**Napster Is Back — and It's Building Your AI Workforce | Sam Huber, Global President Napster Corp — Transcript & Summary | SozAI**
Source: https://sozai.app/transcript/napster-back-building-ai-workforce/

Sam Huber discusses Napster's AI-driven transformation and building AI workforces to redefine human-AI interaction.

## Key Takeaways

- Napster is reinventing itself by integrating AI to create new music and AI workforce solutions.
- Humanizing AI through avatars and live conversations improves user engagement and interaction.
- AI agents can significantly boost productivity by augmenting human work rather than replacing it.
- Successful AI adoption requires a mindset shift to treat AI as labor and adapt business operations.
- The future of AI in business depends on embracing technology to enhance growth and secure competitive advantage.

## What the video covers

- Sam Huber, Global President of Napster Corp, shares his unique journey from physics and Formula 1 to leading Napster's AI transformation.
- Napster has evolved from a music streaming service to a platform focused on AI-driven music creation and broader AI workforce solutions.
- The company builds AI agents and avatars to humanize AI and enable live conversational interactions.
- Sam highlights the importance of first principles thinking learned from physics in problem-solving and innovation.
- He discusses his entrepreneurial path, including founding a company integrating ads into video games and pivoting to metaverse and immersive digital experiences.
- Napster now offers a platform for businesses and creators to hire AI agent teams, enhancing productivity and interaction.
- The company emphasizes augmenting human labor with AI rather than replacing it, fostering growth and abundance.
- Sam explains the challenges and opportunities in onboarding AI agents tailored to specific business ontologies and needs.
- He stresses the shift in mindset required for businesses to treat AI agents as labor and adapt operating models accordingly.
- The video covers the future of AI in business, talent management, and the importance of embracing AI to secure future roles.

## Chapters

1. 00:00 Introduction to AI Snacks and Napster's AI Evolution
2. 02:52 Sam Huber's Background and Physics Foundation
3. 05:13 Entrepreneurial Journey: Gaming Ads and Metaverse Pivot
4. 07:42 Napster Acquisition and Rebranding Strategy
5. 10:23 Building AI Agents and Streaming Intelligence Platform
6. 12:21 Humanizing AI with Avatars and Live Conversations
7. 14:47 Onboarding AI Agents and Business Ontologies
8. 19:09 AI Augmentation, Talent Management, and Future Outlook

Answers

## Questions about this video

How has Napster transformed under Sam Huber's leadership?

Napster has shifted from a traditional music streaming service to a platform focused on AI-driven music creation and building AI workforces, enabling businesses and creators to hire AI agent teams.

What is the significance of AI avatars in Napster's platform?

AI avatars humanize artificial intelligence by enabling live conversational interactions, making AI feel more present and engaging for users.

How does Sam Huber view the role of AI in the workforce?

Sam emphasizes that AI is meant to augment human labor, increasing productivity and growth, rather than replacing employees, requiring a shift in mindset and business operations.

## Full Transcript — Download SRT & Markdown

00:03

Speaker A

[music] Hello and welcome back to A [music] Snacks with Rome and Troy, your bite-sized guide to the big ideas shaping the world of AI and robotics.

00:15

Speaker A

I'm your [music] host, Anastasia, and I'm a founder and CEO of AI Entertainment and [music] AI literacy company. Do you remember Napster, the website that let you download any song for free? And of course on AI snacks [music] we've already discussed AI in

00:31

Speaker A

music. We talked about the creation of the 10th Bhovven Symphony with the help of AI, and we had Paul Lur who created DropTrack, which is a service connecting new artists with all kinds of opportunities, for example with DJs, with

00:47

Speaker A

labels, etc. And today we have one of the key people behind contemporary Napster. And Napster was the company that the entire traditional music industry tried to destroy. But the company is actually well off. It's back and, uh, today it's

01:08

Speaker A

not kind of streaming music, but it's making music with the help of AI for you. And my today's guest is Sam Uber.

01:18

Speaker A

He is French, but he resides in Dubai, which is one of the most exciting places on earth to actually implement, uh, AI and robotics, and I would love to first discuss how did he end up in this role

01:33

Speaker A

because Sam is a very unusual person, and he is a physicist by background if I'm not mistaken, and he was working at Formula 1 not as a PR or marketing executive, but he was actually calculating Lewis Hamilton's engine

01:52

Speaker A

strategy during Formula 1, and then he created a company which was about embedding ads, uh, into video, and it was long before video streaming ads online metaverse spaces online were popular. So Sam, I will just kind of, you know, get

02:12

Speaker A

silenced and listen to you. Could you maybe talk for a little bit about yourself? Uh, describe, uh, how did you end up in your current role and covering maybe a couple of phases in between and welcome.

