The Art of AI Implementation with Rudina Seseri

00:00:00:21 - 00:00:28:00

Ben Hillman

Imagine you're attempting to make your favorite cocktail, The Blender. A crucial tool promises to unify disparate ingredients into a harmonious blend of flavors. However, misusing this powerful device, adding too much of one ingredient or blending at the wrong speed can lead to disaster spoiling the cocktail and wasting precious resources. This is exactly what can happen when it's not strategically implemented within your business.

00:00:28:07 - 00:01:07:00

Ben Hillman

Sure, A.I. has the potential to solve intricate problems, but if not handled correctly, it can create even more complications. In the realm of B2B sass. II is like a high tech blender with the potential to revolutionize your services. A.I. can enable you to process customer needs, streamline your operation and uncover hidden flavors you previously didn't think possible. But the allure of AI, like the mesmerizing whirl of a blender, can be deceptive, leading with the fascination of something being labeled as A.I. rather than concrete outcomes can lead to mismatched expectations and ineffective implementation, creating more problems than it solves.

00:01:07:23 - 00:01:29:09

Ben Hillman

But don't worry, that's where our expert mixologist, Rudy Necessary steps in. Rodina, the founder and managing partner of Glassman Ventures, has mastered the art of blending A.I. into businesses. She stresses leading with outcomes, not with the shiny prospect of A.I. for the sake of A.I.. It's about using AI as a tool to achieve your goals, not as a novelty with readiness guidance.

00:01:29:10 - 00:01:57:15

Ben Hillman

You can avoid the pitfalls of misuse and learn how to leverage A.I. to solve real world problems in your business. Creating the perfect blend of innovation, efficiency, and growth. In today's episode, Radina will unpacked complexities of AI implementation, discussing how to lead with outcomes and avoid the common traps that can turn your A.I. dream into a nightmare. So join us as we navigate the exciting, yet challenging world of AI and B2B sass Learning from Regina's expertise.

00:01:57:16 - 00:02:20:06

Ben Hillman

How to create the Perfect A.I. Blend for your Business. Prepare to transform your approach to A.I. and get ready for a taste of success. From Paddle, it's Protect the Hustle, where we explore the truth behind the strategy and tactics of being a business growth to make you an outstanding operator. I'm Ben Hillman, and on today's episode, Rudi Necessary speaks with Andrew DAVIES about strategic implementing A.I. within your business.

00:02:20:11 - 00:02:39:14

Ben Hillman

They talk about venture capital, artificial intelligence and machine learning, business strategy, data privacy and security, as well as digital transfer machine and tech industry trends. After you finish this episode, check out the show notes or a field guide for today's episode. Then, while you're leaving your five star review of the podcast, tell us what resonated most about our guest advice.

00:02:46:21 - 00:02:55:20

Andrew Davies

Routine or Rice We just jump in here. Why don't you give me a bit of a background to yourself and also your entrance into the investing world and what you've been doing since you founded Glassman.

00:02:55:21 - 00:03:28:01

Rudina Seseri

Thanks, Indra, and good to see you. So my background, I have been in the technology slash venture capital investing world for nearly two decades now. Coming up on my 20th year. I have been investing in early stage companies as the first investor in for the bulk of that period. But my background prior to venture capital was at Microsoft and the prior to that in the technology group at Credit Suisse in the early 2000, like 99 and 2000, where I called my tech bag, as I often say.

00:03:28:01 - 00:03:43:21

Rudina Seseri

So I left the investment banking hours and fell in love with the power of transformation and innovation and really wanted to get on the operational side. Microsoft gave me that ability and then really wanted to help founders build companies. So that's me in a nutshell.

00:03:44:02 - 00:03:49:11

Andrew Davies

Why don't you talk to me a little bit about your investment thesis, a glass wing, the type of companies you invest in, and why.

00:03:49:15 - 00:04:15:20

Rudina Seseri

Glass wouldn't Ventures is an early stage first capital in focused firm, and we're investing in companies that are leveraging. I am Frontier Technologies for the enterprise and security market. It's very much focused on end market adoption, where A.I. or some other frontier tech facet can disrupt an existing market for the better or create entirely new markets and transformations.

00:04:15:23 - 00:04:39:21

Rudina Seseri

In that vein, we have a really sort of in my view, differentiated approach, because from day one we are thinking not simply about how you build a SaaS company, which of course you have to worry about the business model and the go to market and perhaps we can touch on that a bit later. But also how do you build an AI native company, which is not exactly the same as your run of the mill software?

00:04:40:00 - 00:04:59:20

Rudina Seseri

What techniques are better for what approaches? What problems are you truly solving? Are you optimizing your value prop and your technology for those problems and fallacies that I suspect Andrew, you are abundantly familiar with given your background. As we do that, keep in mind, you know, when we're investing, it's typically one, two or three founders and a few developers.

