How Should Entrepreneurs Incorporate AI Into Their Businesses?

uh let me start with my introduction and then we go down the line uh I'm azita arvani I'm the former CEO of rocket and Symphony North America where we disrupted the Telecom Market by creating the very first mobile network that was Cloud native and had full Automation and also uh used open architecture and we that now we're starting to use a lot more of AI in our Automation and we've talked to a lot of startups that that did that as well I'm also on the board of two public companies One is using uh robotics in doing autonomous um industrial cleaning machines and another one that does augmented reality uh that when you com combine AI with ar you get to a lot of interesting use cases on this panel I have these amazing uh CEOs and and uh Chief AI officers for these companies so Shang why don't we start with you you're CEO of dextra you've done this for almost 16 years now so tell us about yourselves Andra we we are the youngest Erp company I I uh make erps erps are the building blocks of uh you know the blueprint for a small company or medum mediumsized companies uh sold to about 100,000 businesses in uh you know Asia India and now in us uh and and I I keep saying Erp is the most uh obvious use case of uh of any intelligence system like AI for uh small or medium-sized businesses because large companies use you know giant systems like sap and article but mediumsized companies which generate a lot of jobs like 80% of all the jobs and 70% of GDP comes from mediumsized and small companies and they don't have systems and they take uh they take uh you know a lot of uh resources money people to get their systems right so uh AI use case is perfect for those companies I've been working with those companies for last 15 years and last few years uh AI has been our main stay again you know we'll qualify AI uh igore has a very good perspective in Ai and I'm a fan of his he'll qualify what AI means and there's a lot of snake oil in this uh you know vocabulary so we'll qualify that but what I'm trying to say is that if we can help small small and medium siiz businesses and you know uh I'm not saying we don't have large large companies they can help themselves but if small companies can use Automotive Systems this can really help our economy uh because they have like you know 80% of jobs so this can help the bottom the pyramid and that's what I do great what you thank you uh so much and Igor I know you've gone back and forth between startups and big companies you with with IBM before and I know your sister's first name Alexa so tell us about that all right so over over two decades ago I was leading the multimodal research team at IBM where we discover the baby version of Watson uh and they didn't want to Greenlight it it was just way ahead of its uh time uh so like Harman from South Park I said screw you guys I'm going home uh and I picked off some of our top uh engineers and scientists stood up our last company uh if you've ever seen HBO's uh comedy Silicon Valley they lampooned a conference called Tech Crut disrupt while went on stage at the very first one I popped a razor flip phone out of my pocket I spoke into it and crickets uh nobody knew what the heck uh they were seeing what they didn't know at the time is we were secretly working with apple on Siri before the iPhone even existed that's how early they were thinking about these natural user interfaces uh to their credit uh and then uh 5 years uh later we were required on the download by Amazon uh to birth what everybody knows as Alexa uh the code name for it was prion and that's why we ended up reing it for this company so yes everybody in the audience can blame me it's my fault you started uh having to pay sales tax on your amazon.com orders because they now had a Nexus in uh Kendall Square great thanks zigor uh next we're going to Stefano Stefano you're the chief chief of AI R&D how's that different from Chief of AI without the R&D Partners well you know it's research and development so researching into it and doing it um so it's probably the same um yeah at Seeker we make AI that um can be trusted it's explainable um and it's transparent uh and we we want to enable others to do so as well and then uh what does Seeker do um yes like I said we we uh have ai driven products uh that across different dimensions we um evaluate content of different modalities think of text images audio uh and we score it qualify it across different dimensions reliability bias um toxicity hate speech profinity and um you know our users customers uh are advertisers um campaign analysts U you know the regular user that really want to discover what is behind content and how reliable and truthful it is so we can call it like the Cara of content with the credit score that you assign to various pieces of content right you can think of it like that okay very good thank you so um you know there's a lot of H AI developments in the recent I mean the every day you get up there's something new going on right um how has the these advancements uh affected the landscape for the Enterprise you know from the perspective of your customers how are they looking at these AI advancements so I'll take the liberty of defining U not just Enterprise but also small business needs small and medium siiz and I said 80% of our uh economy is around small and medium size and that's what that's that's where we