02:25

Speaker A

Absolutely. Yeah, thank you so much. That was a great, uh, great biography. I wouldn't have said it better myself. Yeah. So, you, you covered most of the points. I started, you know, as a teenager. I didn't really know what I was going to

02:38

Speaker A

do next, but I knew that I was interested in understanding how things work. And so it felt like studying physics was the foundation of that because more than teaching you something specific or a specific job, physics really teaches you how to think, how to

02:52

Speaker A

approach problems from a from a first principles and reason like a scientist in anything that you do to go to the bottom of things and really deeply understand concepts. So that's what I did. I was also passionate about Formula

03:05

Speaker A

1. So I managed to get a job at Mercedes Formula 1, and, uh, that was my first job as an engineer, which was very exciting working directly with Lewis Hamilton and a few others as we were the provider of

03:17

Speaker A

the, the Mercedes engine. But I also knew that I really wanted to do my own, make my own mark, have my own company and, and really be in charge of my own destiny.

03:27

Speaker A

So on the side I was looking at what are the big trends happening in the world that are just starting and that will get bigger over time, and I just wanted to be part of that. I always remember my dad,

03:39

Speaker A

who was an entrepreneur, was part of the first kind of web revolution in the late 90s, early 2000, and obviously this was, you know, 15 years later I just wanted to see what could I be a part of because if

03:53

Speaker A

you're part of the right trend, if you follow the right technology wave and you're able to position yourself, this is how you can create something big and actually make an impact because you're essentially carried by the market and at

04:06

Speaker A

the time. It felt like gaming was really going to evolve from a bit of a niche where at the time the, the, you know, the way we would characterize a gamer was someone that would play games in their basement and spend a lot of

04:22

Speaker A

time and maybe not very healthy and just a very niche kind of bad, bad the portraying of, of a gamer was not super positive. I was not a gamer myself as a kid, but I started playing games as soon

04:35

Speaker A

as games were available on my phone. So, I figured maybe there's also a lot of non-gamers that I was going to start playing games because now games are invading the phone. It felt like gaming was being democratized. And so, I

04:47

Speaker A

was playing with video games. I had an agency for a little bit building games, learning how to build a game, how difficult it is to promote the game, to monetize the game. And it felt like there was a big gap in the market in

05:00

Speaker A

creating a better way to help those game developers monetize. The main way that game developers monetize or make money is through advertising. And if you've ever opened a game on mobile, it's a pretty bad experience. You have ads

05:13

Speaker A

everywhere. They interrupt the game. So I had the idea to integrate ads in a way that they don't interrupt the game, like a product placement part of the environment. And so that concept really caught on. We were able to work with a

05:25

Speaker A

lot of games, and I was able to grow the company from zero to about 120 employees. We raised some money along the way. We made a lot of mistakes, but at the same time we were really growing and we were able to align ourselves with

05:40

Speaker A

another wave in 2021, which was metaverse, where every single brand wanted to build a presence for themselves beyond just advertising. They wanted to have their own walls and be able to engage the audience in new way and build

05:54

Speaker A

relationship with their audience. So we, we're able to position ourselves there. We changed the name of the company. But the, the idea was still to connect the brands with their audience. And then through that time I was expanding the

06:08

Speaker A

company internationally. We started it in the UK in London. Then we expanded to the US. Then I found a market in the Middle East. And the more I traveled here, the more I realized that actually this market has found a real use case

06:21

Speaker A

for this whole immersive wave, which in the rest of the world felt a little bit like, you know, it would be short-lived because there was no real substance to it. It was a lot of brands jumping in and not really building anything of

06:34

Speaker A

value. But in the region they thought about it as a digitalization just like, you know, we have, of course, the internet. The idea was everything, everything that has a physical presence one day will has a digital counterpart.

06:48

Speaker A

So every building, every city, every development could also have a digital copy that you can use for simulations, for marketing as an asset. So we tapped into that, kind of pivoted a little bit what we were doing and eventually we

07:03

Speaker A

caught the eye of a company that was doing a rollup, which means that they had raised capital and they were just acquiring multiple companies. We were the cornerstone of that rollup. So we got acquired by this company and then we

07:16

Speaker A

acquired more companies together. Eventually, we acquired Napster, which was still a music company that has been brought back to life by an investment group, but they didn't really know what to do with it. So, we were able to

07:29

Speaker A

acquire the brand. I explain later why we like the brand and what, what was the value, but we basically acquired Napster and we rebranded the whole company as Napster. And, uh, today, this is where we are. So, we have a company of about 300

07:42

Speaker A

people, and we're not just focused on music anymore. We are really focused on, we say streaming intelligence. So we essentially building a platform for to redefine the interactions between humans and AI and provide anyone, any creator, any business the ability to hire their

08:02

Speaker A

own team of agents. But we have a full ecosystem that includes hardware as well and really pushes the boundaries of how we will actually interact with AI because it's not just about the actual intelligence. It's also how do you use

08:15

Speaker A

AI? And for example, with us, it's really all about live conversations like we're having right now with an actual avatar. So that gives the AI a sense of presence. We built our own models to make AI look and feel more human because

08:30

Speaker A

we believe that to be helpful for us humans, AI has to have the right approach, the right look.