00:04:59:21 - 00:05:20:05

Rudina Seseri

The product market fit risk is fully blown. It hasn't been addressed. So the way glassblowing is set up, in addition to being thesis driven and focused on, is that at the partner level, we have a partner for every leadership and executive function of a technology company. So instead of seeing our investment partners, we actually think of it as our building partners.

00:05:20:05 - 00:05:47:16

Rudina Seseri

We have one of our partners is the former CEO of Nuance. One of our partners is the former CEO of a number of companies, including Chorus, AI and Reprise. One of our partners is the former CMO of LogMeIn and Rapid7. So we have that operator group, if you will, in the table making investment decisions all together. Why? Because once we go into an investment, we are not simply, okay, use the capital.

00:05:47:16 - 00:06:16:02

Rudina Seseri

I will come and pontificate every board meeting. It's about viewing ourselves as extensions to the team and really helping them build. Okay, we're going to get the design partners or customers. Let's go and so we're incredibly involved. And some companies are making decisions around what techniques to use. Others are making decisions around or positioning pricing. We want to be able to help them from day one with the ultimate goal of shortening the time and the effort to product market fit eyeballs.

00:06:16:04 - 00:06:17:09

Rudina Seseri

That does it give you a sense.

00:06:17:09 - 00:06:29:20

Andrew Davies

What are we just quickly define for a friend who's listening, then in your words and in your focus as a fund? What is A.I. and what are Frontier technologies? What would be some examples of things that fit within that and some things that would clearly fit outside of.

00:06:29:20 - 00:07:09:22

Rudina Seseri

It in the broadest sense is I think a is, you know, the newly emerging intelligence. Right. And not necessarily mirroring human intelligence. This is a different type of intelligence, but the intelligence that eventually will cut across all facets of technology enable companies, which pretty much will be, I think, any company out there and I often say that five, ten, 15 years from now, we can jointly laugh at the notion of having an AI focused fund because it will be so pervasive, much like we wouldn't have an Internet focused fund today, but for the next set of, you know, iterations around AI, adoption and innovation is a different type of technology because, you know, you're applying

00:07:09:22 - 00:07:37:05

Rudina Seseri

AML and other techniques. Therefore a lot of domain expertise is needed. So the AIS, when you're bringing intelligence, where in our view, the what the how in the current in age has two facets to it in our view, some form of learning, whether it's machine, machine learning, broadly defined, but whether I don't know anything around generative AI so hard right now, generative AI, whether it is deep neural nets, whatever techniques one might be using and data.

00:07:37:05 - 00:08:02:13

Rudina Seseri

So that's the how and the buzz is on the what the how is the more complicated piece, because not all techniques are created equal and not equally valuable to various value propositions in solving problems. So that's how we define AI is very much around narrow AI. Right. You know, in our current investments, by definition, we're not looking for AGI or anything of that nature just yet.

00:08:02:13 - 00:08:26:20

Rudina Seseri

Although the most recent advances have made that much more of a possible reality, Frontier tag is really focused on other very, very cutting edge technologies that can be applied in the 3 to 5 year time window. So, for example, we have a portfolio company called Chaos Search that is really bringing down the cost and complexity of acquiring data in cold storage.

00:08:26:20 - 00:08:55:12

Rudina Seseri

So think about observability needs for security data legs. Why I fund focus on that? Well, data is expensive, is hard to manage and junk in junk out, high quality data in strong outcomes. So from that perspective, that company sits on our frontier tech in that they're leveraging on a whole new approach to separate storage from compute. And more fundamentally, they're really leveraging object storage technology and creating a new type of database wise.

00:08:55:12 - 00:09:20:17

Rudina Seseri

It now deep tech, I feel like I'm asking the questions for you, but just to draw that distinction, because deep tech has a five to 10 to 15 year adoption cycles, whereas the frontier tag that gets leveraged in the innovation that we bag has a 3 to 5 year cycle for adoption. So so differentiation on how quickly can we solve a problem and how fundamental is the technology and transformative that we're using.

00:09:20:17 - 00:09:21:19

Rudina Seseri

Did I draw that distinction?

00:09:21:19 - 00:09:40:03

Andrew Davies

Andrew Yeah, that's super helpful and totally understand given the stage of investments you're making a probably the ticket size of investments you're making, how that time line is really, really important to your investment thesis. My first office software business that I cofounded was a it was using natural language processing and a bit of email to do B2B personalization, and this was ten years ago.

00:09:40:03 - 00:10:00:07

Andrew Davies

So I know the the heartache of trying to work through whether we sit on the shoulders of giants or build a bunch of stuff ourselves and and then also trying to explain what can be some quite complicated propositions around automation and sometimes especially in the early days, unbelievable proposition was around automation, which in an enterprise often aren't received particularly well until well, until you package them.

00:10:00:13 - 00:10:23:11

Andrew Davies

Over the last few years and particularly the last year, in 2022 and early 2023, we've seen the buzz around AGI and the buzz around your open AI and the response from Google has gained news share everywhere. Could you just talk us through what you think has happened for that to gain such public attention in such a quick order, even though this is, you know, the research pathways have been building for the last 20 years or so.

