EP15: AI Strategies in Legal and Marketing with Hacktech CEO Hakob Sharabkhanyan

Discover the future through the eyes of a tech visionary as HackTech's CEO, Hakab Sharabkhanyan, joins us to unveil how AI is catapulting industries into new heights of efficiency and innovation. At the young age of 19, Hakob embarked on a journey that has led to commanding an 80-person team at the forefront of digital transformation.

Our discussion orbits around the strategic application of AI in sectors such as legal, marketing, and advertising, and how these integrations are not just about ubiquity, but about enhancing business processes with precision. From reshaping the traditional roles of sales teams to the necessity of human oversight in AI-assisted legal documents, Hakob provides an invaluable perspective on the delicate dance between AI-driven advancements and the irreplaceable human touch.

We also tackle the burgeoning challenges and opportunities presented by AI. The episode peels back the layers on the intricate relationship between human judgment and AI outcomes in the legal field, especially when the stakes are as high as immigration. As marketing strategies become increasingly personalized, we discuss the importance of maintaining authenticity in a world proliferated with deepfakes and AI-generated content.

With Hakob's insights, learn how companies are navigating the complexities of adopting an AI-first strategy despite obstacles like a lack of expertise and understanding of AI.

Join us for a conversation that promises to enlighten, inspire, and provoke thought on integrating AI into the fabric of industry operations.


About Hakob:
https://www.linkedin.com/in/hakobsharabkhanyan/

About HackTech LLC
https://gohacktech.com/

Chapters

• 0:44 Impact of AI in Industries
• 16:13 AI Applications in Legal and Marketing

Transcript

Speaker 1: 0:44

Thank you very much. Welcome back, fellow travelers. In today's episode, we're going to dive deep into the topic of how AI is going to impact industries such as the legal, marketing and the advertisement industry, and joining us on this topic, we have an amazing guest, hakab, who is the visionary CEO and founder of HackTech. Hacktech is not just riding the waves of digital transformation, they're creating it. They offer tailored AI solutions and custom software engineer to help skyrocket productivity. Hakab's expertise is pushing the boundaries, for the legal and marketing sectors have made him a sought after voice in this digital transformation. So, hacab, welcome to the show. We're excited to have you. I think you're also joining us from Armenia, is that correct?

Speaker 2: 1:31

Yeah, that's correct. Thanks, steve. Thanks for being here.

Speaker 1: 1:34

Absolutely Welcome to the show. So tell us a little bit about what HackTech does and how you became CEO of this amazing company.

Speaker 2: 1:42

Yeah, actually I founded the company when I was 19 years old so I was studying in the second year in my university and I decided that I needed to start a software engineering firm.

Speaker 2: 1:54

So it makes us already nine years and during nine years we reached from a single man company to a team of 80 professionals. And what Haktan does and the way we position ourselves, that we are not a traditional outsource or outsource agency. What we do is engineering partnerships. So with this big team like 80 people, we have right now just nine clients, so all initial clients. We have a dedicated team, full cycle engineering, from idea like, from product management, business analyst to actual development, design, develops all the stuff together. And what makes the most sense in this model for me is that our team is kind of becoming part of our client's team and like just being motivated by the impact they make on our client's project, which is a very big driver in IT. By our logic, we are industry agnostic, but it just happened the way that we accumulated so much knowledge in the legal and marketing and other types of industry that right now it's mainly main focus in terms of digital transformation.

Speaker 1: 3:06

That's incredible. I know that AI has definitely been riding a wave and we've started to see its transformation starting to occur and even some disruptions that we're starting to see kind of in different industries and industries in different segments. So walk us through a little bit kind of what you're seeing from, let's just say, advertisement and marketing. How is AI really kind of playing a role in transforming these very kind of? You can think of them as traditional, very legacy type of industries.

Speaker 2: 3:37

Yeah, I think it's just becoming like you either adopt AI, use AI or you will be out of the market very soon. And it doesn't mean that you need to use AI in every part of your business. It's more about using it the right way to boost your productivity. For me, if I talk to a company now and I'm sure it will be everyone in one, two years when you talk to a company and ask, how do I see you leverage AI, if the answer is we don't use AI, then you look at these companies that are at least 20-30% less productive than their competitors and like leverage AI doesn't mean that everyone is going to build their own AI solution.

Speaker 2: 4:17

It can be like just learning how to properly prompt a talk to GPT, because it's not so easy to learn this. So for me, it's kind of like using the internet, so if you are not using an internet, you can't be in an industry. Now, the same is true for marketing and other item industries. And if you are a product like you have a marketing or advertising product then you need to adopt AI as a solution in every part of your business, because from creating to measuring the success metrics, because AI is not just like GDP or just conversational. So check how your campaigns are. I mean, automatically adjust the campaigns. All these different kinds of AI. I believe everyone reminds me of Matthew.