08:46

Speaker A

exciting journey and I think the place that Napster is today again I found myself at the center of this new wave of innovation with with AI. This is very cool and I think I might have seen some of your offerings uh during your uh time

09:01

Speaker A

as a gamer and implementing ads uh in video games because 15 20 years ago I was the senior VP at T-Mobile International and we were looking into several applications like uh from the mobile ecosystem perspective what was going on and what is actually the demand

09:19

Speaker A

on bandwidth uh what is the prerequisite for latency you know what is the experience on the phone and of course what is the service itself uh service delivered uh to consumer but it's very interesting what you are currently

09:32

Speaker A

describing as a kind of you know the ecosystem of services could you maybe talk a little bit about the technology powering it so you have said yes there is the hardware bit of it and then obviously uh plenty of software could

09:47

Speaker A

you please kind of you know describe a little bit further what does it mean yeah so I'll I'll talk about the the general industry right now in AI and what we see of course AI everyone talks about AI AI is on everyone's talking

10:03

Speaker A

about it everyone's using it a lot of consumers are finding a lot of value it accelerates the way they work and so on if you look at companies small companies and even bigger companies most of them are investing a lot into AI but by and

10:18

Speaker A

large they have not yet hit they have not yet found ROI for for for AI Right.

10:23

Speaker A

A lot of them are still looking for that. It's very clear why they are buying what they are not understanding.

10:29

Speaker A

Exactly. That's that that is part of it and and figuring out why do you need it in the first place? What are you trying?

10:34

Speaker A

Is it about automation? Is it about productivity? Is it about talking to your customers? Right. So there's a lot of use cases and right now we're seeing a bit of a a rush into AI. Some companies have figured it out, of

10:44

Speaker A

course, but by and large they have not. And so our theory and our thesis is that the intelligence that is provided by the large models are incredible and they are going to evolve very fast. So this is not where we want to compete. But we

11:00

Speaker A

think that the way that we engage with AI is there's a huge delta of improvement there just in improving the user experience of how we're going to engage with an AI. Right now most services whether it's Anthropic or

11:14

Speaker A

Gemini or Grock or you name it all of them the way you interact with the model is the same. You have a prompt you have to think about what you want to type it wait a little bit get an answer. It

11:27

Speaker A

doesn't sound like a lot of work right you can use that all day but still it's not a very natural way to communicate.

11:33

Speaker A

This is not how humans communicate, right? The the conversation I have with an AI right now is not natural like the conversation we are having, the two of us. And so we think that for AI to truly be helpful for humans, if we want it to

11:48

Speaker A

be everywhere in our lives and specifically in our work lives where you want to be able to collaborate with an AI, not just tell it what to do and then it does what you said, but it also always says that you're right. It's the

12:00

Speaker A

current way that you deal with AI is not a very trustworthy relationship. So we want to improve that and we want to change the interface between humans and AI. And so there's multiple things that are involved but we think that it starts

12:14

Speaker A

by making AI conversational. So everything that you do within the Navster suite of products is by having live conversation with your AI agent.

12:24

Speaker A

The second thing is we're bringing avatar faces and we are humanizing AI. So instead of just having one chatbot where you throw all of your questions, now the interface is more similar to having a suite of agents or or

12:38

Speaker A

companions that all have their own focus. So if you want to talk to someone about working out and how to create a better health plan, you go to one person. If you want want one person to debug your code, you go to another

12:52

Speaker A

person. If you have a finance question, you have another person. And of course, it can all be orchestrated by having simple conversations. This creates a much more logical interface. This is what we're used to when we are amongst

13:06

Speaker A

ourselves humans. You know, if you have if your kid is sick, you're not going to go ask the question to the same person that when you need to fix your car, right? You have specialists for everything and you know who they are.

