View Auto-Generated Transcript
The below transcript has been auto-generated for your convenience. Please reference the source video/audio for direct quotes or to clarify any errors.
Kobi: AI Is only the beginning of what I call hyper automation.
And that's the holy grail, An AI that creates an action.
What enables this is exactly the same.
Data, AI, and automation connected together. That's the key for any business today to optimize.
Connor: Hello, and welcome to the go to Market with AI podcast. A podcast about AI products, AI founders, and GTM leaders using AI in their work. In today's episode, I speak with Kobe stock founder and CEO of forward.ai, an AI platform helping GTM teams add AI functionality across their technology stack. Kobe and I talk about his path of serial entrepreneurship, his experience at WalkMe as a foundation for wanting to build forward, ways he's seeing innovative GTM teams leverage AI technology, what GTM teams get wrong about AI, and the skill set of future GTM leaders that will be most valued in the age of AI.
Let's get started.
Kobi, Hello and welcome. Good morning or good afternoon. Good evening. How many hours ahead of me? Are you I'm in New York.
Kobi: I am seven hours ahead of you. So it's a good afternoon.
Connor: Like good afternoon. Nice. Nice end of day like beginning of day podcast for me end of day podcast for Kobi.
Kobi: True.
Connor: Amazing. Amazing. Well, I know I'm really excited to talk about Forwrd. I know that that's how we originally got connected and Forwrd as an AI platform, and you'll tell us a whole bunch about all of that.
But before we get into Forwrd, I'd love to be able to start with, like, how did you find yourself in founding and leading and being the CEO of an AI company especially right now? And what was your journey into it? Because as I understand it from talking to you earlier you were maybe around this region before the hype train started, and you were in the right place versus jumping in a little bit late to the party.
So maybe let's start with wherever makes sense to you of what was your journey? How did you get to Forwrd itself?
Kobi: Yeah. First of all, I'm super excited to be here Connor. So, thanks very much. And second I'm a geek. I'm coding since I'm six. And actually I knew that I will do software like from, you know, super early age.
Connor: What were you coding at six?
Kobi: games mostly.
Uh, this
Connor: that was mine
Kobi: of the PC era, right? So we, we, like my parents couldn't afForwrd buying games and we, you had like books of teaching you how to code games in basic, it was back then.
But yeah. Doing that, I was early on the internet, obviously, and really started my career as a developer. I worked for a big company called SAP. And really, you know, kind of what I did, I basically helped to build kind of the, the biggest in memory database, SAP HANA, and then kind of rolled after SAP, I was always like, in my mind, I want to build something myself.
And I actually left SAP to found my first company. It was more than 12 years ago. And then from there, it was a very natural...
Connor: wait, what was the first, what was the first company?
Kobi: the first company was a company that teaches you how to play the guitar, actually. So I'm also a musician in my history. So I basically build an algorithm that understands what you play.
Which chords, which melodics and provides you feedback so you can be a better guitarist. And it was all about playing rock music. So we had like, you know, Green Day and Led Zeppelin and Foo Fighters
Connor: Yeah, yeah, yeah.
Kobi: that. So it's like, it was like an iTunes, but instead of buying the songs, you would buy how to play them.
Connor: Like, so like runs on your phone or what was the hardware? What was the...
Kobi: Phone, tablet...
Connor: my microphone listens, understands and gives you feedback on, on your play style.
Kobi: Exactly,
Connor: Super cool.
Kobi: Exactly. And by the way, this was my first touch in early, early AI, right? How can you, how can you solve that kind of interface that you analyze someone's feedback so fast and you try to kind of say what they've played, right? So, so that's the first kind of time where I kind of touched that.
And after that kind of, I kind of shifted between core engineering to product and to business and to go to market and really kind of discovered what's around me because of that. Then after this company, it was a six years journey. After that, I basically switched to product and took like few product roles at few start ups. And then I started my second company which actually we developed a cool product for mobile companies that helps them to understand when you're happy. And this was the second time that I touched machine learning, because if you think about it, when you are, a bank with a mobile app or an e commerce app, or even Waze, like the, the, you know, the GPS application.
And you want to ask your user something, if it's to write the app or to get them feedback on your next release, you want to catch them at the right moment when they are most likely to kind of respond to your feedback. So today, like, previously, folks are hard coding those points in time, but actually everyone is different.
