AI company,

Within that transition, what gave you the conviction and focus to believe that I will go ahead and build the next AI agent platform for everyone? And what was that trigger for you as a person that, the world needs this and I want to do it?

Tatyana Mamut (07:06.646)

Well, the trigger for me was that I don't think I didn't see, and I still don't see anyone building what we're building because we are creating a platform for companies to really work with AI agents as intelligences, not as pieces of old fashioned software.

Remember, there's no Oracle database that's sitting behind these things that they're retrieving from. Now, they could be using an Oracle database to get some information to help with a particular process. I didn't feel that, I felt that everybody else in this game did not have the anthropology or the organizational design experience that I had.

And so for me, that perspective to like actually bring about something from the perspective of like, we are humans in organizations that we've created to achieve goals. How do we actually create AI systems, right? That empower us to work together with these new, again, new sapiens, right? Like these new intelligences, right? They're not pieces of software. This is the thing that

This is why people get so frustrated with them. expect they expect it to act in a deterministically Just like a piece of software just like a database It's not it's just like an Intel a smart intern coming out of a top university But who really needs a lot of onboarding training management supervision feedback in order to get better

And once people get that, once they wrap their heads around, that's what AI agents really are. They really are interns. They're not pieces of software. That's where we have real evolution and real transformation and real value add.

Tatyana Mamut (09:06.06)

And again, I really didn't see anyone else talking about it this way. Like we own the patents to the first AI agent management platform, AI agent control layers, because again, nobody else was thinking about it the way that we were thinking about it. And again, I didn't see anyone who was trying to bring this about the same way that we are. And so that's why I felt like I have to do this. I have to do this.

And honestly, like it's really great because one investor said to me, he's like, wait a second, aren't you just like building something that's really obvious? And I said, yes, we are. Apparently nobody else is trying to do this obvious thing. I mean, it's, yeah, again, it's, just how you would manage an intern. It's exactly that. And we're just building the platform for it to make them effective.

Alp Uguray (09:54.815)

Yeah.

Tatyana Mamut (10:04.466)

And they're like, well, this is really obvious. And I'm like, yes, but nobody else is doing it. So here we are.

Alp Uguray (10:11.589)

I feel like in one of the previous episodes, got the opportunity to speak to Joy Mountford and then she was saying that finding things that are simple is actually the hardest thing to do. I think it's one of those situations that it takes experience, understanding of companies to be able to see that kind of just a simple insight that everyone has been looking for. So to tie to that point,

How do you see the customers adopting it? their perspective as they engage with this intern that over time gets smarter, work with other interns that are specialized in certain tasks and whatnot. How do you see that?

Tatyana Mamut (10:56.748)

Yeah, it's interesting. right now, all of our customers are using Wayfound to build new custom or personalized customer experiences. A lot of it is because we're working with product leaders, frankly. But what

We're not seeing a lot of labor reduction use cases on Wayfound. You can certainly build those on Wayfound. But what people are really inspired by is they see our website, which is really just an agent. And you get a much better experience asking your questions, getting the responses, as opposed to having to talk to a customer.

Tatyana Mamut (11:42.314)

or talk to a sales rep in order to get those specific questions answered. We also have custom onboarding experiences. Several of our customers are reducing the time to value of getting new customers onboarded by using these personalized, customized agents. We have a customer who's kind of revamping their whole product experience to make it, to have it be delivered through smart, customized, personalized agents so that

not the products, their main products now will not be just the same thing off the shelf for every customer. It'll be customized based on the UI and the experience will be customized by the agent and then delivered in a more personalized way. when people really see the platform, they immediately go from, how do I save time from?

you know, like searching documents, which you can do, by the way, you can do that with WayFound. But like people immediately go from those use cases to like,

Alp Uguray (12:38.559)

Hey

Tatyana Mamut (12:49.804)

I get it now, I can build something for my customer that like nobody, that we weren't able to build before because this agent can call that agent and get a personalized and customized expert answer and then deliver that to the customer. Or we also have multimodal support. like, we could like deliver exactly the right chart or exactly the right graph or exactly the right image at the right time for the customer experience.