13:18

Speaker A

And so we are creating the same the same kind of interface for AI specifically where you have those experts and you can hire those experts and just bring them into essentially the operating system of your life and then you can have

13:33

Speaker A

conversation with them and they can operate some parts of software and application to actually help you automate some of your work and some of them can actually be deployed directly in front of your customers as well to have live conversation with them to

13:47

Speaker A

understand more And this is why we have our hardware. So we have a piece of hardware which is called the station which is essentially a big kiosk that retailers can put inside their physical space. If you're an airport, if you have

14:00

Speaker A

conference, you have people coming in, you can put this kiosk there and it's powered by AI and you can have live conversation with with people that are interacting in your space and all of that contributes to learning more about

14:13

Speaker A

your customer and all of that data gets centralized and is now owned by the business. That's a big part of the Napster ethos, which is that the underlying data is your asset as a company. We do not own the data. We do

14:25

Speaker A

not do anything with the data. not monetize it in other ways. We it's basically property of of the customers.

14:33

Speaker A

I just want to to really understand what is happening in what you're describing. So you have those specialized avatars or agents and for example sales avatar might not be the same as someone who would for example talk to you about

14:47

Speaker A

medical conditions. So are you building ontologies in your data collections to really specialize and train the agent on a specific ontology or how do you do it?

15:00

Speaker A

Yes. So every agent is first trained on general material that is available online but specifically for that field.

15:09

Speaker A

So if you go to your expert Napster expert in physics, they have followed a a curriculum that would be similar to what I studied. They consumed a lot more physics material than marketing material for example, right? So they're not and

15:23

Speaker A

not everyone is trained on the same thing. This is what makes every agent different and unique and they speak like actual experts in the field, not generalists like Chad GPT would because it's just been trained on everything.

15:34

Speaker A

Then the interesting part is actually not this is that you can bring your own data. So if you are a business and you're looking for a customer service expert, you would go to the Napster app and within the 20 plus thousand experts,

15:48

Speaker A

you would find a customer service agent. This is like hiring that person, right? They have the basis. They've trained on general curriculum about customer service, how to express themselves, they speak multiple languages, and they're able to relate to the client. They know

16:03

Speaker A

what question to ask. But then to make it work for you as a brand, as a business, you need to onboard them with your content, with your best practices, with your product, what makes them unique to you. And this is where our

16:15

Speaker A

agent differentiate themselves from general advice that you would get again on chat GPT and any of those large models who are not specific to your business. So when you ask a question, they would give you general statements that you know to retain a customer, you

16:29

Speaker A

have to ask those three question, but this is not specific to your business. So it's very generic. And so we bring that next layer where every company, every user of our platform, whether they are a big government entity or a very

16:43

Speaker A

small business with three people, they can connect their data which could be connecting through APIs or it could be simply uploading documents. This is the knowledge base and this is again because what we're trying to do is to humanize

16:56

Speaker A

the AI. The analogy is that what you're doing now you've hired your digital employee. Now you're on boarding them to be an expert not at customer service but at customer service for your company which is very different. And so you tell

17:09

Speaker A

them this is my product. This is how I want you to talk to my customers. This is what's important to me. And then they essentially start acting with with that that that details and they actually understand what make the company special

17:23

Speaker A

and they can you know be essentially a digital employee that can work alongside your your human employees. So that's that's how we train them. One of the misconceptions about using foundational models like CHGPT or clude uh is that uh

17:39

Speaker A

people believe they have a conversation which happens with real time information and this is simply not doable because so CH GPT version to my knowledge which is currently being in use was frozen on the 31st of August uh last year and then of

17:55

Speaker A

course there were some checks some safety checks adjustments and then it was shipped to the clients uh to the customers in January and for entropic the freeze was I think in January and then it was shipped uh in April. So

18:07

Speaker A

there's always like a six to three months gap in between of uh training. So the the processes to train and then to ship. So how do you keep your data and adjustments current? So on top of that general intelligence that I mentioned

18:27

Speaker A

which is the the base layer as I said businesses can connect their own data.

18:32

Speaker A

So it could be a PDF with a knowledge base of how to respond to a customer in the right tone of voice but it's also an API connection to your fulfillment center or your warehouse. So you know in

18:45

Speaker A

real time how many products are actually available. So if I have a conversation with that agent, they have real-time data and so they combine the general knowledge which is the way to have a conversation right powered by the LLM

18:57

Speaker A

with a unique knowledge about the company and the real-time data that comes from these live feeds. And that's what makes this digital employee someone that actually has all the context, all the knowledge and can answer questions in real time to achieve business

19:12

Speaker A

outcomes. And Sam, what is Napster's data sourcing strategy? I'm asking this question because I sat for six years on the board of directors of Dan and Bradstreet which back then was one of the largest enterprise data companies in the world

19:25

Speaker A

on and how to source data was a crucial part in all the discussions. So how do I get what I need in order for example to provide information on what is the supply chain of this and that company,

19:38

Speaker A

who is on the board, who is the CEO etc etc. So you seem to cover so many spaces across business functions and industries. How do you source this knowledge internally?