So we tried to basically understand what's going on on your device on your graphic card, Excel, a metal jar scope. We took all of those signals, more than 300 signals and computed the happy moment. So that was my second touch in AI.
Connor: Is that just like, if I'm, if I'm button mashing, I'm, I'm probably pissed off is I mean, that's incredibly reductive, but is that sort of like the vein?
Kobi: You're pissed off or playing like hardcore game. So if you're touching the
Connor: Oh yeah. Contextual.
Kobi: and your graphic
Connor: be a lot of engagement. Okay. That
Kobi: exactly. It's it's, it's all related to a context, right? So if you're pushing strongly on the device and the graphic card is working hard, you're not pissed off. You're probably playing something, but if your graphic card stopped and you're pushing games,
Connor: Yeah, yeah, yeah.
Kobi: Probably something wrong, right?
And that's the magic of machine learning and AI, where it can support so many scenarios that when you try to code them manually, you just can't support them.
Connor: Yeah. It's impossible. Yeah, for sure.
Kobi: It's impossible for someone to do that. So, so actually interesting story about this company. I sold it to another company based here in Israel called WalkMe.
I don't know if you've heard about it.
Connor: I do. know WalkMe from, from deep Salesforce backgrounds. A lot of what Happily is doing is sort of inspired by a lot of that Salesforce ecosystem side and WalkMe is one that immediately sort of came to mind. And I think when we originally got connected, I was like, Oh, I know WalkMe.
This is interesting.
Kobi: Yeah. So WalkMe, I joined WalkMe we were around 300 employees, maybe less. When I left, we're above 1, 000 and we're a public company. So I kind of viewed the the business growing. And actually I was sitting in the, in the main junction of the company. I was the SAP of product. So really...
Everything that goes between customers to engineering. I was there. So really amazing journey. Tons of AI initiatives back then. So also touching AI a lot. I always, kind of for me I try to be not too deep, but to really get engineering and really to be in the midst of what's going on. and I left WalkMe around two and a half years ago to solve a problem that I solved internally at WalkMe.
And I said why not going and, you know, productize it. And that's Forwrd my third startup and hopefully last. I mean, that's what I,
Connor: You're like,
Kobi: I tell my wife,
Connor: of my journey
Kobi: that's what I tell my wife, but,
Connor: Okay. Okay.
Kobi: but that's the, that's the journey until here
Connor: So tell, tell me about Forwrd. What, what's, what's the problem? Where does it sit? What are you guys up to? And, and, and then we can talk a little bit. It's a sort of like, Why and where that fits in and I think what's interesting is before we jump into that is I think a lot of folks and it's been really interesting doing this show because I think that there's a lot of people who are very early to the AI boom and getting sort of drawn in by the gold rush and then there's folks who are sort of like, I've been in the mountains drilling and mining and everybody's showing up and I think you're, you're much more in that ladder camp with a lot of this given sort of the background and the space prior
Kobi: Yeah, yeah, for sure. Maybe I will go one step back. So, I'm a product person, right? And for me, product is always around the data that you can really put on your plate to understand what's going on, right? And I'm a big fan of product analytics. I've used all the products until today. And the reason that I started Forwrd is that in WalkMe we had a problem that we wanted to try and predict churn. And the way we wanted to do so is by utilizing product usage. And we said, you know what, let's take all of the product usage, all the features that we release. Let's measure. usage of different users and accounts. And let's try to see if we can correlate usage to churn. So we invented the new metric, we called it adoption score.
And we treated like a credit score in the US. Similarly, you know, zero to 800. So people can really easily understand it. So think about a customer success person trying to understand the health of their customers. They're going and through many dashboards. Salesforce, Tableau, some product data, asking questions on Slack.
It's not scalable. So what we want to do, we want to compress all of the signals and put like a traffic light in Salesforce. By the way, using Walkman, right? So we had the data. Then I built like a big team and we needed to buy a data warehouse. We needed to set up an ETL. We needed data operations people.
We need the
Connor: tons, tons of work to get this whole thing configured. Great.
Kobi: Tons of work. Listen, it took us seven months of
Connor: I mean, that's not, that's not even that bad.