Mostly what we're getting and again, think product leaders are getting really inspired by what they can do and the types of new products they can build because it's a no code platform.

More or less, you still need developers to hook up all the tools and the APIs and things like that, usually. But once you have those hooked up, the product managers can configure how the agents work. So I think a lot of people are really inspired to build new things. So that's how

Alp Uguray (13:48.445)

And how does the, how does the transparency work? like, like from a customer standpoint, when they have like two to three agents, maybe specialized on, let's say I have my product analysts, like getting all the logs from the customers and their interactions. And I have my product manager trying to figure out the future pipeline, what is new. And then trying them in the middle, I have just like customer feedback, people being upset

So like those agents, do they speak together? And then as an enterprise, I have the transparency and visibility into the conversation track? What happened? How did they go from reasoning to decision? think seeing that will be really powerful for all of the customers. And I'd love to see how you approach to

Tatyana Mamut (14:40.966)

Well, perfect. yes. imagine again, AI agents are interns. So imagine you have an intern, right? That's managing the roadmap and they are like, you know, helping the product team, like figure out what the roadmap is going to be. And they're holding onto it and answering questions about it. And then you have another agent, you got to have another intern.

that's manning the customer service calls and writing the transcripts for all the customer service calls, right? And they have all that information. If you were a manager, what would you do to get those two interns to basically to make sure that the customer service feedback is getting to the product roadmap? Like, what would you do?

Alp Uguray (15:23.199)

I will communicate it. will plan.

Tatyana Mamut (15:25.814)

Well, but how would you get like, you would call a meeting, right? You would, you would create a meeting. You would create a meeting and you would invite these two interns into the meeting and you would create an agenda as the manager and say, look, I want the customer service, like you, the person who knows all about the customer service feedback. Can you please look at the roadmap and tell me which items are aligned with what customers are really frustrated about and which ones aren't right.

Alp Uguray (15:29.309)

Yes, yeah, I'll create a meeting.

Tatyana Mamut (15:54.614)

That is exactly what you can do in WayFound. We have something called meetings, right? And in a meeting...

You put two or more agents, actually you can have a meeting with one agent. You can have a one -on -one, right, with one agent and ask them about more detailed questions about their own context. But you can put multiple agents, as many as you want, into a meeting. And then you create an agenda and a goal for the meeting. And then you run the meeting. And it creates a shared context. Remember, you're putting these interns into a shared context.

And then from that shared context that also, by the way, the shared context has the identities of the agents that are also bringing in the content, right? So it's not just.

Alp Uguray (16:40.287)

So then I can bring like a plateau as my intern and then can I bring also like, to bring different perspectives in as well, like setting personalities off, for example. Yeah.

Tatyana Mamut (16:54.318)

So that is set when you build the agent, right? So yes, if you're, yes, if when you, but remember all of that also context comes in, right? And it's like these two interns are really communicating with each other and answering the agenda items together as best as they can. And the output is the answer to you, the answers to your agenda items from the shared context.

Alp Uguray (16:57.523)

Yes.

Tatyana Mamut (17:19.168)

in terms of, so you, everything is transparent, right? To the human, right? Because the human is calling the meeting and running the meeting. And then there's a follow -up discussion where you're like, wait a second, how did they come up with this answer? You can say, Hey, how did you come up with this answer? And they will again, answer it. it's, it's exactly

holding a meeting between your employees exactly like it. You're creating the shared context. You have follow -up discussion. You can ask follow -up questions. And then you can create, you know, the next step, which we're building, is creating next steps, right? Like follow -up action items. So if you have action items that come out of it, then you can assign those action items, right, to the agents, or you can assign it to the manager. You can say, hey, manager now,

We need some global directives. I need you to put this agent on a PIP, right? A performance improvement plan to try to get their performance to be a little bit better because I didn't like what they're doing. You can have all these kinds of follow -up actions afterwards, after the meeting.