19:51

Speaker A

Yeah, the data that we ingest is always opted in by the actual company. So we do not go and source any data. When we when a company decides to use Napster, they get essentially these digital employees that have general knowledge that is

20:08

Speaker A

essentially from the LLM but do not know anything about them. Just like an employee on their first day, they may know about accounting because they're an accountant, but they don't know about your finances because they have not been

20:18

Speaker A

on boarded. The onboarding is when they get these digital employees get access to that specific company data. And that happens with the IT department or the management of that company that decided to use Napster to decide what are they

20:34

Speaker A

going to give it access to. Again, it could be simple documents. It could be full API access. It could be a connection to the HR software. Could be connection to their CRM. They decide how much they want to give access what

20:45

Speaker A

they're comfortable giving access. And that additional data which is obviously encrypted. And we only use for that company. It's not used to train any other models. And by the way, the end customer owns all of the data. We

20:57

Speaker A

actually have solutions on prem as well as on cloud. So there is no issue with privacy there. That data is what makes those agent extra special, go the extra mile, be able to answer real question like employees. And so a lot of these

21:10

Speaker A

companies, they start by giving access to one tool, two tools, and then they realize that as they give more and more access, the system becomes smarter and smarter. And so the value exchange is actually worth it to connect more data

21:23

Speaker A

sources knowing that of course the whole infrastructure is is completely privacy proof. So that's how it works. We do not go out source any data on behalf of our clients. It's our clients that give us the data that they want their agents to

21:37

Speaker A

know about. No, that that was really a question to better understand because I was first on the impression that you kind of you know have those groups of agents for example this is the marketing specialist. This is like the technology specialist and

21:50

Speaker A

they are already trained on some ontologies which are beyond let's say what is provided by the whole internet foundational models available let's say to all of us so I I thought it was a step in between we have to some of course we also train

22:07

Speaker A

the model on you know public data or license free data or tools that are readily available of corpus of data that uh that we can access but most of the value comes from the last part which is when those companies actually on board

22:22

Speaker A

their employees with their own data. This is where we see probably 90% of the delta and 90% of the difference from asking a question to a chat GPT or using Napster is because that first party data has been integrated.

22:37

Speaker A

So that's the part that really provides the value. Yeah. And can we now kind of you know leave this data realm and go into the governance because what I'm always asking how do you actually audit the agentic workflow if there are so many of

22:51

Speaker A

agents working on behalf of your business within your business and maybe touching your business from somewhere else. So what is your solution? Because for example there is to my knowledge not one single kind of you know registry or

23:03

Speaker A

several registries of agent where you can say oh okay that was avatar x y z doing this and that. So how do you track uh because you know the world is fuller and fuller with AIS uh but we still need

23:18

Speaker A

to have some transparency on what is going on. So what are your solutions there?

23:23

Speaker A

Absolutely. So there's two answers to this. The first one is a technical answer and then there's the more philosophical answer and I think that's actually the most important. So Napster's vision is not that AI is replacing jobs. Far from it. We do not

23:38

Speaker A

buy to that narrative that everyone will be out of a job and that you know companies are going to fire the employees. We think this is very shortsighted because the goal of AI should not be to do the same with less.

23:52

Speaker A

Right? If you have a company that's doing a million dollars and you have 20 people, one strategy could be, well, now with AI, I only need 10 people, so I can do a million dollars with 10 people. We

24:01

Speaker A

don't think that's very interesting. This is something that will be capped very quickly because there's a limited amount of people that you can let go of.

24:10

Speaker A

What would be more interested is to think how can I go to 5 million or 10 million using more agents and then of course to manage these agents, I need more humans. And then because I make more profit, I can hire more people to

24:22

Speaker A

also manage the clients better. So it's more of a vision of abundance where AI helps everything grow. And it's you will have more employees and that will involve more humans and more agents.

24:33

Speaker A

Everything will grow. So that's our vision. And so part of that is that AI is not going to replace people, but it's going to augment people. There's a big difference between a job and a task. I think sometimes it's conflated in the

24:48

Speaker A

press. A job is what we do as humans and that involves relationships. That involves managing deciding what to do first, then doing it and then making sure it's done the right way. The AI really does the middle part which is

25:01

Speaker A

doing something. It doesn't really decide what to do. And it certainly does not know if that thing has been done the right way or not. Of course, there are some agents that can do some of that, but I think the beginning and the end of

25:13

Speaker A

a job is is a very human, you know, it's it's the judgment, is the taste. It's really what makes us good as what we do.