Speaker 1: 5:05

Yeah, I would think that with marketing, you know, there's a real push towards, you know, a really big push in marketing towards more personalization. There's more things around data analytics. There's more things around trying to build more creative and interactive customer engagements, and I really think that you know, and I think I see the data is starting to move this way. I really think that you know and I think I see the data is starting to move this way too. Is that you know? Is that there's marketing companies are starting to they need to embrace kind of an AI first strategy. Is that right?

Speaker 2: 5:36

Yeah, yeah, that's right, but again like they don't need to fully rely on AI because it will kill the personalization. If you like, just use GVD to write a copy and send it to people. They will not read because even from the first line, you can understand that this is AI generated. So it's all about using AI in the right way, like still putting your fork on creating your campaigns, creating your copies, but using AI whenever it's best, whenever it can hit the most value.

Speaker 1: 6:08

So definitely, dr Andy Roark. So definitely okay, so we need to. I think that there's kind of a crawl, walk, run type of perspective, right Is that? I think it starts with educating the right type of AI and being very prescriptive about how it's being used with things that are customer facing. I think that's where you're trying to go right, okay, and what have you been seeing, I know, in terms of kind of you know, when you're getting involved with certain customers in these industries, are you seeing? Is it that there's a lack of adoption due to lack of understanding and how AI works, or understanding and expertise with an AI? Where do you start to kind of see more of a skill or a technical deficiency with them, kind of moving forward with an AI? Where do you start to kind of see more of a skill or a technical deficiency with them, kind of moving forward with an AI approach?

Speaker 2: 6:49

Yeah, actually, for me, some of these industry companies having like CTO or CIO, even if they don't have a product, even if they don't have like internal software, they still have this role to keep them up to date with technology. And in this case it is much easier because these people understand the value. These people know that at some point they need to build something for them. And the other scenario, where the founder or CEO is not technical, you need to do a lot of job to educate them. Sometimes calculate like return on investment. So, hey, you know like you are losing 20% of productivity here and we build these custom software or implement this AI solution for you that like in a month or in six months or in a year, and we'll cover your costs. So the way we typically work when starting a new relationship is opening a deep logistic product manager. So a product manager spends a month or two with their company, sees all their day-to-day processes and then comes up with a product solution that they'll solve their needs.

Speaker 1: 7:57

That's interesting. I mean, talk to me a little bit about what you're seeing across the C-suite. I know you mentioned CIO and CTO. Those tend to kind of be more of the technology and information leaders inside of an organization. Do you see organizations that don't have CTOs or CIOs that have the right type of expertise? I mean, is that typically where HackTech comes in? Talk to me a little bit about how you kind of bridge these two roles together with this new kind of disruptive form of technology.

Speaker 2: 8:28

Yeah, that's exactly what Haktek does. If a company is not having a CTO, they need to go to digital transformation. It's very hard because you can't just hire, like even senior developers, and expect that something will be built. You need the domain knowledge, you need the tools, expertise, all these that CTO usually brings to the table, and with Haktek, that's what we bring to the table. So we are not just outstaffing, because for me, outstaffing is not an IT business, it's more an HR business. So we are not just deploying, like saying that, a local, we are deploying the team who knows the processes well, everyone in Hacktech aligned with the processes that we are working and they are coming with these processes implementing it to the Hmm, what do you, what are you seeing and kind of going to dovetailing back here is, you know I want to work through, you know, kind of the the the role of AI in startups.

Speaker 1: 9:18

I know we talked about CTOs and CIOs. You know we we see how most established companies kind of in in in the space have kind of got a very mature C-suite of your board of directors. Startups, you know they're trying to get a product out to the market pretty quickly, right, so they're trying to scale for growth and efficiency. What do you see as some potential pitfalls when it comes to startups? When they start to leverage things like AI, where do you start to see some of them not really hitting their mark?

Speaker 2: 9:49

Yeah, actually, again here there are two types of AI. One is just actually a couple of levels or just two like. First level is just leveraging the GPT APIs or any other model APIs to bring some AI components into your product, Like it doesn't matter if you are an AI startup or any other industry startup. I mean, if you have a text box where someone should type something, it's a very low-hanging fruit to integrate GPT and give the user possibility to generate it automatically. The next level is prompt engineering, to deeper understand how all it works and come up with prompts that will give the most value. And the other level is training the models itself, getting the most value. If the R&B data and the other layer would be like going and creating your own models and social models.