Kobi: with a 20
Connor: Okay. Okay. So a lot of people making that happen on an, on
Kobi: a lot of work, right? To build the model. Then after we built the model, we needed more people to integrate the model back into Salesforce. So the CS folks can really see the data back in Salesforce. And after that, I realized that it's great that we built the first model, but the data is keep on changing. We need to constantly update it. And I stopped there and I said, there must be a better way. How can I productize it, so I can reduce the barrier? So I can reduce the number of people who work to do that.
I can reduce the cost of the infrastructure. And instead of a year, take it down to a week. And that's exactly what Forwrd is. Forwrd is basically automating this very, complex, heavy process that most chances that internally you can't solve it. And even if you can solve it, you're not focusing on your business.
So you don't want to solve it internally,
Connor: Do you think so? Okay. So if I'm going to go and install Forwrd, do I, what, what do I need in place already? So if, if you're sort of describing, Hey, we went and implemented this thing. We had to set up data warehousing. We had to connect all these systems. Is it, you're going to go and connect to everything that already exists.
I already have to have a data warehouse that's provisioned and then you're connecting to that, or where, where do I have to be on that maturity scale for something like Forwrd to be a value to me?
Kobi: the goal of Forwrd is basically to improve your internal day to day business processes. Let's pick a process. Let's pick, I don't know, lead qualification, for example, a very common process. Most companies solve it via a very simple lead scoring that they implement in HubSpot, right? Salesforce, Marketo, whatever, right?
And lead scoring up until now, it's basically a set of rules. I call it the casino model. Why the casino
Connor: casino model.
Kobi: The casino model, because you just guess.
If the title is a vp, if you have more than five content, if they answer an email, you, you just give them points. Right? But as I told you before, you don't know the context. If someone answered an email in an enterprise company, it, it, it should be weighted differently than if someone answered an email in a, in a small company. This is why it's the casino model because it's nonsense. it's not the truth. So, in order to improve the process, you already have the data in Marketo, right?
In Salesforce, whatever. So you connect the data that you already have to Forwrd, and you define what you want to optimize. So, generally, lead scoring or lead qualification, you want to optimize the number of sales qualified leads that you produce. So you provide Forwrd with this specific goal, how you define a sales qualified lead, and Forwrd will reverse engineer your entire data set.
will clean it, and we'll build an AI model that will show you which factors and values are relevant, are impacting your sales qualified leads, and will help you to predict on a daily basis, which leads will become SQL, and the same goes to sales forecasting, the same goes to identifying churn, upsell, cross sell, so think about all those processes that you have today, and you implement them manually, think about adding AI to these processes and think about what's the optimization that you can drive.
Connor: So are you guys building, so do I, do I bring my data warehouse or do I bring my amalgamation of business systems and start connecting those?
Kobi: it depends on how you operate. We don't want to move your cheese anywhere. If you are operating as a company and your kind of architecture is you funnel everything through your data warehouse, we will work with your data warehouse. If you don't have this kind of architecture or maybe you have it in One department and the other department you don't have it.
We will work in the exact architecture and nature and process that you already have. So we can use a data warehouse. You can just use all of the systems. And the cool thing is that if you don't have those systems connected yet, you don't need to. Forwrd can do this for you.
Connor: Question for you just on being a serial entrepreneur, being somebody who's built stuff in sort of the GTM arena. And when I think about what Forwrd. Is when I think about what you've just described is, it is a B2B SaaS product that is oriented around insights and data and is made possible because of AI versus like, this is an AI product that does AI stuff.
And it makes me think and wonder. If like a lot of people think about and a comparison that we've made with a couple of other folks is sort of like people are looking at AI, like, oh, it has open API's or it runs on the cloud. And as this sort of modifier, and I'm curious if you sort of look at it the same way, where you really think about it as we're building a software product, and that product is possible because of the AI that's inside of it versus like, we're selling and marketing an AI thing.
And, and if that like resonates or aligns to your sort of framework of thinking.
Kobi: That's a very good question. We are actually enabling folks to build their own AI products. So we are a company that sells AI, not a software that AI is embedded in, because I think that every software today, every software that people do today is using AI, and if not today, Maybe tomorrow we see more and more companies
Connor: Certainly.