Alp Uguray (18:26.879)

That's really cool actually, like putting an agent into a performance improvement plan based on the next step. that then in a way does the reinforcement learning of improving their performance over time and making sure they're more active and more engaged in meetings.

Tatyana Mamut (18:45.878)

Yeah, we already basically have that. So we have two types of learning and feedback loops in Sideway Found. So the first one is self -reflection. So every agent, every night.

runs a job where it basically reflects on how well did it perform, right, given its whole history, but including the history of that day. And then it provides suggested directives for the human operator, as well as more information that's required. So basically it says, hey, these are things I didn't do well. I actually think I should have these directives added to my, you know, to my prompts.

essentially, right? Please add these directives, because I think this would help me improve my performance in terms of accomplishing the goals that I'm supposed to accomplish. And then it also says, hey, I couldn't answer these questions, or I couldn't do these things that users wanted me to do. Can you please give me either access to this type of tool or give me more information on these topics? So we have that kind of self -reflection. Then we have the manager.

sort of, you know, kind of analysis, right? We call it the manager assessment. It's a self -assessment and then a manager assessment. So the manager assessment then looks across the agents and it basically has a leaderboard, right? These are the top performing agents based on the manager's opinion, judgment, right? So the manager also every day looks at what is the performance of the agent team that's under

and says, what are my top performing agents? What are my like worst performing agents?

Tatyana Mamut (20:33.662)

you know, stack ranks them in terms of the normal metrics that you would expect, like usage, You know, conversations with other users, number of actions performed, those types of things. You have all of that transparency. But then if you go into like one of the lower performing agents, you click on one of them, again, you can have a conversation with the manager, say, what's going on, right? Like, why do you think...

this agent is not performing well. Why did you, why are you ranking it so low? You know, and then you can, and then it's like, are there any directives, right? That should be included to make it perform better. The next step is to say, then please add these directives to the agent and tell me next week whether the performance has improved and by how much that's essentially a performance improvement plan.

Alp Uguray (21:24.775)

I love the simplicity in how it's structured and it made me think about a lot of the enablement side of it, like enabling a customer, especially an enterprise to your original point of the organizational behavior. Because especially in some of those meetings as they're assessing maybe the product roadmap or maybe assessing like different parts of the organization to make decisions on top of

setting up that meeting to encompass different perspectives or different points of views because then agents may not be complementary all the time to their ideas but maybe sometimes argumentative to be able to keep bonds of different views. How do you look at that?

Tatyana Mamut (22:17.984)

We haven't seen that yet. agents seem to not really have egos or be worried about losing their jobs when they go into these meetings. So if you actually, like I've been in a lot of meetings where customer service or sales was giving feedback on roadmaps and people get really defensive.

Right. And then there's a lot of blaming afterwards. The product team didn't listen to us. The customer service team doesn't know what they're talking about. You know, like that kind of stuff. They just don't do that. Right. They don't have egos. They're not worried about losing their jobs. They're not losing worried about losing, not getting their next promotion. so you have, you get pretty good results. Actually, you don't have that type of,

emotional, you don't have the emotional context, you just have the intellectual context right now when you put agents into a meeting.

Alp Uguray (23:18.911)

And then I think that's interesting because my point was from the configuration standpoint, configuring two agents that try to solve a single problem, but approach it to different perspectives to collaborate on it. And then that will bounce off those counter arguments and then maybe the blind spots that they could be missing.

Tatyana Mamut (23:39.915)

that's interesting.

Hmm.

I'm going to have to think about that. Thank you. That's an interesting thing about like, so what we do have right now when we have workflows, what we can do is that the way to manage complex multi -step workflows, the best way to do it.

and reduce hallucinations is to have very tightly narrowly defined agents that are doing no more than four steps of the workflow at a time and then passing it off to others. you could actually have, and then in between the handoff, you could have the manager actually assess the work and either accept it or not, accept it, like approve, it's like an approval step, right? You can actually have the manager approve or not approve. And if it doesn't approve, it sends it back to the agent to

and create and redo the work. But you could actually have multiple agents working in parallel. And then, and we don't have this yet, but we could actually have multiple agents working in parallel, right? And then the manager could choose the best.