Speaker 1: 10:42

That's interesting. You know, you mentioned data modeling and I really wanted to kind of double click on this one here, and this is just maybe just kind of helping our audience just understand a little bit. More is, if you've got an idea with a startup, right, they're leveraging some sort of language model, they're leveraging some sort of AI. Where do they get the data to start doing the data modeling? Where does that typically tend to come from, right? Where do they leverage that same type of you know the amount of data, or do they? Or do they not? They don't use the data.

Speaker 2: 11:17

Most AI startups don't, and I was making this joke. I was recently in a big conference like Web Summit in Qatar and, like I would say, 80% of startups were having an AI next to their own name. So it's like something AI, something AI, and you see that everyone is just leveraging the apis or calling it an ai product, which actually is a. It's just a software using like basic integrations. So most of the companies because, like for very, very easy, very basic model training, you will need at least like, let's say, if it's conversational ATA and we will have thousands of conversations and we know that we're already adding more For something like more to get a better result, you will need like hundreds of thousands of them. So most companies just either new ATAs or, if they are like, getting stronger skills, they are doing some problem.

Speaker 1: 12:09

Hmm, so what do you think that the three key, what would you say are three key? Three key, uh, or three key assets that a startup needs to be successful when they're looking to leverage things like AI and not just basically saying, hey, we're X company and that we've got, you know, ai next to it. What do you start to see? And when you evaluate startups in terms of kind of three key components that they need to have in order for them to say there's a potential for you to have great success or you're just never going to go anywhere?

Speaker 2: 12:40

Yeah, I fully. First of all, again, if it's a technological startup and they want to adopt AI, they need to have a CTO. It's a must. They need someone who understands. First of all, they need AI or more, because sometimes you can just go train models and do all this complex stuff and meet us at dollar, while you can just use some open source model.

Speaker 2: 13:01

And we saw a lot of things after this AI hype A lot of startups who were doing their internal engineering, like doing some complex researches on their own business model and all investments that they were getting was related to this uh core, uh ai research and then something like from this giant tech company something open source coming and killing their business. So it happened a lot. Like many companies before gpt became available to everyone, many companies were were doing research in a lower scale and getting investment in this research. Then, with GPT, I guess 90% of them just get out of the market. So the first is having proper texting there. Second, understanding the vision, making sure like you need that AI solution, you need that to go and actually research, do the research in that field. And the third one was check that nothing exists in that space, nothing these tech giants are doing in that space.

Speaker 1: 14:04

It definitely seems okay. So you definitely got to have the right leadership in place to kind of help steer and kind, of course, correct kind of where the vision is actually going to be and then setting the journey for how we're going to achieve that as an organization. Sounds like those are some key differentiators every startup needs to have. Is that key leadership here? I wanted to kind of just dovetail back into the topic around AI's influence around the legal industry. I know that there's a lot of potential for AI to disrupt traditional type of legal practices. What are you seeing from this particular angle when it comes to leveraging AI with regards to legal industry, legal type of contracts or even if, in fact, if it's just kind of mixing these two together, is it a good combination or is it you start to see that there might be some other way that there might be?

Speaker 2: 14:58

some some other other way that this might go.

Speaker 2: 15:01

Yeah, actually, a little tech, little industries being affected with AI.

Speaker 2: 15:03

And again, it's not going to change people, it's not going to replace human, but it's going to make them more productive, starting from like just research that any other rebuild should do, or they really like just research that any R&D build should do on their daily job, from to creating contracts, creating some documents, and we have a very successful case study right now working with a big integration company who is dealing with O1, h1b visas. If you know, it's very complex, like five, six hundred pages of PDF per weekend and they are handling it in a scale. And we started working with them about two years ago. We came up with the custom software which accumulated all the data and it was the AI just streamlining the processes, the data from practitioners, et cetera, et cetera, and with this AI we started to automate a lot of. I think like 50, 60% of the petition can be automated by AI and then just immigration specialist or a common agent will approve and send to the next teenager. So and this is just one example from immigration law, but it's a completely different way.

Speaker 1: 16:13

No right, you would think that at least I know. I probably would think that you know planning low price. You would think that at least I know. I probably would think that you know leveraging things to be able to look at things within legal such as looking at things such as you know contracts and then looking at things such as you know keywords to be able to look at and assess certain levels of risks within contracts. I know you mentioned H-1Bs and visas and things like that. There's a lot of documents that has to source through, so I think AI is great application there. Where do you see as a potential pitfall when it comes to AI? Where do you see that there needs to be some sort of boundary or guardrail where you might need to have some sort of kind of human interaction to kind of say, hey, I need to make a call on the field. This was a false positive on AI's part, or where do you kind of see this?