25:20

Speaker A

The middle part though, which is when you know exactly what to do, you have the guidelines, then you can get you can offload that task to an agent or more agent. So, the vision that we have is yes, workplaces are going to change for

25:33

Speaker A

sure and people do need to adapt. If you're again I like to take the example of an accountant but that can apply to to many jobs. If you are in spreadsheets all day doing people returns and so on

25:45

Speaker A

and you just want to keep doing that it's likely that an agent will be able to do a better job. But if you're more of a creative accountant in the sense that you know exactly what people should be doing, how can they maximize their

25:58

Speaker A

returns and you know fill their tax return in the right way and all the other things that you have to do and make sure that this is done according to the right tax laws and so on. Then there's a very human element to that job

26:11

Speaker A

and you can essentially have a team of agents that would do the middle part of it but you would tell them what to do and you would also check that this is done in the right way. So that person

26:21

Speaker A

will grow from an individual contributor to more of a manager of a team of accountants that will do the work for them. But now they can do the work of 10 accountants and that's just one person and so another person could do the work

26:34

Speaker A

of 10. So you used to have two people if you own the accounting firm. Now essentially you have two people but they do the work of 20. So with that you can hire a third one to the work of 10 more.

26:44

Speaker A

So this is the the way that we're thinking about it. And so because AI is not meant to replace humans end to end, we think that the idea is to give humans the right amount of controls on top of

26:56

Speaker A

the agents. Making sure that they decide what to do. They are able to instruct the agents, stop them at any time. But the fact of validating what is being done is also part of the human job. And so you know it's definitely the it's the

27:10

Speaker A

responsibility of the agents the humans managing them to make sure that they are doing a right job and obviously documented because uh you know in every company whenever you you do an IT audit you discover a jungle of

27:24

Speaker A

uh files and if those people are leaving no one can really understand what was done and and why. So this is obviously something which you envision to be a part of a human job actually to create the trace that people can understand

27:38

Speaker A

what was happening there and by the way I'm completely subscribing into this vision that AI is augmenting rather than replacing but I must say uh obviously I fund uh my consumer activities and this were very podcast because I work for

27:53

Speaker A

tons of businesses and uh I'm being asked more and more to bring AI into businesses who are lacking people because of the retirement issue and uh especially in Europe the demography uh is actually in a very bad state. So

28:09

Speaker A

people can't find apprentices to fill the jobs of workers who are currently 55 and 60 and they're going to retire in the next five years and then what uh so actually kind of you know redesigning all the processes even on the factory

28:26

Speaker A

floor and understanding how to plug in automation and AI to still keep the business up and running. So this is a huge task and uh we can't do without automation as simple as that. So what is actually if we kind of you know go to

28:43

Speaker A

the start of this conversation so you are still in creative industries so I totally understand the enterprise game and the vision but uh you are helping businesses to be more creative with your services. Could you describe what is

28:58

Speaker A

going on there? Maybe uh come back to music AI generated music. So, how do you envision this future with more creativities and services for maybe smaller companies who can't afford to hire Edelman or other huge agencies?

29:14

Speaker A

Well, most companies, especially the the small ones, if you ask any business owner, they would tell you that the the limiting factor for their business is is resources and people, right? If they had more people, they could do more. That's

29:28

Speaker A

just that's just generally the the rule of thumb. And this is even true for for very big companies. And that relationship has been between people and and resources has been true forever since the early days of business. This

29:41

Speaker A

is the first time where this is changing. And companies are no longer going to be limited by headcounts.

29:47

Speaker A

You're not going to need more people necessarily to do more things. And so we call that the concept of elastic organizations. a little bit like the cloud enabled companies and and made essentially compute and infrastructure elastic. Before the cloud, you had to

30:04

Speaker A

buy servers and the amount of people you could get on your website was limited by the capacity of your servers. So, you were limited by that. If you wanted more, you had to invest more. With cloud, it's elastic, right? If you have

30:15

Speaker A

people, you'll spend. If you don't have people, then your cost is going to be very low. So this is the same concept that for organizations where based on the demand you'll be able to have more employees and you'll have your base of

30:28

Speaker A

human employees and then you have your digital employees that you can hire very quickly and on board very quickly by connecting your data and now they can do tasks for you. So if you need something there's a rush finish the quarter and we

30:40

Speaker A

need more customer service or more sales or more creative people because we have this big pitch you will be able to bring those agents very quickly in literally a matter of hours and then start outsourcing and offloading to task to

30:53

Speaker A

them of course with humans supervising that. So the more agents you have the more humans you need as well and and your organization scales up and down based on the demand. But the day where you do not need that many resources, you

31:05

Speaker A

just stop interacting with the agents and with the Napster model you do not pay for it because you only pay for for the live interactions. So, so that's the idea and it applies to every industry.