Tatyana Mamut (24:55.264)

right, the best approach or the human operator, right, in the approval step could choose the best approach before the next, before the work is actually completed. We could do that. yeah, interesting.

Alp Uguray (25:06.717)

Yeah, duh.

Yeah, that could be something interesting. I think to like, starting from there, right? Like I think the agents, there's the approach of going through more process automation route of like the enterprise works at a certain process and a task, and then we execute those tasks. And then and some companies are robotic, and then it's become the stitch on a wound.

Right? Like, and over time the wound grows because the problem didn't really get solved. So one thing that I've been hearing a lot is that actually the, instead of approaching it as like, is a process and here's an enterprise that executes on it. And how can we retain the entire workflow and then the engagement of people within the company. And I think what, the way you are approaching to that problem.

Tatyana Mamut (25:44.141)

Right.

Alp Uguray (26:06.397)

resonates really well with that approach. And I'm thinking that just from the customer standpoint, what will be the certain steps to enable them, right? Like as they are adopting it, maybe the product, like product leaders or like different, let's say CFOs of this month, they'll adopt it for financials or adopted from different perspectives. What would be from their perspective

they should change their thinking. Like they are still thinking robotic, right? Like, but they should start to think a little bit more creative. Like how, how do you, what do you think about that?

Tatyana Mamut (26:46.74)

You said exactly the right word, robotic. So most internal applications of AI agents right now is basically just RPA. You've defined a workflow and you want a certain set of steps to be completed.

with the bots inputting certain pieces of information. And essentially you're just trying to like take the RPA process and put in an LLM. But then the LLMs don't, again, they're not drawing from a database. So they don't exactly input the exact things that you want. Remember they're interns, right? So they might get the inputs wrong sometimes.

And this is why I actually think improving those processes is probably the wrong place to start for most companies because they put in so much time and effort into optimizing the current workflows that they have. And it's really hard to rip something out of that workflow, even a small piece of it.

and introduce uncertainty into the process. Because remember, these are interns. They're really smart interns. They can do really creative, amazing things. But if you need them to write the exact same words in the same exact form field, if you need them to do data entry, they're going to get it wrong a lot more than just like an RPA system. So

In terms of enabling customers, and I think this is why most of our customers are actually building new customer experiences that create really amazing value as opposed to trying to retrofit existing complex workflows that they've optimized with RPA. Now we will get there, right? Like definitely AI agents will be able to do that, but right now it's too

Tatyana Mamut (29:00.622)

hard and people are spending a lot of time and money and being disappointed because they don't understand that this is not just a software system that is deterministic every single time. Right. It's a really smart intern that create create amazing new value and new experiences for your customers or employees if you let them and if you don't constrain them too much. And so the way that we enable our customers is

If there's a process, if there's a complex process that is currently working for you.

Tatyana Mamut (29:37.939)

Why change it, right? Why do you do like,

Alp Uguray (29:39.761)

Why do you change it, right? It doesn't need additional AI just to be cool.

Tatyana Mamut (29:45.994)

Yeah. And they, they're like, well we can save, you know, money later on. And I'm like, okay, or you can create a better top line right now. Right. Like which one would you like to focus on? Right. I actually think. So here's like, everybody has seen the chart of how quickly Chad GPT got adopted. Right. It was like the fastest adopted, you know, consumer software experience ever in the history of planet earth. Okay.

Why are companies not taking that learning and saying, Hey, there's something people find really valuable in the chat GPT experience. How can we take that insight and get that kind

upswing in our adoption or upswing in our customer satisfaction, right? By using this technology, like the it's really weird to me that what people are actually excited about is the fact that they have seen the chat. GBT creates amazing experiences for the end customers. And yet they're not building products for end customers. They're trying to like.

automate away their employees in some way. it's data driven. It's not data driven. The data, the data driven approach would be to say, my God, look at this massive adoption of this new UI paradigm. How do we drive our adoption with our customers with this insight?