Speaker 2: 17:00

Yeah, again, like, for example, O1 or H1B is super important, like for some people, it's like life's important thing. So what we are doing every part of AI generation should be checked by a foreign or immigration specialist. So AI is doing. Every part of AI generated should be checked by a foreign or immigration specialist. So AI is doing a lot of mistakes. Again, it's helping you to boost productivity, but you can just ask AI to generate 600 pages of petition for you and you hope that you will be approved. So whatever AI doing in any industry where there is importance to be always correct, it should always pass through the human.

Speaker 1: 17:39

Okay, so that's where I was getting at. So, yeah, so it seems like you're offloading the undifferentiated heavy lifting to the AI model to do kind of more of the work or the analysis, maybe more of the consensus mechanism, where it kind of then gives you some sort of predictive outcome, where it kind of then gives you some sort of predictive outcome and then kind of that last check or verification kind of goes through. An attorney or somebody that then looks at everything to kind of says, yes, this basically does look good, does look accurate or no, this was completely inaccurate. We need to run it through an AI model again. So there definitely needs to kind of be human element, or the human in the room, so to speak, to make that last verification on whether or not this is good or good, good decision or bad decision.

Speaker 1: 18:19

Right, okay, now also, I want to kind of touch on the part of I go back to this again in marketing is you know, I am very, very curious to kind of see where we start to see future trends with AI and market-driven strategies. What are you seeing from your perspective in this space? Have you had a chance to work with a customer or a client where they've been able to say hey, we've taken the traditional model of marketing and advertisement and we've been able to completely game-change it using AI strategies. Have you seen that yet?

Speaker 2: 18:55

No, not yet. Actually, I think marketing will be changed totally because the way we look at the content right now, it will be changed totally in a couple of years. Because we see the content, a piece of content like just posting LinkedIn right, we used to read all of it. Now you lose trust towards LinkedIn posts because I can go and post about, like astrophysics, that I don't have any knowledge about it, and the same is starting to happen to creatives like images, and same thing starting to happen to videos.

Speaker 2: 19:28

So all these things which make a huge difference in marketing. For example, if you want to bring bigger attention, you should create an e-joke and it takes a lot of effort and it resonates with the higher views and nowadays it's becoming just accessible to everyone. So I believe there will be a huge shift how we look at the content, how it attracts us, in terms of our clients using AI to make their sales and marketing strategies better. Right now, we are working on an AI product which already, in eight months, we created a POC which is already able to replace low-level sales teams. So, basically, as soon as the lead is coming to the CRM, the bot is able to start the conversation itself. Go to the booking. Actually it's like sporting industry class booking type of CRM Actually push to the booking and when it sees that there is a positive answer that hey, yeah, I want to book this class tomorrow at 8 pm it will go back to the CRM border class. So it completely replaced these low-quality salespeople that were not able to close the leads.

Speaker 1: 20:40

Hmm, Interesting. I know you mentioned around content creation not being authentic or being inauthentic and I think that's kind of a concern is. You know, you don't know whether or not whoever generated the content actually has the credentials to be actually doing that from maybe some cryptographic keys to maybe NFTs or some sort of way to kind of encrypt or digitally sign something as an authentic creator. Are you seeing that as well? Do you start to hear this from people saying, as we start moving into marketing and advertisement, we need to find a way to protect our brand, to make sure that it's authentic and we've digitally signed that that it's authentic and we've digitally signed that.

Speaker 2: 21:26

Yeah, it's becoming very popular, like digital signage, because I can just take an image of any celebrity, like 30-minute talk, create an AI-generated video of any goal there, and right now, I guess 70% of people will not understand that it's fake. In the very future, it will be so realistic that maybe 90 95% of people will think that it's real. So this digital signage, these things that can able to understand deepfakes a sign that is getting very popular and also in terms of just regular context, like LinkedIn, Google all of them starting to improve some logic there to be able to understand the AI generated content and just rank it very low.

Speaker 2: 22:13

So you can just go and create 30 log posts in your website with AI, hoping that Google will raise your SEO rank.

Speaker 1: 22:24

What would you say would be things that everyday listeners need to be aware of when it comes to AI-generated content.

Speaker 2: 22:32

I think they just need to take a look at how deep the context is. For example, for me, if I see a couple of keywords like delving into the sound thing, there are just keywords that if you use ai daily, it's very easy to understand. This context was written by ai or no other than that. Just you need to analyze and filter. How deep is the context. So that's how google boards working as well.

