Latest / Glovisor Podcast / Ep. 57: Your Company Has AI Chaos — Here’s How to Take Back Control | Intrascope (Vladimir Krsic)

Transcript

Ep. 57: Your Company Has AI Chaos — Here’s How to Take Back Control | Intrascope (Vladimir Krsic) · Glovisor Podcast

• 0:00
  Speaker 1: They built a tool that finally stops AI chaos. One workspace,
• 0:04
  Speaker 1: every model, total control. Welcome to the Glowvisor podcast, where
• 0:10
  Speaker 1: we explore business growth, communication, and personal development through real
• 0:14
  Speaker 1: conversations with people that make an impact. Our guest is
• 0:18
  Speaker 1: the founder and CEO of Interscope, an infrastructure platform for
• 0:23
  Speaker 1: generative AI inside companies. After years of building software and
• 0:28
  Speaker 1: IT systems for businesses, he launched Interscope to help teams
• 0:32
  Speaker 1: manage AI models, share context, governance, usage, and costs. Interscope
• 0:39
  Speaker 1: is a member of NVIDIA Inception. Will Weiser welcomes Vladimir Khristush.
• 0:45
  Speaker 2: And thank you for the call.
• 0:46
  Speaker 1: Vladimir, let's begin. Please share with us a little bit
• 0:48
  Speaker 1: more about what you do.
• 0:49
  Speaker 2: Our startup is helping companies to use AI in a
• 0:53
  Speaker 2: structured way. have a full governance around the AI. So
• 0:59
  Speaker 2: we provide them infrastructure layer for using AI inside the companies.
• 1:06
  Speaker 2: Before that, we started as a software company and we
• 1:10
  Speaker 2: built these solutions for ourselves. And after we realized that
• 1:16
  Speaker 2: solution is good for our use cases, for the generative
• 1:21
  Speaker 2: AI tasks in the company, we share it with customers.
• 1:25
  Speaker 2: our friends and our colleagues, and in that time we realized, okay,
• 1:33
  Speaker 2: we need to launch it on the public. So we
• 1:36
  Speaker 2: had our public launch on Product Hunt, finished nine products
• 1:41
  Speaker 2: of the day, and the journey started.
• 1:43
  Speaker 1: There's a lot there to go on, but before we
• 1:46
  Speaker 1: go there, tell us a little bit more of how
• 1:48
  Speaker 1: you got to what you're doing today. What made you
• 1:50
  Speaker 1: choose this path?
• 1:51
  Speaker 2: I have that inside myself. that some feeling that I
• 1:59
  Speaker 2: need to do something more than the regular job. So
• 2:03
  Speaker 2: I started my first company when I was 20 years
• 2:06
  Speaker 2: old and working with my corporate job. And at one
• 2:14
  Speaker 2: point I decided, okay, I cannot do the same stuff
• 2:18
  Speaker 2: every day. I need to do every day something new
• 2:22
  Speaker 2: to be thrilled to do. go happy on the workspace
• 2:26
  Speaker 2: or very sad and afraid. But I'm looking for that
• 2:35
  Speaker 2: feeling that makes me full. And that's why I started
• 2:41
  Speaker 2: working with startups.
• 2:42
  Speaker 1: And that's so common that I see with a lot
• 2:44
  Speaker 1: of founders like yourself is these are individuals that have
• 2:48
  Speaker 1: so many ideas and so much creativity because you have
• 2:51
  Speaker 1: to be very creative. And you have all that energy
• 2:54
  Speaker 1: inside of you and it has to come out, right?
• 2:56
  Speaker 1: It has to find a channel to go through that.
• 2:59
  Speaker 2: Because you have a lot of ideas and you need
• 3:02
  Speaker 2: to define the right ones that you will focus your
• 3:06
  Speaker 2: time and money and everything. So it's, you have ideas,
• 3:10
  Speaker 2: but you need to choose wisely. Sure.
• 3:14
  Speaker 1: So now, before we go into Interscope, let's talk about,
• 3:17
  Speaker 1: can you get us on the same page? What is
• 3:19
  Speaker 1: the difference between regular AI and generative AI?
• 3:23
  Speaker 2: Generative AI is artificial intelligence, yeah. So it's general term
• 3:28
  Speaker 2: for all the whole process that AI can do. But
• 3:35
  Speaker 2: generative AI is that part of AI that you are
• 3:40
  Speaker 2: using to create something. So It can be text, images, presentation, right?
• 3:46
  Speaker 2: So that is the main difference. Gotcha.
• 3:50
  Speaker 1: Now, what made you go into that direction with Interscope?
• 3:55
  Speaker 2: What did you.
• 3:55
  Speaker 1: Find that was missing in the marketplace that you're filling
• 3:59
  Speaker 1: with your startup?
• 4:02
  Speaker 2: Companies are adopting AI faster than they can manage it.
• 4:07
  Speaker 2: So we find that market gap, it's It's a very
• 4:12
  Speaker 2: new market because generative AI is a few years old.