play uh and even in Enterprise because this is basically a superet of all the needs AI has limitations it can can't do math we all know that it can't do math so you know you you write write an ELO where it uh it tries to create a program to do math whenever you ask it to do you know a complex calculation it'll basically you know turn out a program to do math in real time that's what it does um if if if I told you um that uh there was a chance that the answer that I give you in a business setting is going to be wrong I'll be fired so if there's a 1% chance that my boss I don't have a boss but let's say I had a boss um um you know if I ask a question and somebody says look I say what's my sales and they say it's $100 billion and I know it's not $100 billion because if you know it was then I would be alone sitting here alone with you guys but it would be wrong so they would be fired and that's the that's what AI does there's a chance that it'll be wrong the answer will be wrong so you have to check every output and there are ways to do that you know Arin was here in the morning uh his system works on rag which is retrieval augumented generation it checks for answers checks if your answer is correct or not the point being AI it there's a chance it it is going to be wrong in the answer it gives you which means it doesn't work in a business setting so then you need to quantify what it what it will do and what it can do now what it can do is have a conversation with you which is what we see in chat in the case of a business user Enterprise 70% of how many of you know that 70% of small businesses and you medium business don't use software except QuickBooks that's a big Challenge and you've heard of sap and Oracle and Erp companies but they haven't been able to penetrate even 5% of that market it's because small businesses don't trust software and look what are we talking about right if AI cannot be trusted they're not going to trust you know whatever you sell them so they don't trust QuickBooks they don't even trust the data that comes out of QuickBooks because they feel that that the numbers that they see in those reports doesn't represent their business and anyone who runs a small business knows that they have some you know they have a view of the business that is not seen in the numbers Andi can't solve that but it can have a conversation it can do onboarding it can do a lot of other things and I I have a t you know I spend a lot of time writing about what it it can do but these are the limitations and I you know I leave it for the rest but I think we have to you know we have to have this uh a resar of expectations on AI as far as it concerns about Enterprise and small businesses it can't do everything it can do a very small amount of things uh actually it's interesting I did a search on one of the companies I'm on the board of and it said that the revenues was one not a search engine it said the revenues are 1.6 billion and I knew that's a little higher it was actually 1.2 billion and I checked so so you're right it's 1.2 billion yeah but it it was saying it's 1.6 billion which is you know it was a nice thing and I said oh I don't even know the revenues of the company I'm the board of that's strange that's why I checked anyways the point being that you obviously have to check these numbers as they come out uh igore what do you think uh how should the Enterprises think about you know they they hear all these things they get excited and they come to you and say hey give me the latest gen I want to you know take that and brush it all over my Enterprise yeah it's actually worse than that um pre pandemic I remember giving a talk at the chief digital officers uh forum and they said hey what's one thing that you're going to now uh tell us uh it's very important that that we should know I'm like here's what's going to happen you're going to waste my time for one year because it's going to be build versus buy you're all euphoric the picks and shovel crowd are knocking on your doorsteps and your it teams are going to try to figure this out and then I have to channel my inner Japanese uh soul to figure out how you don't lose face you know I can't let them use face so I have to do all this theater for for the better better part of a year before you realize there's no way you can build what we've been building for 20 30 40 years so that's the first thing uh that we typically encounter with these Enterprises is the stages of grief until they finally realize it the other thing is I mean uh I I Adore that he's starting on smbs and bringing capabilities that would otherwise be unyielded to them like the saps and the rest of them that that they would never hope to touch I did uh What uh my coo calls SAS backwards when he en countered me he's one of the original Oracle exacts that benof brought in to help him run uh go to markets from year 2 through IPO and he kind of uh was was laughing his head off he's like you're going after the biggest baddest organizations on planet Earth the largest government agencies and things like that why are you doing that instead of going SMB and I'm like here's why if you want your kid to be an Olympian you hang out with Olympians and you get trained by by them and if we go and we can service and install things in nuclear reactor sites and and government gencies and things of that sort then I know everything is down market and the team uh you know