Alp Uguray (30:59.529)

That's very interesting.

Alp Uguray (31:18.451)

Yeah. And I think they have the lack of, and I think this happens and this ties to my next question really well, because all of the enterprises and all of the teams at enterprises are there to execute, tend to innovate. And that kind of results in more of a short -term thinking. And in a way, the incumbents, like a new startup comes that maybe is doing a very niche, very specialized way.

of automating or replacing it with AI and then can completely blow the incumbent away. So my question is, this one I think is, if you think about the long -term vision of Wayfound, like maybe the interns over time, that the foundation model is getting better, fine tuning gets better, specialized AI gets better, they're gonna go from an intern to senior, to manager, to maybe a director level

intelligence and execution power. So like to the board as well. Exactly. Exactly. So at that level of scale, there are two parts. One is at that stage, how does an enterprise that's 10 ,000 employees stay innovative and stay lean? And do you see in opposing to that startups coming

Tatyana Mamut (32:18.03)

to the board, to the board of

Alp Uguray (32:43.647)

two employees, three employees, but 200 agents focusing on certain tasks tend to dethrone those incumbents. How do you see that layout?

Tatyana Mamut (32:56.687)

No incumbents ever get dethroned because a startup is more efficient. Incumbents get dethroned because startups create better customer experiences. If you are not losing your customers, it doesn't matter if somebody is more efficient than you. If you are continuing to win, this is the other thing, if you are continuing to win customers because you are creating great customer experience,

It literally does not matter how efficient you are or how efficient your competitors are. Efficiency is not why people win. Nobody wins based on efficiency.

Now you might win because of being of lower cost and efficiency might be one of the factors in being a lower cost service. But there's no, nobody buys something like, do you, as a customer say, I'm going to buy these, I'm going to buy Nike's or instead of Reeboks or the opposite, because Nike is a more efficient company.

Alp Uguray (34:00.799)

Because they deliver faster. That doesn't happen.

Tatyana Mamut (34:03.97)

Well, because they deliver faster, well, then you have to say that's what you're building. And if AI agents help you build that customer value, then that's the question. But the efficiency question is nobody's going to win based on efficiency. That's nonsense.

People just need to start thinking about why people just start need to think about why do I buy things? Do I buy things because Tesla is more efficient than Mercedes? Do I buy this thing versus that thing because this company is more efficient than that company? It's nonsense.

Alp Uguray (34:23.699)

What about the customer experience?

Alp Uguray (34:41.929)

And what would your take be if that was purely customer experience? Because AI agents can really drive improved customer experience.

Tatyana Mamut (34:51.094)

Yes. So if you are building amazing customer experiences that are with AI, AI agents, LLMs, anything else,

and customers in the market choose you over an incumbent, yeah, that's going to make the incumbent lose. Now, the incumbent at that point might try to get more efficient to please their shareholders. Remember, efficiency is something that pleases shareholders, not customers. But that only lasts so long, right? Because there's only so much

Bottom line growth that you can squeeze when your top line is not growing or decreasing. The game name of the game is always to increase the top line in a, in a business. If you want to run a successful business, your top line is really the thing that matters. Like I learned this at Salesforce. Salesforce had an executive training program and it was called leading for growth and the top like insight there.

was at the end of the day, you've got to stay focused and as an executive on the top line, because that's the only way to make the bottom line work in the long term.

Alp Uguray (36:13.097)

And I think that will take like a lot of enablement education and advising so that companies change their mindset on how to adopt an AI agent so that their thinking wouldn't be so much on bottom line, but more on top line and more customer experience driven metrics to measure their outcomes versus more efficiency focused KPIs instead.