Kobi: So, so we are a pure AI company that actually our goal is that everyone in the, company, and especially because we are targeting operations like revenue operations, marketing operations, CS operations, we want them to have a new tool that they don't have now. And up until now, they needed a lot of technical depth for them to use it.
They needed a lot of people. They needed complicated infrastructure. Now they don't need it. Now they can use a platform like Forwrd to create their own AI applications.
Connor: yeah, to your point, you're, you're adding that, that AI functionality into and on top of their existing infrastructure. And so you, you guys are essentially helping any of these organizations become AI powered.
Kobi: true, within RevOps will up until now. It's not there. I mean, will people think about ai? They mostly think about, ChatGPT or LLM, right? It's a prompt that you write something and it gives you feedback, Like the new model that OpenAI released recently, Soma, right?
That creates videos. That's amazing, right? But at some points, for some problems, you don't need content. Content is not, the answer,
Connor: I think to your point, like content's the most relatable. It's the thing that a lot of people just immediately understand. And so I think it's the one that it gets everybody both excited. Well, I think I was, I don't even know if it was recent, this is like. Saw it on a tick tock. And so I'm like, Oh, this happened recently.
Could have been a while ago. I have no idea. But Neil deGrasse Tyson with Stephen Colbert talking about AI and being like, this has been around for a while. This isn't new. It isn't novel. Like, the only thing that's novel is all of a sudden they can write essays and do liberal arts stuff. And now everybody's like, holy shit, this is very scary and intense.
And the reality is, that's just a permutation of what was already there versus being sort of like this whole new modicum on top of it.
Question for you that, I'm sure you get from investors and other folks all the time. And if I'm going, and I don't mean it to be super hard hitting, but how do you think about sort of like Forwrd as a solution that's connecting to sort of what you have versus what all the platforms are sort of adding and doing themselves, right?
So you have Salesforce is adding AI, HubSpot's adding AI. How do you think about not, not just Forwrds positioning, but sort of like. What functionality of the platform itself is going to do and then where other solutions that you might be connecting, adding in sort of adopting. In addition to that, where did those fit?
Kobi: Great question. So, to my opinion Salesforce adding AI and HubSpot adding AI it's a must. They will add AI, but to solve your problems, your organizational problems, you need more than a Salesforce or a HubSpot. You need something that will do cross organization, cross application.
Connor: this is something I think in, especially in the HubSpot ecosystem is it's, it's nascent, it's growing, it's getting more enterprising, but I think a lot of people, especially coming from sort of a Salesforce universe know this already is even if you're going to fully utilize and fully deploy any product, any platform, no one software is going to do all of your things.
And so to your point, there's always other stuff in the stack and there always will be.
Kobi: Exactly. And that's only one point to my opinion. Second point, you have business processes that's unique, that are unique for you. When Salesforce design a feature or HubSpot, they don't design it so it can support anyone's business process, right? And by the way, Salesforce Einstein is a very good example.
Salesforce can predict stuff, right? Almost like Forwrd, but in order to use it. You need to change your internal processes so they fit what they have designed, and you don't want to do that. So this is why an external piece can really connect the different voids into a solution that is customized to you.
And if the solution is customized to you, and, even more when talking to AI and data, the accuracy. and the results that you will get will be much, much better that will fit your needs.
Connor: Do you think that there's a place for both platform and tool native AI functionality and then sort of like A, AI wrapper around all of that as well and like both of those coexist or do you think one is just sort of a better approach and strategy than the other?
Kobi: I think that they are already coexisting. And I think that, you know, companies like, Airtable or Miro or ClickUp or even Monday proved us. That even when you have Salesforce, you sometimes don't have all the modules that you need for your own company, like a specific process within cs, a specific process within customer support.
Like you don't have everything. And I think those voids will always be filled by startup companies and exactly the same as and exactly in the same way. AI companies will do the same. And again, a company are different, right? So Forwrd is not a company that's using LLMs or an external models.
Why? Because we want you to build your own models on your own data. You need also to differentiate between the two. I'm, Certain that enterprises, moreover, in the rev ops kind of area, will use more AI for process optimization, let's call it this way. and I want Forwrd to be the simplest way for them to do that.
I think that obviously more companies will try to do those things. But another things thing to mention about ai, I think that if you don't. include AI with automation you're not doing anything and I think that AI Is only the beginning of what we will see I call it hyper automation.