00:10:23:13 - 00:10:49:20

Rudina Seseri

I would love to build on something you said around we sit on the shoulders of giants and then we iterate, innovate and disrupt from there for context, open A.I., at least in this non for profits, you know, side has been around for many, many, many years. I think what makes it particularly relevant to sort of give us the sense that it happened overnight, a decade that feels like an overnight phenomenon.

00:10:50:05 - 00:11:15:07

Rudina Seseri

It's actually a technology back, you know, breakthrough in that fundamentally two things happened. So the emergence of embeddings, continuous vector representations of text that are now must haves in in any natural language process thing type offering. And secondly, the transformer capabilities, the occurrence of both of those phenomena on the technical side led to openai. I said, you bet.

00:11:15:12 - 00:11:35:00

Rudina Seseri

And and not I think we're version three. We're waiting for version 3.5 or whatever the case might be. But for those evolutions and revolutions to occur and is a confluence of those two phenomenon, those two breakthroughs occurring at the same time and the adoption that really overnight it made it feel like as though overnight you know, the landscape changed.

00:11:35:00 - 00:12:01:16

Rudina Seseri

So that's one piece. The other piece is that this notion of separating hype from reality and what I mean by that is over the last since 2010, 2012, we've been seeing different facets of AI actually gain traction in the enterprise. I shared with you that we are the thesis driven investors. My first AI focused companies were right around the 2010 timeframe.

00:12:01:16 - 00:12:23:14

Rudina Seseri

We developed a thesis and we could no longer talk about events that did our advanced analytics. Something else had happened and that something else was really deep neural nets. So 2006 you had deep neural nets, the giants of shoulders, right? Or the shoulder on the shoulders of giants, rather. Let me start. Or still in 2006, we had the the advent around deep neural nets.

00:12:23:14 - 00:12:49:19

Rudina Seseri

So your notion of sitting on the shoulders of giants, then we have continued to iterate new techniques. So we've come about and here we are again another leap on the in directions, on the, on the broader awareness. What Chen has done is it allows in the average person to go and type something and get an outcome. And the results, at least at the consumer level, feel quite impressive.

00:12:49:20 - 00:13:18:13

Rudina Seseri

Now between that and the adoption for the an end and prized, there is an ocean in between. And in fact if we come back to the everyday reality of founders building AI first companies in many, many, many instances. Andrew you don't need generative. There are other techniques that are a lot more useful and a lot lead to much better and concrete outcomes and value creation without needing the sophistication of or the buzzword related to generative AI.

00:13:18:16 - 00:13:19:10

Rudina Seseri

Although I'm a fan.

00:13:19:13 - 00:13:47:09

Andrew Davies

What if the most what what are the triumphs has been the interface and that some people have described JPT as direct to consumer AI, this idea of bragging rights to every single person and certainly, you know, the eyebrows have been raised in my family when I've opened up that interface and tried to give it instructions, have like my dad's a chartered accountant for small businesses and showing him prompts that would immediately replace a huge amount of his day work certainly were very eye opening to him.

00:13:47:14 - 00:14:07:11

Andrew Davies

But I think the really interesting counterpoint there is of confident wrongness, and that's probably a flaw that is being exposed as people are starting to dig into playing with chat. And also we saw the launch of Bard from Google in the opening advert, a very confident wrongness. Is that part of that gap you see between what we've got now of what is needed for enterprise adoption?

00:14:07:11 - 00:14:13:23

Andrew Davies

And you know what, Angela, you taking on helping founders think through overcoming that confident wrongness that we're seeing? And it's actually been.

00:14:14:01 - 00:14:29:18

Rudina Seseri

A myriad of questions there. Let me let me decouple them. First of all, I will compare your dad to my nine year old. I have a nine year old daughter who, you know, got a bit of exposure and said, but of course, this is table almost like the table six is my choice of vocabulary. But this is expected.

00:14:29:18 - 00:14:49:20

Rudina Seseri

It will be fascinating to see the generational shifts between what we are used to are our parents to the AI native generation and what they will perceive, not just the use of something like change. You bet, but the human in the loop world, the newly defined relationship with death. So I'll leave that out there is it's it's an interesting notion to ponder.

00:14:49:23 - 00:15:15:02

Rudina Seseri

How do I guide my founders on the question of, hey, we're see we're seeing the tragedy. And then the other solution for that matter, that's AI powered is prone to error. I ask beg the question why humans are perfect and or why is technology in end prior to AI perfect? So there is because it's almost mirror in certain ways human like interactions.

00:15:15:02 - 00:15:37:22

Rudina Seseri

We're trying to humanize what's ultimately technology and have expectations that are or are not realistic. Again, I'll leave that there. When it comes to the enterprise, it's less about all this will not be adopted because we got this. There is much more about predictability overall before someone uses, you know, builds on that foundational model, whether it's an app or some sort of solution for solving a problem.