31:15

Speaker A

Creative is is one of them where you know we have agents that have been trained on a lot of UIUX material for example and so they're able to come up with concepts for designing an app or website. So you can apply you know the

31:29

Speaker A

at the core we're not trying to become a a verticalized agent. We think there are a lot of companies that are doing agents for marketing, agents for finance, agents for this, agents for that. And the problem with that is when you start

31:42

Speaker A

plugging that into an organization, the organization looks like Frankenstein with a lot of these kind of strapped on agents that do a little bit of this, a little bit of that, but actually they don't talk to each other. The marketing

31:54

Speaker A

agent doesn't know what the finance agent is doing and no one has really control over what they are what the results are, right? The agent could deliver great results for finance, but maybe somehow it hurts another part of

32:06

Speaker A

the business and no one knows it. There's no single view of that. They're not, you know, the the bottom line is like all these agents are pieces of software. They're not yet employees or labor. And so Napster's vision is to

32:19

Speaker A

rethink that completely and to enable these companies not only to add agents but actually to rethink their operating model. And we're saying the operating model in the age of AI is an organization that is a mix of human a

32:33

Speaker A

employees and synthetic employees which are the the the digital you know the agents and that put together in the same organization. So your agent are part of the organization they have context on the organization. So they know what team

32:47

Speaker A

is doing what, what are the reporting lines. And that is why you can have live interactions with them. I could go to my agent and say, I want to email, I want to get all the US salespeople on the

33:00

Speaker A

call tomorrow to talk about a strategy. It's not going to ask me, okay, who are the US salespeople? And you have to provide a list. No, it knows because it knows who's based in the US. He knows who in the sales team. It knows their

33:11

Speaker A

email and he can just send an email. That's why that's when it becomes useful. No other agents can do that because they do not have the context on the on the organization. So we think that this is how you get to the next

33:23

Speaker A

level. Sam, I really love this concept of elastic organization and I I must admit uh you know I mentor executives as well as part of of what I do and I talk a lot to CHRO's uh globally and I'm always

33:36

Speaker A

asking this question. Could you please describe what is the vision for whatever next generation R or people function because we now have uh the world of automation and AI and it's very very rarely that I hear something which is

33:52

Speaker A

convincing it's all kind of you know very old traditional thinking so we have like hiring we have of course firing and then we have this on boarding processes some esoteric concept of talent management uh working with some mentors

34:07

Speaker A

coaches providing courses but there is no this idea of what is the business strategy where do we want to go what is achievable with this current workforce and what could we add on top of it uh with the help of automation and AI and

34:23

Speaker A

how to bridge it and how actually to educate current people so real bodies to use automation and AI and actually actively spotting those applications and maybe kind of you know calling out for those providers who might bring value. I

34:40

Speaker A

guess what you are describing put you in front of a lot of um executives and those people function or not. I am not I do not know may you talk what is your level of discussions with them? what are

34:54

Speaker A

you hearing from the market and I assume that Middle East might look differently from I don't know Europe uh traditional European Union and then Switzerland where I desire the UK and then United States.

35:07

Speaker A

Yeah. So I personally I lead our efforts in the Middle East and the great thing about this region is that on top of the desire to implement AI there's also a strategy a government strategy which pretty much fixes and determines what's

35:22

Speaker A

what needs to happen for government services for public sector for private sector over the next couple of years. So within two years there needs to be a 50% automation of most services and then there's another deadline three years

35:35

Speaker A

later which is kind of the the the first or second checkpoint for this AI strategy. So everyone is scrambling to figure out what to do and there's even more momentum even more pressure to implement the solutions. So the region

35:47

Speaker A

is not necessarily where the technology gets built but it's where it gets rolled out and this is why it's a very exciting place to to be and the conversation we have at the highest level with chairmans or sea level of very large companies

36:00

Speaker A

sometimes sovereign funds which are kind of parent company that own multiple companies in the region and they're all trying to figure out how can we build AI native organizations how can we really think what the organization is going to

36:15

Speaker A

look like in the future So in the region the interest for agents for marketing here or agents for finance there is very little. They're really thinking long term about what is the workplace going to look like? What is agents everywhere?