And in your perspective, what, like based on the customers you're interacting with, of course, there are the ones that are already thinking ahead, right? They get it. Yeah. They already get it. So, so from the ones who don't get it, right? Like, or, they could get stuck in their approaches and then so much process focus, so much efficiency focused.

Tatyana Mamut (36:51.586)

They get it. Look at someone who will love Weigh Sound, already get it, right? I don't have to keep this from anything.

Alp Uguray (37:11.527)

what would be the one thing they should do to think differently.

Tatyana Mamut (37:21.678)

going to answer this slightly differently. I were started, so you know, I've, was a consultant for the first half of my career. If I were starting a consulting firm right now, this is all I would talk about. I would talk about the fact

The reason why we're entering this trough of disillusionment with generative AI is because people started in the wrong place, right? They actually were not taking a data -driven approach to where generative AI really shines right now. Again, just look at that Chat GPT chart, the customer adoption of this amazing new UI, of this amazing new product experience is the thing that excited everybody.

Right? And so I would be starting a consulting firm.

and talking all over the world about how to take that adoption curve and apply it to your products and to your company in order to win in this next era of transformation. The next era of transformation, we'll see a lot of incumbents lose, but they're not going to lose, again, they're not going to lose because they're not efficient enough. They're going to lose because they're going to be native generative AI startups.

that are going to be creating incredible customer experiences with AI agents. And those are going to be dramatically better alternatives for customers than what they have without generative AI.

Tatyana Mamut (38:54.444)

And you can just look at the internet. Like imagine if when the internet happened, everybody was focused on just making their companies more efficient with the internet as opposed to building the Amazons of the world and building the Spotify's of the world and building the net. Like who were the winners in the internet revolution when the world shifted to the internet?

It's the companies that were thinking about how to create transformative customer experiences with this technology, not the companies that were, you know, trying to become a little bit more efficient with the internet by taking away memo writing and replacing it with emails. I mean, again, it's as if we're entering the next internet age and everybody is focusing

how to like implement email communication as opposed to memo and phone call communications in their companies as the way to win. Like what? Like so why would you do that? Like it's nonsense. Just look at history. Look at history. Look at history, right? The companies that won in the 1990s were the ones that were creating the great customer experiences, not the ones that were focused on.

Alp Uguray (39:49.244)

Yeah, why would you do that?

Tatyana Mamut (40:04.212)

automating their internal workflows with, you know, with the internet. Later on, once people figured that out, then people started to say, okay, now how do we automate some of our internal workflows with the internet? But again, nobody won. Nobody won because they did that. Blockbuster could have automated the crap out of their, you know, out of their internal operations. You think Blockbuster would have won versus Netflix?

Alp Uguray (40:32.167)

Yeah, that's a little lost story.

Tatyana Mamut (40:36.47)

No, but I mean, honestly, imagine if block, mean, I'm sure block blockbuster did right. They, automated a lot of things, right. And on their internal operations with the internet, right. They, they, they use the internet for inventory management. use the internet for, you know, better, like, you know, basically getting you your whatever blockbuster membership updates and things like that. did a lot of that backend automation. Did it matter?

Alp Uguray (40:43.199)

Dated.

Alp Uguray (40:47.593)

but they didn't change the experience.

Alp Uguray (41:06.399)

Yeah, I think that efficiency at the end of today made them forget what the customers really want or how that customer experience evolved to become and which Netflix came and then serve a completely different way to engage. that ties actually really well to the question that the previous guest would like to ask you. So I have this new format. I ask every guest to ask a question to the next guest.

So the question was, as companies think about customer experience, and if you think about the way foundation models present themselves, so like the chatbot interface, and the way a customer adopts it, how does the user interface design would evolve from more of a chat kind of approach to another way to interact?

with the computers, maybe there's more of an imaginative thinking that what do you think it's become in the future, as well as more of maybe how chat GPT and clouds take on it. Where chat GPT is more conversational, cloud has this new artifacts feature come out that maybe render different applications people are building or rendering that code. What is your take on this? What do you think about it? And then given

I would love to hear how do you see that UI interface change over

Tatyana Mamut (42:41.436)

you.