00:15:37:22 - 00:16:00:10

Rudina Seseri

That's it. Customer success for the enterprise. What data is it being trained on? Are you you're in Europe, you have to deal with GDPR and all those joys. Are you in violation of regulations? Is it rely able, is it going to understand, you know, but if it's being leveraged for different sets of domain expertise, is broad, is it deep enough or what does the app need to do?

00:16:00:10 - 00:16:17:10

Rudina Seseri

Appliance matters If it's regulated industries, we need to take what's a raw technology and turn it into a product with all the compliance and expectation requirements for an integrated enterprise grade product That's expected. And I can go into some examples as well, But does it make sense?

00:16:17:14 - 00:16:20:11

Andrew Davies

Yeah, it totally does. And yeah, once you bring it to life, the few examples.

00:16:20:11 - 00:17:06:06

Rudina Seseri

So I'm thinking of a particularly one company that we have called VeriSign in the really inventory optimization space. They're using a myriad of extra combinations of ML techniques. But what they are doing is fundamentally bringing visibility into the MRO, which is really the replacement parts or the machine parts to produce goods. They are creating on average, 30 to $40 million savings for any given large enterprise because they're able to identify poor ordering, reordering, overstocking, last minute ordering and unnecessary premiums that are getting paid in that entire statement while I said that they are combining different AML techniques and didn't get into the, you know, nitty gritty of how we can we can have a feast

00:17:06:06 - 00:17:39:22

Rudina Seseri

in that regard. But I started with what the value prop is and immediately what problem is solving and the magnitude of the problem in solving. And I share that because there is a real danger for founders that, you know, our leveraging or building a native companies to get the sense that there is a they're there because they're getting a huge response from possible customers on account of positioning themselves as as a company or generative AI is the flavor of the day and it goes both for their prospective customers and their venture and venture capitalists that they're pitching.

00:17:39:22 - 00:18:04:21

Rudina Seseri

And I warn my own founders to be careful of what I call the fake interest syndrome, which is every enterprise in some fashion or another has his I or didiza design assets or digital transformation is a one, two or third priority. Strategic priority from the boardroom and the CEO down. So guess what? The likes of you and me will get on the phone who for part of those enterprises.

00:18:04:21 - 00:18:33:05

Rudina Seseri

And we will want to learn. If someone says, I have an a generative AI solution that will solve your problems, it's much less about the actual solution and about the possibility of leveraging it. It's much more of an educational journey. So I encourage founders to actually start thinking and leading with what problem they're solving, how they're solving. And by all means, ride the generative AI wave, lead with it, and then make sure they know that you are the for our own frontier of AI.

00:18:33:07 - 00:18:38:20

Andrew Davies

It's not just the fairy dust you can sprinkle over your investment pitch. There's got to be some value proposition that's making sense to the enterprise buyer.

00:18:38:20 - 00:18:57:17

Rudina Seseri

I pray to God, yes. Yeah. Especially if you do. I you know, Becky, I companies for a living. That fairy dust all disappears very, very quickly. 5 minutes in, we're talking about depth of techniques and data sources and structured, unstructured, etc.. Said rapid. We're talking more about go to market and how do we build a bid company.

00:18:57:18 - 00:19:22:15

Andrew Davies

I almost see that there's a continuum here and the best companies kind of solve that false dichotomy and have both. On one end, you've got the value proposition of a real solving, a pragmatic problem. And on the other side, we've got the fact that that tech is yours and the data set is yours and it's proprietary. We're seeing a huge amount of companies spin up based on a small interface or layer or application on top of open air or other sets of APIs.

00:19:22:15 - 00:19:33:18

Andrew Davies

Clearly, you'd want both of those to be true in most of the companies you'd invest, but which which one do you look for first and which you have more bias towards a very clear and pragmatic value prop or something proprietary and owned by the business.

00:19:33:18 - 00:19:54:02

Rudina Seseri

Guys don't make the world so binary. Those are hard choices. And then this is a good obviously that I am. I will give you an answer that doesn't align on that paradigm, but I will go for execution every time. Let me start with that. So the idea around the value prop can be big or the technology can be incredibly differentiated, at least in its vision of what's being billed.

00:19:54:03 - 00:20:24:22

Rudina Seseri

The number one question is on either one of those. Can you execute and build? I find that technology or, you know, leveraged and products built because this a big distinction. Wonderful that you have great tech. What's the product technologies leveraged and products that are built to solve a real world problem have a much bigger chance of commercialize able success then amazing technologies in search of a use case from that perspective, especially if you don't want to be in research for 2030 years.

00:20:24:22 - 00:20:50:08

Rudina Seseri

I think having a value prop, or at least the thesis around the value prop, whether you're building a platform or a tool, is fundamentally important. If you're able to have proprietary tech proprietary early data and create this value prop kudos to you. If using, you know, foundational models that, you know, the incumbents have built and then you build your own layer of technology and your own layer of, you know, perhaps the proprietary offering on top of it.