And that's the, the holy grail, right? An AI that creates an action. Hey, Connor, this customer is not responding. If you will send them an email about X, Y, Z, they will increase their likelihood of buying. Boom. You trigger the action. You send this specific email, you get the result. I mean, the entire funnel and process will be boomed.
it reminds me that I've read, I'm not sure where, that Sam Meltzmann said that soon we will have a unicorn of one person or whatever. what enables this is exactly the same. Is data, AI, and automation connected together. That's the key for any business today to optimize.
Their day to day. I think that the companies, you know, like HubSpot and Salesforce doing a lot by developing the modules, but connecting them differently. It's always, you need to look back from your own needs, which are custom and then just execute that.
Connor: I just saw, maybe this is coming from this, but I, I went to. Lord of the Rings in concert, which I highly recommend is amazing. But as you're talking about this, what sort of like reminded me of this is you think about kind of like what those big platforms are and they kind of have their armored plates.
And the reality is, is that if they were fully armored there, there would be completely immobile. And so in those gaps, there's always opportunity.
And if you are an entrepreneur and you are looking at sort of like, what can you be doing and where can you be adding? And I think the reason that I love. B to B entrepreneurship is it's, it's a lot easier to find those.
Like you're going to go ask people like, Hey, where, where are the gaps that you guys are feeling? And they'll just tell you, whereas I think in consumer, it's a lot more like you just make stuff and see if anyone likes it which I think is a lot harder and a little bit more lightning in the bottle, but.
I think to your point even all of the platforms that they add that native functionality get to a place where they need to be able to have something that supplements those gaps and even sort of wraps around the entirety of what they're doing, because ultimately that platform only has eyes on what is in that platform itself.
And even as you get fully deployed, you always have sort of additional pieces.
Kobi: 100 And it's always has been this case, right, back from the 90's PeopleSoft, Sibyl, SAP, you don't have one software to run everything, you, you, you just can't.
Connor: So you talked about the sort of churn prediction and customer scoring components with with Forwrd at the top end is an example. What else are you seeing customers do where you've sort of seen anybody that is using the product and you're, whether you're like, Whoa, that's an incredible, or they're having positive outcomes or whatever sort of resonates to you that you've actually seen some of your customers leverage Forwrd or maybe even Forwrd plus other stuff in their stack.
Kobi: So actually we are playing a platform play. Right. We don't sell an an application. We sell a platform. And when we started, we always talked with customers about what they want to build using the platform, right? Because we ourselves don't know exactly what are the limits, how big we can go.
And on those discovery calls, we learn all the time what customers want, want to do. And you know, it was, I think, like a few quarters ago when we had customers talking about us, about territory management. It's a subject that I never thought that can be connected to AI or predictive model. And we just build it using Forwrd.
So now they have a better territory alignment and management using a predictive model that suggests them how to kind of, divide the leads on a specific regions that all the AEs within this regions get high quality leads. And you don't have one AE that gets all the, all the gold
Connor: Are you doing that like on assignment or are you doing that? Like let load up all the accounts here. Like how does that, what, what's the input and what's the output for that type of a solution?
Kobi: Yeah. So. It varies, but usually we select a specific segment of accounts. It can be a product line, it can be a few regions, whatever. And then with the customer, we actually don't build anything, the customer builds. So the customer knows, usually in this place with a very talented RevOps person that kind of had this idea, and I'm like, yeah, let's do it.
So he defined exactly the goals. He defined what a qualified account is. He defined what a qualified contact is. And we use actually two models, one for the contacts, one for the accounts. At the top of the funnel, we combined both of them and then we pushed the predictions back into Salesforce. We did an aggregated score of all of them.
And then based on that, He implemented the routing function in Salesforce that actually routed the contacts and accounts to AEs based on what the model suggests. So this is like a very good cool example that, you know, I never thought it
Connor: technical does your buyer have to be? So if you have somebody is it the RevOps person that does. Low code and no code platforms and is comfortable in that arena. Is it somebody, maybe you have that plus maybe a data engineer, like what level of, of sophistication does somebody need to be, to be able to work effectively with, with Forwrd specifically?