Tatyana Mamut (42:57.9)

Technology has always been able to do a lot of cool things. Two years ago, if we were talking about this, we would say, how is the metaverse going to change app experiences?

And I would have said, I don't think it's going to change it very much at all. I think it's a shitty experience, right? It's like not something like who wants to wear a heavy thing on their head to interact with an application? Like, do you want that? Like who wants that? Nobody wants that. Right. So here's the thing. Human culture has evolved in many ways. And if you believe in evolution, the way that humans interact with each other,

is how we would want to interact with other intelligences. So do we interact with each other? Do we like to interact with each other through interpretive dance? No. We like to interact with each other through verbal communication, primarily. Primarily. And yes, we like to talk to each other. There's going to be some...

places where voice is an interesting modality to get input from other intelligences and applications and companies. However,

As you know, you look like a crazy person if you're talking to your phone in public, right? And there's nobody else at the end of the line. Like, you know, it's like, it's just, it's just not, you can like, you can do anything, all that stuff, right? But it drives people crazy when Siri turns on by herself because it's not anything that anybody really wants unless they're driving, right? Or something like that. So,

Tatyana Mamut (44:53.086)

I actually think that what we're going to see in terms of a UI shift is that it is crazy to me that websites require users to find the menu item and the button and then scroll down to the thing to try to find the information that they want. That's not how anybody wants to interact with anything, right? They want, they want to come in and they want to do the chat GPT thing, which is like, tell me about your company. Tell

about what you do. I have this type of company. What can you do for me? I have this problem. What can your company do for me? I actually think the chat interface, like the actual human type of interaction is exactly where people are going to be going. So if you go to wayfound .ai, like we're trying to

Telepe... We still have to build the normal website for frankly Google SEO purposes, but the better experience is just to say, we have starter prompts too, so if you don't know what to ask, we give you something where you can just with one click get the conversation started. you want to actually, don't... Why do websites require us to scroll and click and find the button and find the menu option to find the information

want. No, I'm here, right? Like, it's like you're walking into a store. When you walk into a store, the greeter doesn't come to you and like throw the new merchandise in your face, right? She doesn't do that, right? She says, can I help you find something?

Right? And you say, yes or no. Yes, yes, I'm looking for this thing. And then she walks you to that thing and gets you the thing that you want as quickly as possible. Or you say, no, I'm just browsing. In which case she says, okay, let me point you in the direction of all the things that you can find here. That's how humans like to interact. Websites are shitty. And the reason websites are shitty is because they're built more for Google to SEO.

Tatyana Mamut (47:01.206)

right, than they are for customers to actually have good experiences, right?

Alp Uguray (47:06.303)

Yeah, that was fantastic. Yeah, I fully agree with that. I think the one thing in the beginning of the episode you mentioned was also the sales side of it as well. Because like when I opened the way found, like I didn't see a lot of like sales and that's like trying to sell me something, but it was more for me to find answers to. I really liked that experience.

And I think then over time, we will see maybe the applications, or the chats, be the interface between us and the internet, right, more and more over time. So that finding information or what we want out of a website is just easier.

Tatyana Mamut (47:56.384)

Yeah, we've been talking about conversation. I designers have been talking about conversational UI for over a decade and it has, there's never been a good experience to really bring conversational UI to life. Now there is, but we're, but again, our workflows are so optimized for like, if you go to another, like, you know, chap, like a traditional chat bot on a website, the first thing we like, what is your email? Give me your email. Give me your email.

Alp Uguray (48:20.896)

Yes.

Tatyana Mamut (48:21.974)

I want to ask you a question like, give me your email, give me your email, give me your email. And you're just like, F you, I'm leaving, right? You're walking into a store, a website is a virtual storefront. Imagine if the greeter was like, give me your email, give me your email, buy this, buy this, buy this, let me show you this, let me show you that. Like that's literally what like what most B2B websites are. It's like the greeter shows

Alp Uguray (48:27.707)

Yeah, just close the tab,

Tatyana Mamut (48:47.648)

and yells at you, give me your email, give me your email, give me your email. It's a terrible experience. would like you would walk out of that store? Why are you doing this to your customers? Why?