• 4:17
  Speaker 2: And companies want to find the ways that they can
• 4:21
  Speaker 2: achieve better performance and be more effective. And they're just
• 4:29
  Speaker 2: finding so many different tools and providers and start working
• 4:35
  Speaker 2: with them. And they forgot about company context, company knowledge,
• 4:41
  Speaker 2: and the cost that different models can accumulate into one place.
• 4:50
  Speaker 2: So we are helping them to use it like every
• 4:56
  Speaker 2: other system in the company. So you can use it
• 5:00
  Speaker 2: in the, we can say, better way.
• 5:04
  Speaker 1: Because one of the things that I see when I
• 5:05
  Speaker 1: talk to a lot of CEOs who are just infusing
• 5:08
  Speaker 1: AI into their companies is that they know they have
• 5:11
  Speaker 1: to invest into AI because everyone else is doing so.
• 5:14
  Speaker 1: But they have one system. They have another system for
• 5:18
  Speaker 1: something else. They just know they need to pour money
• 5:21
  Speaker 1: into it. But now they have this chaos of AIs.
• 5:25
  Speaker 1: And that's what you're doing. That's basically just putting it
• 5:27
  Speaker 1: all together into one.
• 5:29
  Speaker 2: Yeah, yeah, yeah. Because everyone... there is a battle between
• 5:34
  Speaker 2: AI providers like Anthropic, CGPT, Google, etc. But there is
• 5:42
  Speaker 2: no clear winner which model is better. Every model is
• 5:47
  Speaker 2: good for something. So, you're basically using Gemini for the
• 5:52
  Speaker 2: web search because it's a model from Google and using
• 5:58
  Speaker 2: Cloud for one task, CGPT for others and we are
• 6:01
  Speaker 2: just connected onto one place. So you can use it
• 6:05
  Speaker 2: from one workspace. You can use it for different aspects
• 6:11
  Speaker 2: of the company. Sure.
• 6:13
  Speaker 1: Yeah, because you have to find the right one for
• 6:15
  Speaker 1: the right task. So you're not doing too much. So
• 6:20
  Speaker 1: who would be the ideal client for your company?
• 6:23
  Speaker 2: We had our ideal client before. some marketing agencies between
• 6:31
  Speaker 2: 20 and 100 users. But in this moment, we a
• 6:38
  Speaker 2: little change our perspective to enterprise clients because we can
• 6:44
  Speaker 2: give them the best value. So if they invest in
• 6:49
  Speaker 2: the structural use of AI, we can make them. Cost reduction.
• 6:56
  Speaker 2: better than on the small company. And for the bigger enterprises,
• 7:01
  Speaker 2: data is more important than the smaller companies. So we
• 7:06
  Speaker 2: are balancing between companies between 20 and 100 and big SAs.
• 7:15
  Speaker 1: If one of our listeners wanted to work with you,
• 7:18
  Speaker 1: what would that journey look like?
• 7:20
  Speaker 2: We made a discovery call and helped them to choose
• 7:24
  Speaker 2: which model they they need to use because we have
• 7:29
  Speaker 2: a SaaS model for using our platform and tailor-made model
• 7:34
  Speaker 2: for enterprises. So we need to decide with the client
• 7:38
  Speaker 2: what type of service they need. So if they have
• 7:44
  Speaker 2: some confidential data, don't want any other AI agent outside
• 7:53
  Speaker 2: their organizations use that data or just they want to
• 7:58
  Speaker 2: go on other servers, we can provide them the full
• 8:02
  Speaker 2: closed system on their servers, connecting with their small language models,
• 8:10
  Speaker 2: made it just for them. Or they need it just
• 8:13
  Speaker 2: for a regular task for writing to content of the
• 8:17
  Speaker 2: 10 people team. Okay, you need to choose SaaS and
• 8:21
  Speaker 2: if You need something bigger than that, we can talk later,
• 8:24
  Speaker 2: but that is the main difference of our approach. But
• 8:30
  Speaker 2: Discover Call is the first one.
• 8:32
  Speaker 1: You can do all of those, but can you give
• 8:34
  Speaker 1: us an example? Why would an organization want to have
• 8:38
  Speaker 1: their own system in-house versus using a SaaS?
• 8:43
  Speaker 2: It's given more control with the context tool and the privacy. Also,
• 8:52
  Speaker 2: we are living in the era where companies compete with
• 8:57
  Speaker 2: each other. This is just the starting of AI. We
• 9:02
  Speaker 2: can say that this is the most affordable time to
• 9:07
  Speaker 2: use AI because it will only get expensive time by time.
• 9:13
  Speaker 2: And companies will start to realize that it it's better
• 9:18
  Speaker 2: to them to get one investment in the hardware or
• 9:23
  Speaker 2: virtual private and deploy their own LLMs so they can
• 9:30
  Speaker 2: use it locally. To not to be dependent on the
• 9:35
  Speaker 2: bigger ones or anthropic or open AI or things. So
• 9:43
  Speaker 2: they have that one closed cycle. It's interesting.