Kobi: Yeah. Any rev ops, any rev ops that operates within Salesforce, Marketo, HubSpot can build this. And by the way, average build time takes three hours, the average go live is three weeks. So really it's a new tool. Sometimes we do training for an hour explaining the concept. and we also get feedback from customers.
Hey, I don't understand this, this, this, and that. So we constantly change. The product. So folks will understand it, but, and we also have a, we also like a no touch funnel to the product where I see people just go in, connect HubSpot and just starting to build. HubSpot is our main integration, by the way I see people from the store go in, install it in minutes and build models.
and that's really amazing to see and it happens. So. Any rev ops marketing ops, CS ops can use Forwrd we, and, and by the way, the more technical side of the business, like the analysts and the scientists also use Forwrd to accelerate their own work.
Connor: What do you think people are missing? obviously you guys are working with GTM teams that are innovative. They're, they're using some of the AI functionality and sort of new and exciting ways beyond just Forwrd. Where do you think people are either, I mean, everybody's sort of talking about, Hey, we're investing in this, we're doing something here, but what do you think people are either not investing in that you think that they should, or where are they sort of misaligning their energy versus what you think the expected upside actually is?
Kobi: I think that, you know, Prior to OpenAI, to ChatGPT most people thought that AI is a buzzword, it's not working, you don't need it, blah, blah, blah. After people saw OpenAI's ChatGPT management started saying to all employees, Hey, listen, use AI, right? Because we want to be more efficient, more over with the global economy and what's going on.
like globally, right? People want to be more efficient. But I think that I see companies wasting time of building internal solutions that don't scale and don't provide what they want to do just because they want to play with something. I think it's wrong. I think that often, by the way, 90 of the machine learning projects in enterprises fail.
And I think people are wasting. Too much time about things that are not directly impacting their core business. And this is why I think that people should adopt external solutions to support as they do for HR, as they do for a database. No one builds their own database, for their own purposes.
So I think that with AI, especially people are playing with it a lot A lot of companies comes to us after they failed with their internal projects. So I think that people are missing that the idea is not to play with it. The idea is to make an impact on the business and really to try to do it as fast as they can, because this is how you win.
So. Yeah, I see, I see a lot of it. I see tons of people wasting time. I'm not saying that they shouldn't, I'm just saying that as long as it's really focused around their core business, yes, great, right? But if it's orthogonal to the business or a peripheral project, use a different technology, use a different product and just solve it.
Connor: Yeah, what what do you think is what makes you excited either about well, maybe we'll do both sides So so maybe specifically where Forwrd is going next what sort of gets you excited about what that roadmap looks like and then maybe a little more macro than that Sort of how you see some of AI's impact on on GTM software and how you sort of see things changing
Kobi: So with regards to Forwrd, I think what excites me is that everything is new. We're always doing like new stuff that we don't know if they will work. it's being super creative all day long. that's amazing for me. I love to build, I love to create, that's awesome. So with Forwrd, as I told you before, I believe that the key is hyper automation. And we are in the center of helping employees and helping companies essentially to hyper automate their own business processes, right? So if I can achieve the same business goals with a half of the team size, or if I can expand my business outcome by 10x instead of 2x in the same workforce, that's amazing.
and the way to achieve it. Is using A. I. This is this is already agreed. Now the question is how you would implement this AI And that's what's exciting about Forwrd because Forwrd helps teams to implement Ai And to embed it into business processes quickly and to see the results.
Connor: Is speed like? So you talked a little bit, do you think just speed is the most important thing because the faster you start getting those, those returns, the faster they compound? Or is it just that like, if when you quickly seems to come up repeatedly for you and where do you think that the speed is the most important sort of metric that you're looking at?
Kobi: When I'm saying speed is when you're building AI And A. I. Models. You want to be successful, successful in your first model. Maybe in your second model or your third model, right? In order to get to your second or your third model, you need to fail fast in your first model. If you will do an internal project within your company with, you know, like, I try to do, and you will fail after a year, you won't start your second model. So I think because it's super new and everything that we build is new, you need to be able to execute and see results fast. It doesn't mean need to be a day or an hour. It can be a week. It can be three weeks. It could be a month. But you need to control it. You need to own it. And by having something like Forwrd, it lets you own it and build it.