gets that experience in a lot more aggressive way so over a half decade ago we said hey there's an intersection of AI and Knowledge Management that doesn't exist look we're all going to these conferences and all the big Tech folks are showing up hey look at all this stuff that I now have for Enterprise and Ai and stuff like that 2017 is when we started who was talking about Enterprise AI then no nobody who was talking about having conversational interfaces on on these systems nobody you know why because I because every time I said hey this is what it's going to be like I was saying no the interfaces are going to be more Tableau like more looker like all selectors and things of that sort that's how Enterprises wanted to use that that stuff and I'm like that's peculiar because all of us know how to talk to each other before we learn how to read and write as well so I'd rather focus on putting the human at the Center of these experiences and then everything will work itself itself out in the wash and then what happens with chat GPT 24 months ago that gets revealed 100 million people start using that and then everybody starts paring the same thing and I get to roll my eyes uh Stefanos uh can you talk about how Enterprises are seeing the advancement of AI but in particular maybe uh touch upon the fact that since you come from the sort of the R&D side of the this the technology has democratized across you know whether you're large or small you can use AI um you know but for your customer service whether you're a $1 million company or a$1 billion doll company how do you how do you respond to that and and what we heard from shash Shang and Igor yeah so I guess I think democratization is one one way of of seeing it you know there are there is open source there's open source AI but there's also I think the vast majority of of you know companies companies are not using open source AI they're using closed Source AI um and you know the big challenge with with um really um open- Source AI is Computing costs uh it's in order to make AI in order to build Ai and maintain AI um you need computers you need power and um you know it's there's going to be and that's expensive right there's going to be a correction I think where where we with where things are going and there's going to be some um kind of change in the stat status of the of the hardware providers with more competitive AI accelerators and and chips that are coming out uh but yeah generally I think democratization is it's a interesting uh way of putting it because there's only still a very few uh that that can operate open source due to the prohi prohibitive costs great thank you um eigor which uh specific strategies since we have a lot of people here that really want to put these take the ideas and put it into action right so what specific strategies what Lessons Learned uh do you have from your work at at pron that has worked for you in the Enterprise to grow your revenues um follow the pain you only go to the doctor when something hurts right oo I have a AE right and then you go get get it handled and I'll give you a specific example of a piece of pain all right so here's a big Energy company and for years they've been pouring millions of dollars uh to construct an AI for their outage and maintenance services now why do they care about that because they predicted if they had such a thing they could reduce the downtime of nuclear power plants by half if they had such a thing meaning they don't have to spin up fossil fuel burning plants it makes the grid more resilient especially in deep summer and uh deep winter now there's a dark side uh to what I just said if you ever watched any documentaries on Three Mile Island when Congress did an investigation we were only within 30 minutes of a Chernobyl like event that would have irradiated the Eastern Seaboard and and here's what they found six major issues and I'm oversimplifying one was a design flaw one was a faulty valve four out of six issues isues were Knowledge Management issues where engineers and technicians were not rapidly getting the the pieces of information uh that they had so for years they've been trying to uh you know figure that out we had them in production for the spring refit cycle of the reactor plants in less than five business days so those are some of the outcomes so Frankly Speaking you're going to have to build the the Rapport and relationships and and trusted uh to become a trusted advisor to these organiz ations and essentially follow that pain I know it seems very simple uh but it's amazing how how many of us uh think that we show up and we know their use cases in terms of hey this must be your sales enablement pain hey this must be your your technical pain and things of that sort and then we're not actually sitting there keeping our our our mouth shut and actually listening uh to them and following that pain ourselves how do you get to the point where they trust you like uh from the moment you you go to an prise uh how do you make them trust you enough to give you enough people for you to see how that paino then uh goes down the line and creates this um massive cost that you could potentially reduce right because I tell them I understand you're a serious environment that's highly regulated and there's certain controls that are necessary so my team may not even be ready with all the things that you need and so you're not allowed to become a