00:20:50:08 - 00:21:08:02

Rudina Seseri

You know, just because you're leveraging a foundational model doesn't mean that you can have no proprietary taken on top of it. And you're solving a real problem and you're training your algorithms for that particular problem. So talk about the depth of data training in a differentiated manner and then executing. You have every chance of success.

00:21:08:02 - 00:21:15:09

Andrew Davies

You know, we we we're tracking about 28 billion of our off through our metrics products. We see a huge point of visibility into the success.

00:21:15:13 - 00:21:18:19

Rudina Seseri

Is that a pitch? Is that a pitch right there? What is the product?

00:21:18:19 - 00:21:47:12

Andrew Davies

We can talk for hours about that, but most of the people listening to this will know we have a free metrics product that you can plug in your your stripe dashboard or similar to. And so we're seeing all of financial data from about 25,000 different software businesses. One of the things we've seen spike in our dashboards boast in the and that free metrics promote and in our core payments platform, what we see the real revenue in the churn and always rates is the massive growth of tools built on some of the best of the technologies we've just been discussing, particularly on the sets of APIs from opening, etc..

00:21:47:13 - 00:22:13:00

Andrew Davies

One thing that is remarkable to me as someone who's in a go to market role, we both of you both use the phrase standing on the shoulders of giants is how easily strong value propositions can be built on top of what is there and therefore how difficult the go to market becomes. Because you suddenly, if we take Jasper as an example of a company that is done extremely well, there are, I don't know, 500 competitors for approximately what they're doing instantly.

00:22:13:00 - 00:22:26:18

Andrew Davies

The spun up very quickly afterwards. Talk to me a little bit about the go to market, the distribution that you look for, and any good examples of how companies have gotten to market with a complicated or a base value proposition in a way that's won the attention of the end user.

00:22:26:19 - 00:22:54:05

Rudina Seseri

A couple of thoughts in that regard. One, it begs the question, you know, you've referenced just for a couple of times, is there truly a first mover advantage? And if it is in that first mover advantage may not be on the, you know, foundation model itself, because, you know, they're using openness capabilities or if it's not that the core product itself is highly differentiated per say, is it in execution of go to market?

00:22:54:05 - 00:23:14:17

Rudina Seseri

So if you have hundreds and hundreds and hundreds of customers and I wake up one morning and I say, Oh, I love what Andrew's doing, I think I'm going to go after what is in the question is for you, how do you protect your stake, continue to innovate and back to sort of our paradigm discussion and shift the paradigm so it doesn't become a race to the bottom because there are 50 other players.

00:23:14:17 - 00:23:37:21

Rudina Seseri

So now you're competing on price. And that may mean that incremental products, that may mean if it is a platform additional and you use cases, you've got to figure out this is truly sort of an exercise in strategy. What are the barriers to it that I have and how can I make the switch and costs very, very high for my customers by, you know, delighting and continuing to expand the depth of my offering and the breadth of my offering.

00:23:37:21 - 00:23:57:12

Rudina Seseri

So that's one. But to use the term for these, you know, companies that are leading with a I please don't lead with I lead with, you know, outcomes, I go back to that earlier comment of these are the outcomes I deliver and I deliver to them because I leverage I and it may sound like a subtle but is a distinct difference.

00:23:57:12 - 00:24:22:17

Rudina Seseri

So you know, is is I reflect on our own portfolio. I'll keep I'll pick on a company that's in the security space called Black Kite. They essentially what they are doing is for the, you know, your supply chain ecosystem. So that could include your vendors, your partners, you all the way to your customers. To be honest, if you're an enterprise, they are leveraging AI to give you real time assessments of your vulnerabilities and exposures.

00:24:22:17 - 00:24:48:17

Rudina Seseri

So as the CEO said, the chief information security officer of that enterprise, you can both make some decisions in assessing risk and your appetite if it's a very important customer and they're exposing you do X, Y, Z risk, do you want to mitigate? Do you want them to mitigate? Is there a middle ground? And it can be leveraged all the way for risk and governance management at the public board of directors, staff level, the underpinnings, all sorts of techniques and data and training.

00:24:48:17 - 00:25:13:01

Rudina Seseri

But the value prop did they create is incredibly differentiated. Now you may say, okay, but why can't Andrew Rudy now wake up and do the same thing? Well, funny enough, these guys actually had incumbents in the markets, is still exist. The difference, though, is it comes from the amazing tech and what they have done around API, around this real time and continuous sort of ability to accurately predict vulnerabilities, deliver better outcomes.

00:25:13:07 - 00:25:49:05

Rudina Seseri

So that's where your tech and you know, and the continuous learning that occurs through email techniques gives you an edge. Customer support thought, leadership, your domain. Right. And you know, amazing sort of all end to end experience for the customer. So it's a NAND and it's not a silver bullet that addresses all you leverage tech to either gain advantage or disrupt and then you leverage execution on the go to market and continuous innovation thinking to make it difficult for the new comers who don't have your customers, who don't have your possibly your capital, and the ability to command pricing make it difficult for them to respond.