• 9:48
  Speaker 1: And I would imagine also, because you don't want your
• 9:49
  Speaker 1: data floating out there somewhere, you want to protect that
• 9:52
  Speaker 1: as well.
• 9:54
  Speaker 2: Yeah, for every provider that we support, they have in
• 10:00
  Speaker 2: terms of search that they will not use data through API,
• 10:04
  Speaker 2: but you'll never know. True.
• 10:08
  Speaker 1: Yeah. It's the reality of the world nowadays.
• 10:10
  Speaker 2: Yeah.
• 10:11
  Speaker 1: Now, can you share a client success story of someone
• 10:14
  Speaker 1: who's worked.
• 10:14
  Speaker 2: With you in the past? Yeah, I have one. one
• 10:18
  Speaker 2: very good example. The company that they didn't use any
• 10:24
  Speaker 2: paid AI, they're just using some free versions. They are
• 10:31
  Speaker 2: a tax consulting company. They're more than 50 employees. And
• 10:38
  Speaker 2: they're using JGPT free version, Gemini and everything. Because we
• 10:44
  Speaker 2: are working with them for other projects when we made
• 10:49
  Speaker 2: it through scope and everything we approach them and tell
• 10:52
  Speaker 2: them okay can help you to start using AI because
• 10:56
  Speaker 2: we don't we didn't do transformation of their business you
• 11:00
  Speaker 2: just implement it the right way and they start using
• 11:04
  Speaker 2: and they see next day quality of the answers because
• 11:10
  Speaker 2: they use AI to support clients or reply to emails
• 11:17
  Speaker 2: and that is just another dimension for the CEO and
• 11:22
  Speaker 2: for the management we provide them that dashboard that they
• 11:26
  Speaker 2: can see how much every employee is spending on AI
• 11:30
  Speaker 2: and savings it's there is no savings because they didn't
• 11:37
  Speaker 2: use paid AI but if they are using paid AI
• 11:41
  Speaker 2: versus our solutions It's more than 70% of cost savings. Wow.
• 11:48
  Speaker 1: And how are they reaching those cost savings?
• 11:50
  Speaker 2: When you use the AIHU API system, you're paying only
• 11:56
  Speaker 2: for the usage, not some price that you may never
• 12:03
  Speaker 2: come to that price. So, like, example is JGPT +,
• 12:10
  Speaker 2: it's costing $ 20 per month and you will maybe... do
• 12:14
  Speaker 2: that month 10 prompts and you spend $ 0. 001 for that
• 12:21
  Speaker 2: month and you will pay $ 20. There is some power users,
• 12:26
  Speaker 2: but that is the other case. But with Interscope and AI,
• 12:32
  Speaker 2: we have automatic orchestration of the models. So when you
• 12:37
  Speaker 2: type the prompt, our inside system see and choose the
• 12:46
  Speaker 2: right model for that task. So if you need translating something,
• 12:50
  Speaker 2: it will be using the most cheapest model on the market.
• 12:53
  Speaker 2: And if you do some very deep research, it will
• 12:58
  Speaker 2: be using the top tier models, but not every employee
• 13:01
  Speaker 2: needs top tier models. So that's how we do that cost.
• 13:07
  Speaker 2: That makes sense.
• 13:09
  Speaker 1: Now, if listeners remember only one thing from today's conversation,
• 13:13
  Speaker 1: what do you hope that message is?
• 13:15
  Speaker 2: They need to start using AI in a more controlled way.
• 13:23
  Speaker 2: They don't need to use our product, but they need
• 13:27
  Speaker 2: to find something to be steady and to scale it
• 13:35
  Speaker 2: and to track the cost and protect their data. So data,
• 13:41
  Speaker 2: it's very important for every company. If some of your
• 13:45
  Speaker 2: employees are using your personal accounts and upload there your
• 13:49
  Speaker 2: tax report and ask free model something, that tax report
• 13:59
  Speaker 2: will be used for the train of AI. This is
• 14:05
  Speaker 2: the time where data is for sale. Everybody's buying data,
• 14:11
  Speaker 2: selling data, and companies need to start to protect that data.
• 14:16
  Speaker 2: That's good advice.
• 14:17
  Speaker 1: And now finally, where can our audience connect with you?
• 14:21
  Speaker 1: Find out more about your work and continue this conversation.
• 14:24
  Speaker 2: We can connect on LinkedIn. This is my name here,
• 14:27
  Speaker 2: so you can find me, Vladimir Kristich, or visit intrascope.app,
• 14:33
  Speaker 2: our website. Excellent.
• 14:36
  Speaker 1: Now, Vladimir, thank you so much for joining us. We
• 14:38
  Speaker 1: really appreciate your time and the value you shared.
• 14:41
  Speaker 2: Thank you for this opportunity. You're welcome.
• 14:45
  Speaker 1: Thanks, everyone, for listening. Did this episode inspire you? Make
• 14:48
  Speaker 1: sure you share it with someone who needs it. And
• 14:49
  Speaker 1: don't forget to follow the Globalizer podcast for more conversations
• 14:52
  Speaker 1: that make an impact.
• 14:54
  Speaker 2: Till the next time.