36:30

Speaker A

And so the shift in mindset is not thinking about agent as software. It's about thinking of them as labor and also being willing to change your operating model. not just the software stack and and how business is done but actually

36:45

Speaker A

the operating model for example HR we have a lot of use cases around H and helping to automate performance reviews or on boarding and and you name it and this requires a drastic change in the way that the HR profession actually

37:00

Speaker A

works people will be displaced and probably will have to grow into the role of managing agents as I was describing earlier but most organizations are not willing to do that because the people that are in power might be threatened by

37:15

Speaker A

by the agents. So I think this is something that will happen over time. You know people and organization can resist change for for a while but ultimately in the you know in a capitalist society the companies that have done the right decision will

37:29

Speaker A

benefit from it financially and they will grow and they will hire more people and they will become the the company that everyone want to work at and I think eventually everyone will have to change their operating model and in the

37:40

Speaker A

future every organization will be elastic. Well, I sometimes wonder whether some European companies are indeed capitalistic companies or companies from the realm of socialism because the sense actually is really leaving the building there and sometimes decision are made which have nothing to

37:57

Speaker A

do with uh let's say the laws of you know how to generate profits but this is probably another discussion and uh the explanation why Europe is uh in such a state uh in in which it is uh is

38:10

Speaker A

actually quite a very large geography right with if we don't count just the European Union but everything it's like 900 million people uh so huge missed chance Sam I am really absolutely loving what you are saying so you introduced uh

38:26

Speaker A

a lot of interesting concepts to my audience uh to think about elastic organization which is not just about companies I think we need to think about educational institutions in the same way because uh you know an agent is is not

38:41

Speaker A

just uh let's say a sales clerk, it might be someone who is explaining chemistry to you or for example how to do a lab experiment. We really need to rethink governmental services and of course it helps if the Gulf governments

38:58

Speaker A

implement uh a clearly defined strategy and they have goals how much to automate in what year. So of course it's very very helpful. So Europe does not have any of that. And I think it's very helpful to remind people that AI is not

39:13

Speaker A

about replacing, it's about augmenting. But those who don't know AI and who are simply afraid of it might actually fail to secure uh good positions, good roles and uh be actually in charge of their own future. If you are to speak to

39:30

Speaker A

students who are currently entering the university and are studying something so what would be your recommendation to them studying today whatsoever they want to study but in the age of AI I think the first thing is is really to

39:47

Speaker A

adopt the technology and and you know learn it and use it for different things but always keep a critical eye on on the results because again those uh this technology ology is not bulletproof.

39:59

Speaker A

It's an algorithm that is based on past data and essentially putting data together and and coming up with the most likely answer and not always necessarily capable of coming up with drastically new concepts. So I think you know if

40:14

Speaker A

you're a student now you have to understand how the technology is built because you want to retain a a critical eye on the output and make sure that you use it again to augment you not to replace you. We actually have an

40:29

Speaker A

education product at Napster which we are rolling out in multiple universities and enable the professors to essentially create an a digital copy of themselves.

40:38

Speaker A

So students have access to the professor 247. they can ask questions and it's all AI an AI professor trained on their data so it has exactly the same amount of context so I I we talk to a lot of

40:50

Speaker A

students and professors but I think a lot of the students would take sometimes the information at face value you know and so it's very important to know that this is again you you do not want AI to replace you and if you do that that's

41:04

Speaker A

essentially what you're doing if AI is doing your entire exam AI is replacing you this is not the best way to use AI AI I should augment you. You need knowledge in the field and the more knowledge you have then you can use AI

41:16

Speaker A

to your advantage as a super tool so you can do even more things. But if you do not have any knowledge, AI won't help you because it would just give you a broad answer and you wouldn't be able to

41:27

Speaker A

tell if it's good or not. So you still have to be an expert more than ever. You still need to understand topics. You have to be ahead essentially of of the AI in the in your field of interest and

41:38

Speaker A

then you can use this AI as a tool to do even more things. Sam, thank you so much. And that was another episode of AI snacks. Romy and Robbie, I'm tremendously grateful for your time. Our audience grew to 12,000

41:53

Speaker A

listeners uh this day. So once again, thank you for tuning in. And uh there's an a reminder in the first two weeks of November I'm offering two free webinars uh because I am going to reveal the results of the global AI literacy study

42:06

Speaker A

which I'm carrying out on behalf of the house of lords in the UK. The results uh will be available on the 22nd of October and I will be discussing in those webinars some interesting figures and insights across four categories.

42:20

Speaker A

educators, students, young people, professionals aged 45 years and older, and then small and midsize businesses.

42:30

Speaker A

Once again, thank you, and I will see you next week. Thanks for having me.

Topics: Napster AI workforce Sam Huber AI avatars music AI streaming intelligence human-AI interaction AI agents metaverse digital transformation

Study this video

- [Flashcards from this video](https://sozai.app/tools/ai-flashcard-generator/?from=transcript&id=77169)
- [Quiz on this video](https://sozai.app/tools/quiz-generator/?from=transcript&id=77169)

Free, made by AI from this transcript. No signup.


---
This is the markdown twin of https://sozai.app/transcript/napster-back-building-ai-workforce/ — the same content, without the markup.
Published by SozAI (https://sozai.app). Reuse and quotation are allowed with attribution and a link back.
Machine-readable index: https://sozai.app/llms.txt · data API: https://sozai.app/api/