00:25:49:10 - 00:26:10:04

Andrew Davies

Do you have portfolio companies that are taking that product led routes to market? I think one thing that we're seeing again in our data is how particularly generative AI brings a value proposition so close to the customer, the freemium and the margin option there becomes really interesting. And usage based pricing allows people to show the value proposition before any kind of sales cycle.

00:26:10:04 - 00:26:12:15

Andrew Davies

So is that a mode you're seeing within the businesses you're funding?

00:26:12:16 - 00:26:36:12

Rudina Seseri

So it's interesting. I don't have a good answer for you there, and I'll tell you why. Yes, I do. But we have two opposing forces happening in the current market environment. One is a big push for product lead. I mean, we are investors in Ripley's, which is automating the demo function of software. I'm I'm a big, big, big, big fan and believer in that company as the market leader and market maker and category maker of course we do.

00:26:36:12 - 00:27:07:01

Rudina Seseri

By the same token, what I am seeing in the current environment, Android, it's really the early signs is that is enterprises are pulling back. You have layoffs. You know, you hear about the layoffs in the tech space, but you have layoffs and cost cutting across, you know, sectors. And the scrutiny around costs comes to bear. Two things become true customers require more hand-holding and, you know, a clearer and more sort of support in seeing the value proposition.

00:27:07:01 - 00:27:31:03

Rudina Seseri

And the cycles are slower. So if you have this, you know, something can go. What goes along with product led growth is typically shorter sales cycles and honestly, a several, you know, 20, 30, 40, 50 K not hundreds of thousands in invasive. Well, wait a minute, when cycles get longer, when more is needed and more touchpoint is it really does you know, product led growth hold?

00:27:31:03 - 00:27:58:18

Rudina Seseri

I don't have a good answer but these are two going hand in hand and equally strong forces. Again, you made the point generative. I could really push the product led growth movement farther and accelerate. And yet it's happening in an environment that's a recessionary nature. Which one will win? We're watching it very, very carefully because fundamentally, if it starts at four, I'm making it up 430 K a CV It's now taking you five, six months.

00:27:58:19 - 00:28:13:17

Rudina Seseri

Is that the red pricing? Is there, the red go to market? What do you need to think about the other? You want to have a profitable customer, you want to have a business, you know, mid to long term. So let's keep an eye on that. And I'm sorry I can't give an definitive answer, but this is the beauty of what we do.

00:28:13:18 - 00:28:28:16

Andrew Davies

Let's go to a couple of different tacks before we come back and finish here on on AGI and some other future gazing that I'm sure you'll be willing to do. What have the you spoken about a few challenges in crossing that chasm from smart technology to something that is enterprise ready? You know, we've talked a little bit about you.

00:28:28:17 - 00:28:55:23

Andrew Davies

You mentioned compliance is one of those. What are the challenges that that I've seen in a few ensure tech based startups around pricing of risk was genuine. DG Diversity equity inclusion challenges when trained on open data sets that haven't had full context. And so you've got decisions being made abstracted from real persons lives that can actually, if it's not, if you're not got if you're not putting guardrails in place, could have devastating consequences because of biases that are in the datasets.

00:28:55:23 - 00:29:07:10

Andrew Davies

Is that something you've seen? Is that something you've got companies that are actively addressing? It feels like as we give more and more power, decision power to two different forms of AI, that's a known consequence.

00:29:07:11 - 00:29:35:23

Rudina Seseri

I think for a long while. Still we will be in the human in the loop mode precisely for that reason. So we are making decisions in formed by our agent counterpart, not not to humanize technology, but for sure. But the decisions still lie with us. That's very important. Secondly, to your very valid point, single black mom automatically denied by an insurance provider.

00:29:35:23 - 00:29:56:09

Rudina Seseri

Or, you know, we can have countless of other examples, but that's a real challenge. Why noise in the system and biases in the system. So, Andrew, I think it was around 2020. I wrote an article with a provocative title called This Email, and I talked about the challenges that we have biases and noise, and those are two different things.

00:29:56:09 - 00:30:21:09

Rudina Seseri

But with the algorithms, if it's only man coding, if it's only women coding, what issues come with those? But even more dangerously, on the data side, even things like Why Series Voice a woman, Why was Watson a male's voice? Is, you know, is it because women have a secondary assistant like role? We can challenge it and say, This sounds so feminist, but why is it somebody did some testing?

00:30:21:09 - 00:30:40:16

Rudina Seseri

I assure you to find which one appealed. So it's more a reflection of what the market will support and adopt rather than someone started with the notion of, Oh, we will make Siri a woman because this is what I believe about women. Just think about that piece on the data sets. If we've had biases in our decision making, insurance, stare complex, whatever, recruiting.

00:30:40:16 - 00:31:04:07

Rudina Seseri

I mean, Amazon had a huge issue because for the first passes in resumé evaluations, they use prior datasets. Well, what kind of developers do you think they recruited? There was a lot of homogeneity, not just in gender but in backgrounds of these folks building compliance and building capabilities to address those biases and building what you know is commonly known as ethical AI is something that we care about.