You control it yourself. You don't have like a big team that kind of, you know impacts you, right? You control it. And I think that that's the key. The key is that we enable folks to own it completely end to end. So they can fail fast on the, on the first model and see the results and see the improvement on your second model, on your third model, on their fifth use case.
And I think that that's a very important aspect while adopting new technologies that you don't know exactly what you will get. You need to see something. You need to feel something. It's like an MVP of a product, right? You don't expect it to be perfect, but you expect to feel something, to see something.
And until you don't feel it, until you don't see it, you can't really impact anything. So the The key here is how can I do something on my own without so with without a lot of resources and how can I actually do an impact fast? Because if I can do an impact fast, that will increase the top line by, I don't know, a percent maybe in
Connor: you might learn something along the way.
Kobi: exactly and it's the same for me by the way, as I build this platform, as we, as the team, we're changing, we're adding constantly features to the platform and, and I always say to the team, we need to deliver new features on a weekly basis. That's very, very important in order to see, in order to feel what's working, what's, what's not working.
That's very important for a startup or for any project that's Starting out,
Connor: So normally I, I, I end, but I feel like you just answered my ending question, which is for folks that are wanting to get started. They're, they're eager to jump into some of the eyepieces, what, what, how should they just start getting into things? And it sounds like your advice is just do stuff and experiment and learn something.
And, and that alone starts compounding value.
Kobi: Exactly, but maybe before that try to think of the problems that you want to solve because again, the context, try to think about the problems that are really urgent for the business or to yourself personally. Right. are in a business context. So I always try to think about what's important and what's not important, and if I can impact somehow on what's important, right?
And if I can impact it, I would definitely go and try and experiment with AI. If you're in operations, I would urge you to read more about predictive AI models, obviously about Forwrd, but, to be curious and always try to improve your current
status. Always
Connor: do you think the point of diminishing returns is? So if I'm, if I'm in operations and I'm. Let me go and see if I can use AI to solve this problem, which is like a big generic statement. Right? But if you're thinking, how do I go and jump into this? At what point do you sort of say? And maybe I'm, it's kind of the age old, like engineers will spend, you know, 10 hours trying to automate the task that takes one.
At what point do you think you sort of run into that? And then maybe it makes sense to not necessarily implement AI into that workflow. And and maybe you're spending a lot of time trying to solve a problem that Isn't necessarily it's not that it's not solvable, but but solving it in that way doesn't make a lot of sense
Kobi: It's a, it's a good question. That's how to answer. But my rule of thumb is if you see, if you know that you have enough data to build your predictions On top of it, if you know the business processes in HubSpot or in Mercado in Salesforce and you know that you have enough history to learn from, I would go and definitely check it out.
Usually those are growth companies, startups, I mean, especially like, you know, products like Forwrd is not, is not the way to go. But if you're a growth company if you're scaling, if you don't have, by the way, also, if you don't have resources. If you're looking to solve a problem, but, but you don't have an analyst or you don't have a data scientist, but you have the tools and the data and you want to do it yourself and you want to impact yourself, go and do that.
Connor: yeah, I mean, I think to to your point the the superpower that I think a lot of people in GTM and operations are gonna have is the discernment of when and where can AI be leveraged and add value and then knowing when to apply that as a solution versus When you shouldn't do that. And I think it's kind of the next extrapolation of is this automatable?
Should we automate this? And I think as you get into more levels of seniority and experience and strategy that most of it applies to, should we even invest the resources in trying to automate this? I think that extends to can AI add value to this workflow and the folks that are able to identify where AI is going to add value and where it doesn't make sense starts to become one of the most valuable skill sets in GTM teams.
Kobi: I a 100 agree with everything that you just said, Connor.
Connor: Fabulous. Well, Kobi, thank you so, so much for joining us. I could talk to you all day. I appreciate you sharing your afternoon with us. Hopefully I'm not leaving you to too big of a backlog of emails. But I look Forwrd to catching up with you more soon. Thank you so much for joining us. And for anyone listening, go check out Forwrd.
ai. It is Forwrd with no a, so F O R W R D dot AI. And Kobi and his team would love to show you how it works.
Kobi: Thanks for the opportunity to to speak, Connor. Super excited. Always a pleasure speaking with you.
Connor: Amazing. I'll catch up with you soon.
Kobi: All right. Bye bye.