Alp Uguray (48:56.541)

Yeah.

Alp Uguray (49:01.129)

And I think it's just purely SEO driven design that and websites copying each other to be standardized and they follow their own standard, right?

Tatyana Mamut (49:11.51)

Yeah. I mean, again, we still had to build that standard SaaS website behind the scenes for Google, but we know that customers Google wants it. So we built it for Google,

Alp Uguray (49:22.405)

Hahaha.

Alp Uguray (49:27.487)

Yeah, that is very true. So I have one question to you. So for you, if you were to ask one question to the next guest, what would it be? What are you curious about these days?

Tatyana Mamut (49:45.39)

If you were 13 years old right now, then you were looking about how the world is changing.

Tatyana Mamut (49:53.774)

Would you be optimistic or pessimistic? And if you were to be optimistic, what would you be studying? What you would you, and when people say, what do you want to be when you grow up? What would you say? And by the way, all of us are 13 year olds right now because I just turned 50 literally a few days ago. Thank you.

Alp Uguray (50:17.232)

Happy birthday.

Tatyana Mamut (50:19.374)

And I still have at least 50 years ahead of me. So if I think about the fact that I do have to like the question of what are you going to be when you grow up? What do you want to be when you grow up is relevant to all of us because the world is going to be so incredibly different 20 years from now. And all I'm still going to be working 20 years from now.

For sure, I'm still going to be working 20 years from now. So what do I want to be when I grow up? What do I want to be 20 years from now? So for your guests, what do they want to be 20 years from now? And if they were a 13 year old right now, what would they want to be 20 years from now when they grow up?

Alp Uguray (51:00.927)

I love that question. I love that question. It's, it is deep in the sense of, right, like there's a timelessness, right? Like of what we do, what we love to do. And there's also the macro conditions of, like people being pessimistic and optimist based on the events that happen. It always goes back to the times of, like in different snippets of time, if they were to ask that question, I think this will be my answer.

But I think in each time, I think in each, like if you go back in time and think about like the bad times, right? Like World War II, World War I, and when things were great and ask the people, like, are you optimistic about future? I think it's the human nature to find the path outside, to not to fall into despair.

Tatyana Mamut (51:31.318)

Yeah, what would be your answer? What would be your answer?

Alp Uguray (52:00.095)

So I would still be hopeful that things will get better, right? Things will get better in some sort of way.

Tatyana Mamut (52:11.944)

There's a saying in Russian, there's a saying in Russian I want to tell you

Alp Uguray (52:15.763)

Yes, please do.

Tatyana Mamut (52:17.87)

When you think you've hit rock bottom, somebody knocks from below.

Alp Uguray (52:25.789)

There's always a worst of every situation.

Tatyana Mamut (52:26.882)

There's always works. You think this is not as easy for me, but there's always something works.

Alp Uguray (52:37.003)

That's so true. Well, thank you very much, Tatjana. It was a pleasure to have you again on the episode. And I hope we will meet again, maybe for round three as well. But very exciting things are ahead. I'm very excited about what you're building. And I think a lot

Tatyana Mamut (52:56.032)

I love that.

Alp Uguray (53:04.767)

customers and enterprises, this is what they need. So I'm excited about

Tatyana Mamut (53:09.03)

And I would like to offer to your listeners, you know, one month free of the full Wayfound platform. If they want to just go to wayfound .ai, chat with our agent, you know, say, you know, I want, know, Al told me I could get a free one month, you know, access to Wayfound and we will get your listeners hooked

Alp Uguray (53:33.907)

Yeah, perfect. Thank you very much. It's very kind of you. Thank you, Tatyana.

Tatyana Mamut (53:36.492)

Yeah. All right. All right. Thank you.