client of ours and so we start this journey years ago with them to start building a rapport in relationship and I start saying though hey start I don't want you to be a client but let's have our R&D team starting to work together think of us as a trusted adviser on the outside start bringing us into into the environment where we can start working together that in Insurgent style um uh relationship building actually works because then we become deeply embedded in the organization look whenever you're servicing these entities you're not Johnny on the spot you're not always there so somehow when you leave and your competitors show up they have to become Teflon coded and they know that you've put in that sacrifice you've worked with them you solved their problem and your product management teams keep absorbing these requirements because at a certain point in time uh the the two entities are going to meet and say yes you now have let's say in DOD the impact level five and six that you need for classified uh information you now have some of the Hippa stuff or PCI stuff that you actually need to operate in their environments but you don't just show up on the first day that you have those certifications I probably was there years ago yeah I was probably there years ago that makes a lot of sense uh Stefanos how do you see um you know what kind of lessons have you learned at Seeker What Works what doesn't work uh working with the Enterprises um what works and doesn't work yeah I think that um one big lesson learned is doing the cost benefit analysis of you know what model you're going to use for which application really understanding your application uh and doing the cost benefit analysis of where where you're going to apply this model what Hardware you're going to use um can go a long way uh large language models are not a Panacea right they're they're not going to solve every problem that you may have on your business or your use case and they're also very expensive to operate uh and you know even even like I said earlier like when you even if you don't operate uh your own kind of large language models and use um a closed Source One you have the issue of trust and control you can control it you can trust it you can change it you can understand it so um I think um really Lessons Learned is understand the problem do the cost benefit analysis choose the right solution for the problem and um focus on your Roi okay uh sa as while we're at uh you've got the microphone we could also talk about what is the biggest challenge uh that you see with companies deploying gen it's kind of flows from what you were just talking about if they decide to do gen what what are the problems that they would yeah challenge again it goes back to uh cost it's it's very expensive to build it make it build it deploy it um that's why you have to be smart and always do the cost benefit analysis uh of what is the right solution for your problem and again the other challenge is trust and control uh you know yes close source is a little bit less expensive but uh you can trust it and you can control it okay thank you thank you um shash any anything in terms of the biggest challenge that you see with them um deploying GNA Solutions yeah so the one thing that we have seen work is uh AI agents uh and I want to speak about agents let's say you are a small company and you know you had few employees and you you know you want to model a business where you if you had more employees what would you do with them and you know every small and medium SI business has this this this hard uh bottleneck where they can't hire more people so they don't think beyond that but with agents let's say you can deploy agents and we have agents which run on you know a very tight guard trail of Erp so you have agent that can do accounting you have an agent that can do purchase uh agent that can buy manufacture and so on so you can then simulate your business so let's say you simulate your business on you know J and you say look I have 50 employees and how how does it look like how would it look like if I grow at this space for the next 3 years and if I you know invest This Much from my uh business from you know take taking a loan or raing a financing AI agents can do a very good job at simulating your business and I think that's what we have seen a first of all it's very very less harmful than you know operating your business so it's just simulating it but B it gives you ideas that would not necessarily come from your operations and there's something more that I I'm I'm writing a paper on uh generative uh agents which are uh sort of you know there something called as Gan adversary agents um adversary Network sorry so we writing up I'm writing a paper on adversary agents so let's say you have two agents you have uh you know two employees and their goal is to you know achieve some they have individual goals but they adversarial goals and you're just simulating that those goals you have a manager you're an employee you have a manager and two employees and five employees and you're then seeing process improvements you're asking questions that you would not necessarily see answers to in your regular operations for example if you have three or four offices and you buy stuff from you know overseas you sell in us and you have a you know you want to see what's your labor cost going to look like for the next 12 months you can actually run a