00:31:04:08 - 00:31:34:03

Rudina Seseri

In fact, not only do we have it part of our lives, we investment evaluation process, but we have incubated a business that speaks directly to this notion called flux, not A.I.. It's still in stealth mode, but I will be sure to let you know when it comes out of stealth mode. But it's exactly speaking to the need not simply to address this problem, but also automate them because the way that they will be addressed or are being addressed somewhat haphazardly now is, you know, human based.

00:31:34:03 - 00:31:40:23

Rudina Seseri

And there's a lot of opportunity to leverage AI to drive better outcomes around against biases and noise in the system.

00:31:40:23 - 00:32:05:13

Andrew Davies

We've got, you know, thousands of SAS founders listening to this and many of them won't have A.I. as part of their core proposition but will be quickly Googling searching have been probably testing out chat GPT for the last two months daily. What advice do you have for businesses that are looking to reinvent what they're doing with some of the available, you know, APIs and technologies now, or looking to add ancillary services or products?

00:32:05:13 - 00:32:08:15

Andrew Davies

What's your advice for a founder looking in on what's emerging?

00:32:08:16 - 00:32:33:04

Rudina Seseri

First, remember that the air wave is much bigger than generative AI and for goodness sake, don't let on to that. I think retrofitting A.I. is not an easy exercise. I like this notion of think about how can you adopt it for products going forward. Secondly, start with I think I'm becoming a broken record, but what problem are you solving and what technique it's best it might be?

00:32:33:04 - 00:32:59:01

Rudina Seseri

General TBI Don't get me wrong, but it may be that, you know, OBIS and neural net is the best outcome or, you know, are an NS or some other techniques. I mean, they're supervised, unsupervised. Don't just rush in for the sake of checking the box because you have to picture this year after speech of respected customer and you have to check the box that you have AI that has real consequences and or you will get snuffed out somewhere in that process that you don't have.

00:32:59:06 - 00:33:20:11

Rudina Seseri

Instead, think about what am I problem for, you know, if I'm doing drug discovery? So the techniques are way better than other techniques or, you know, some basic techniques may lead to very, very strong outcomes or not. So start sinking by. What sort of if is a drug am I using? The CNN is probably best is probably better than general debate I will ever be.

00:33:20:11 - 00:33:30:00

Rudina Seseri

I don't know. So you figure out what techniques are best and start with the fundamentals and then put the lipstick, so to speak, around the pitch and how you're leveraging.

00:33:30:04 - 00:33:58:05

Andrew Davies

I you mentioned your your daughter earlier. I've got a 11 year old daughter and an eight year old son. Based on what you're seeing in the market, the automation, and there is obviously a consequence there on jobs and careers and what people will have valued careers doing in the future. What's your advice to my kids, age 11 and eight, about the kind of careers they should be looking out for, and particularly the kind of things that they shouldn't, because there's going to be automation, you know, eating those jobs over the next ten years or so.

00:33:58:06 - 00:34:17:15

Rudina Seseri

It will be interesting. I mean, first, are children provided that they have they're fortunate enough to have the ability to pursue. They need to pursue what they love. There's just, you know, a human view rather than a domain experts view. But I think chances of succeeding in something that you love in and have a talent for ideally are much higher.

00:34:17:15 - 00:34:52:02

Rudina Seseri

I don't think we should box our children based on what's coming down the pike or our perspective, you know, perspective like that. Also, I fundamentally believe that I will create more jobs ultimately than it will automate, and there is plenty of research that that supports that where we will see a negative disruption. I usually use the term disruption for the positive, but I think where we will see a negative disruption and on this, the societal consequences is around this notion of will the doctor of today remain relevant in the same position 30 years from now?

00:34:52:02 - 00:35:18:19

Rudina Seseri

Why do you and I do you and I want a very experienced doctor mean in the U.S. we have, you know, residents. I don't want to see a resident. I want to see the head of the department. Why this notion of experience and having seen it all, what happens. We'll still value experience, but what happens if seeing it old is now done by his copilot or her copilot, where the model has seen it all and can give you the outliers and what becomes relevant.

00:35:18:19 - 00:35:55:04

Rudina Seseri

And more importantly, is it more relevant that the doctor, him or herself see it all or that they know how to interpret the data and, you know, with their copilot? So I think the concern that I have is more how the definition of what constitutes an attorney. You know, the do they need to read all the law regarding their legal copilot and do so a medical doctor who can go, you know, field by field how their role will get partially automated but really changed definitionally in now it operates and we've seen this happen right in generally lower skill industries.

00:35:55:04 - 00:36:25:22

Rudina Seseri

When the world shifted to either got automated or shifted to cheaper labor. We've never really seen this for highly trained high income earning categories of labor, and that's where I think there will be societal consequences. What I don't know is, is it happen? Does it feel like it happens overnight and we have upheaval? Does it happen gradually? So the, you know, new generations catch on, kind of like you're seeing computer science majors now learning lots about AI as opposed to it being two separate these, you know, fields entirely, data science versus computer science.