Speaker 2: 22:58

If it goes to your blog, post it has an ability to understand if it's deep research, deep technical, unique content or not. So the same thing should our brains too, if you see content that it has an ability to understand, if it's deep research, deep technical, unique. All that and all. So the same thing should our brains do. If you see it all, then that like very AI. Then you're just ignoring.

Speaker 1: 23:17

Yeah, I definitely see that. Where you know there's that, you start to see similarities amongst different language models that kind of produce certain type of keywords or phrases that you kind of hit on and if you've seen it and you kind of know it, you go well, that's definitely written by AI, it's definitely AI generated, non-authentic, and things like that. You know what is your outlook when it comes to you know kind of looking forward at you know what's your advice for a certain type of startups, or even you know some founders who are trying to journey on this navigation within AI.

Speaker 1: 23:48

How do you stay ahead of the competition with these technical advancements? What's your message for them?

Speaker 2: 23:53

First of all, it's very hard because, if you are daily checking AI news, it's growing so fast that you can just spend a couple of months on some research and then see that it became an open source, killing your business. So it's just predicting the future, trying to understand what is going to become a mainstream, like if you are trying to create, uh, for example, voice cancellation thing, I think what will be your unique proposition when it becomes an open source. It's's very hard and it's not easy, but that's what every AI founder should do to understand what will be their advantage compared to the others.

Speaker 2: 24:34

And also, very specifically, I see recently that startups who are very specific are doing a much better job than the startups who are trying to cover everything and actually not covering anything. That's quality that will attract users.

Speaker 1: 24:50

Yeah, so they're basically going too broad. They're too broad, they need to go deeper. Right, you can't cover everything. You need to have a very hyper focus on a particular niche. You need to have something that is actually solving a real world problem. Basically, right.

Speaker 2: 25:07

Yeah, exactly. I mean, if you want to build the next generation startup, then you need to find a niche. For example, if you want your product to create perfect LinkedIn posts, then you need to concentrate only on that, understand all the aspects and who are on that way, only that way, you can like beef chat, gpt Lima, all these other models. Yeah.

Speaker 1: 25:34

What do you think about the new Facebook model? Was it the Lama 3 that's just recently come out? They're talking about how it's going to be like the chat, gpt and open AI killer. What's your perspective on that? I've not had a chance to dive into it, so I wanted to ask an expert. So what's your thoughts on this new Lama3?

Speaker 2: 25:50

I haven't tried it in engineering level yet so I can't tell too much, but it looks very promising and, considering all the data Facebook has and it is a huge data it has the chances to beat GPT in terms of conversational AI, For example, like Microsoft definitely will be stronger in all generation because it owns the GitHub, so it has also trained. The model basically has access to all these conversations, so it will have a completely different conversational level.

Speaker 1: 26:24

Yeah, it's going to be interesting to see what happens. I know that everyone's watching and everyone's kind of waiting to see, kind of, what's the next big wow factor. What are some things that you're working on at HackTech that you're very passionate about? What do you see on your radar over the next 12 months?

Speaker 2: 26:41

Yeah, we are now fully concentrated again on digital transformation type of things, because it's super interesting to go to a company, traditional legacy company- find out where you can give the biggest value to increase the performance and start working from there, and our main target right now are, like Lego and marketing agencies, just because, except that involve these engineering skills, we also have a very strong domain knowledge there, which is very important.

Speaker 1: 27:09

That's incredible and for our listeners. Where can they follow you to gain more insights on HackTech and on your travels and your amazing journey through this incredible landscape that we're navigating today? Where can folks follow you?

Speaker 2: 27:24

LinkedIn. I'm very active in LinkedIn. I would say 24-7 in LinkedIn. So, yeah, that's a place to find and chat with me.

Speaker 1: 27:34

I really appreciate this. Thank you so much, jacob. I really appreciate our discussion today. I really enjoyed diving into this topic with you. I know our listeners are so more informed than you know now that they've had a chance to kind of understand a little bit more around AI marketing, advertisement and how it's going to change the legal field, would definitely love to have you on as a guest on our show. Again and again, thank you so much for taking the time. I know you're all the way out in Armenia, so it's very late in the evening for you there, but so it's very late in the evening for you there. But again, thank you for taking time to come on the show. Really much appreciate your insights on this topic.

Speaker 2: 28:09

Thanks.

Speaker 1: 28:10

Steve, I enjoyed it. Thank you so very much for joining us on today's show. Join us next time as we venture into the realms of technology and, until then, stay curious, stay informed and, most of all, happy travels. Thanks so much for listening to the Tech Travels Podcast with Steve Woodard and, most of all, happy travels.