very good simulation on agents and that's what we have seen small even mediumsized companies with 50 employees or 100 employees use and then get more ideas on process improvements the second thing is that the process boundaries are going to dissolve so an accountant can also do purchase a marketing person why kind of marketing person do uh you know better operations So In traditional you know companies medium siiz small Enterprise doesn't matter traditional companies like manufacturing industrial companies they have been stuck uh you know with the business process uh for like tens of years years even hundreds of years in some cases like Industrial Age right they've been doing the same thing over and over again for hundreds of years and the reason they can't change that is because the risk of changing that in real world is very high me you can't raise Capital as easy as you know a West Coast entrepreneurs can I we have raised a lot of capital so I know you know how a small company struggles with raising Capital so if you can't raise capitals you don't have use cases you you can't expl you can't take a risk if you can't take a risk then you don't have success that can then Define the next you know generation of companies or next generation of business business processes so gen can help you simulate a lot of use cases in you know environments which were hither to untouched and that's the most exciting thing so you can you will probably have I mean people will talk about uh you know uh jobs going away and you know people getting uh lesser jobs I don't I don't believe in that I think AI is a good tool for humans to use just like a calculator is as I said it has limitations but it can be useful and I see agents I see uh you know very helpful uh co-pilots that can enhance a small business and then again I keep saying that Enterprise large companies can use AI or not use AI they are still going to be you grow at the same rate but if small companies can grow even 5% faster than they are growing at our economy will grow 5% faster it's a very straight mathal calculation thanks shank um Igor um if you look at how rapidly the the AI uh landscape is moving as a as a startup company um very successful startup company but still a startup company um how do you make sure that you have an enduring business value that your business model and your competitive Advantage you know stays relevant and that that you're not getting um competed out by by the new startups that are coming uh only two out of uh um there's 10,000 Reasons a startup could fail only two out of 10,000 are probably some competitive threat the rest are self-inflicted wounds you're not taking care of your clients you have the wrong pricing the wrong packaging your your your coat is buggy and things of that sort there's a lot under your your control you really do have to find equilibrium between R&D and and uh go to markets this generative stuff I mean we were using a lot of these things didn't know the words right we were doing Cloud before Cloud existed using neural networks before people you know knew these things existed um in some ways um I wouldn't oversell it because I can tell you in in the environments that that we find ourselves in you have compliance and legal officers that when Genai shows up in their environments they shank them they literally shank them they don't want any part of them because do you want hallucinations in a nuclear power plant in a hospital in an air base and things of that sort so the things that actually operate the world as we know it um they don't want any part of that they want determinism in some ways and and some of you are like well you know let's figure out how to guard real all these these things um and uh I mean they're developed with stolen content let's call a spade a spade right now the the regulatory environment and and the legal environment has to figure that out between the Publishers and in the AI industry uh but trying to gu guardrail these things with no matter whatever fancy agents you want to use to try try to do that is like trying to keep Donald Trump on a teleprompter you may be able to get away with it for about 30 minutes or so while he delivers his speech in the Oval Office but later that night he's on social media and letting you know what he really thinks about what he previously said and it could be even contradictory okay Igor let me push back on that a little bit right so as a as a consumer right even as a consumer um I I see chat GPT before is doing great so I'm paying for that and then I see anthropic it's got claw three that's awesome I'm getting that and I've got perplexity and that's great I'm getting that but at some point like these models are like I'm not going to pay for all these models right but as as a startup that I'm putting my future on one or two how do I know because it seems like it's a race right uh Chachi PT is good and then uh CL three comes in and then GPT 5 is going to come in and then you know what I mean like it's it's moving way too fast yeah as a child of the 80s I remember watching war games and what's the Lessons Learned there what's the last scene the only way to win is to not play okay so there's two ways to make a hamburger you can take a cow and turn it into a hamburger or you can be inspired by a cow that could make a hamburger but you go over there and you get some potatoes and some carrots and some turns and turn it