00:36:25:22 - 00:36:32:15

Rudina Seseri

So we will see. But that's an area to keep an eye on. But tell the kids to be happy and enjoy being outside, not in front of their devices.

00:36:32:18 - 00:36:44:00

Andrew Davies

My eight year old son will be very pleased to know that his dream of being a professional soccer player is not going to be automated away by by some bots. I'm sure he'll he'll be delighted to know that this evening. And I do think.

00:36:44:05 - 00:36:45:21

Rudina Seseri

What position does he play? Hold on.

00:36:45:21 - 00:36:54:10

Andrew Davies

I think he has two dreams. He says he's either going to be the world's best goalkeeper or the world's best striker. They're the only two career options to him, so I'm glad he's aiming high.

00:36:54:13 - 00:37:16:05

Rudina Seseri

So I have to tell you, we actually had last week did an interview with Briana Scurry about six months ago, who was the goalie for the women's U.S. soccer team. It's a thankless it's a thankless job. So go for the striker. If you attempt to score five times and you score to your star, if someone attempts to score five times and you're the goalie and they succeed in do, you're a loser.

00:37:16:05 - 00:37:17:23

Rudina Seseri

So just just all in perspective.

00:37:17:23 - 00:37:42:15

Andrew Davies

And I think it's super interesting also to to just go back to your point of being a copilot, having conversations with our content team, who are some of whom are concerned about what they're seeing from all of this generative AI and recognizing that they're not going to lose their job to, you know, an API, but they might they might struggle to compete against someone who is using that as their copilot if they don't, you know, learn those new tools and techniques.

00:37:42:17 - 00:37:51:12

Andrew Davies

And I think certainly that's the conversation with my daughter. She's an avid user of Dali, too, in her in her design work at at a few of the other the other tools.

00:37:51:12 - 00:38:08:17

Rudina Seseri

Yeah. No, yeah. We use my journey. So it's interesting that notion because on the one hand it's like saying, hey, Excel got invented. Now I'm scared that I have to use Excel. Well, you know, we got over it, we got trained. That will happen over time. The bigger question I have is where is the push and the pull?

00:38:08:17 - 00:38:33:07

Rudina Seseri

If you're a creative type where your work is being used to train the algorithm for that broader so you can get, you know, lookalikes. And where does your proprietary and copyrighted work sort of begin and end and where does the innovation or the was being generated by the copilot begin? I think that will make that will come with some sort of best standards and or the regulatory environment.

00:38:33:09 - 00:38:36:15

Rudina Seseri

Another area, by the way, that should be interesting over time for an investment.

00:38:36:15 - 00:38:59:03

Andrew Davies

We've we've almost run out of time here. This is a fantastic discussion. I'm really enjoying it. Let me ask you to put on your future gazing hat one more time. We're seeing a lot happen right now. You bollards being released by Google as well as Microsoft's increased investment and entitlement with open A.I.. If we've rolled forwards a year and we're now in February 2024, what it's going to be hitting the news cycle concerning AI and what what things should we be looking at coming down the pipe.

00:38:59:05 - 00:39:32:16

Rudina Seseri

I mean, I think the the war of the incumbents will still be ongoing. There will be probably some continued breakthroughs and some early disillusionment because somehow we have this notion that whether it's Jedi beauty or other capabilities are to be perfect or perfect, like because there will be this push to now commercialize and build on top of these platforms, then that's where reality will settle and will set in that, Hey, you've got to really productize whatever you're building and it's great, but it's not all the way there.

00:39:32:16 - 00:39:43:13

Rudina Seseri

So I think we will have some healthy doses of reality. But the train has left. The station is just the question of is it going to go, you know, at 100 miles an hour, 50 or something else?

00:39:43:13 - 00:39:55:22

Andrew Davies

Well, that's a great a great note to end on. Thank you so much for your time. I've really appreciated the conversation and as I said, will be cutting an intro to an outro to all of this. We're making sure you get that and you get a chance to review it before it goes live, but I really appreciate it.

00:39:55:22 - 00:39:58:05

Andrew Davies

Thank you so much for talking so openly about it.

00:39:58:06 - 00:40:00:03

Rudina Seseri

Thank you, Andrew My pleasure and great to read you.

00:40:02:18 - 00:40:24:07

Ben Hillman

Shout out to Rodina for being on the show. Now you have a better understanding of strategically implementing AI within your business. Today we talked about venture capital, artificial intelligence and machine learning, business strategy, data privacy and security, as well as digital transformation and tech industry trends. Make sure to give Protect the Hustle A five star review, and tell us what lesson from today's episode was your favorite.

00:40:24:13 - 00:40:31:23

Ben Hillman

Thanks for listening. Subscribe to and tell your friends about Protect the Hustle, a podcast from Battle Studios dedicated to helping you build better SAS.