into a hamburger right so so an Alm could just model the language it doesn't have to actually produce the the answer and you're not going to be able to use any answers that an llm emits in any sort of uh highly regulated industry and and serious uh construct so that's how you can still uh remain relevant and look there's two ways for startups to be born let's not think that there's only one way because when I started the last company everybody was looking at this famous uh company uh in uh UK called spinvox that had I remember G Kawasaki being at that that crunch he's like eigor I use their product it's Fantastical you have 1 and a half million in funding they have almost 300 million in funding from Goldman Sachs they cratered BBC did an investigation and found that they were a Mechanical Turk they were lying about doing speech recognition when this company started everybody was saying whoa why are we going to get this thing funded uh luckily we count you know Steve case and his team is some of our biggest uh earlier uh backers as well and I appreciate um him working with us but everybody was looking at LMN aai right the fame team up in Montreal right 150 million in funding they were going to Corner the market on Enterprise AI they got cratered then Reed Hoffman shows up with his inflection thing with a billion plus and everybody's like oh my gosh that's going to you know take the wind out of your sales there's two ways to make a company ladies and gentlemen you can grow it in the rainforest of the West Coast where there's plenty of resources food and water to attend to the creation of these organisms or you can grow in a desert and and economics as a study of of the all allocation of scarce resources I always bet on the desert creatures the ones that are rise of the rest that grow up in other markets know how to be more resilient creatures and adapt to these market conditions and I don't need you know billions of dollars of Jensen's parts in order to operate okay thanks igore okay so since we only have a very few minutes left uh very succinctly Stefanos what are you looking forward to see uh for Gen for AI in general not just geni in the Enterprise in now and going into 2025 yeah what what I think uh is going to be amazing to see in 2025 is basically businesses and Enterprises being enabled to build their own AI Solutions uh build their own models that on their own data that they can trust they can control they can explain they can change and um you know I think that there is an out of competitive cost uh I think there there are going to be some amazing opportunities to do that um and looking forward to that great thank you shashan what's your same discussion build versus buy for the next 12 months build veres bu half or close to 70% of teams will build and then fail and then they'll go and buy and this is you know till Tail as all as time Cloud you know everybody wanted their own cloud and then eventually settled around three or four Brands so build verus buy is the biggest biggest conversation right now in Enterprise and you know your recommendation of uh you know you'll go around and then learn your lessons will probably hold true so that's one thing but I personally believe for small businesses and even for Enterprise agents that can be guard railed uh there's a lot of reporting Solutions I've not seen any promise because as I said you know if it's 1% faulty it's 100% faulty but we'll see more progress in reporting great thank you Igor what do you see coming up in rest of this year and in 2025 all right here it is I'll unpack it there's regulation coming isn't there and so let me share something with you thanks to EU first and then us a a a good uh country uh lawyer knows the law a great country lawyer knows the judge all right so some of us are going to be doing what waiting for these things to show up the EU aai act what's happening in California the executive order and and some of us are are going to be part of the negotiations and the constructs and then figuring out how how these things mutate again to have proper you know responsible AI being birthed ensuring uh as as uh made in his previous comments that we can democratize access to AI so that there's opportunities for companies of all sizes small medium and large um and so that's the thing that that over the course of the next 12 months uh we should be paying attention to right and also what I've seen from Enterprises is that they have been very ambitious in the last 18 months you know they they start up uh we were having a conversation with a good friend of mine that's here uh with like 500 AI projects and that they have to filter it down to a reasonable number that they can actually put the the Investments that it deserves in order for it to go from you know proof of concept to the actual deployment thank you very much for your attention and your time really enjoyed it thank you to my amazing panel and have a wonderful rest of the [Applause] day for

Channel: Forbes

Views: 6,937

Keywords: Forbes, Forbes Media, Forbes Magazine, Forbes Digital, Business, Finance, Entrepreneurship, Technology, Investing, Personal Finance, Small Business, Small Businesses and AI, how can AI help my business, How can I use AI in my business, LLM, language Learning models, Adobe, Apple, Amazon AI, big tech AI, how to make money in AI

Id: ehgkfmsdOzg

Length: 34min 39sec (2079 seconds)

Published: Thu May 02 2024