Hello there! Welcome to AI Gamechangers, your weekly guide to the people putting artificial intelligence to practical use across the games industry. In every edition, we sit down with someone working at the intersection of AI and games, from the tools being built and the problems being solved, to the bigger bets on where all this is heading.
This time, we’re chatting with Sam Barberie, who joined Google earlier this year to build out its AI practice for gaming, media and entertainment. Sam’s spent a career watching platforms either join the dots or fail to, and he has insightful things to say about why most studios are still leaving value on the table, what transmedia has to do with it, and why the real opportunity in AI isn’t just making games faster, but making player relationships smarter.
Scroll to the end, as ever, for a curated round-up of AI news from around the web. And remember to read all our other conversations with industry leaders.
Sam Barberie, Google
Meet Sam Barberie, a specialist in AI-driven customer experience at Google, where he’s building out the company’s practice for gaming, media and entertainment.
Sam has a rich background in the games and interactive media space, with previous stints at SuperData Research (acquired by Nielsen), Animoca Brands, and blockchain platform Sequence. He’s spent his career at the intersection of data, platforms, and player engagement.
In this interview, we explore what a truly joined-up AI stack looks like for games companies, and why Sam thinks retention, not just monetisation, is where the real opportunity lies.
Top takeaways from this conversation:
Google’s Gemini Enterprise for Customer Experience is an end-to-end stack for games companies, covering everything from pre-sales and content creation through to in-game player support, connected through cloud infrastructure.
Sam argues that AI’s biggest underexploited opportunity in games isn’t development efficiency but player relationship management: personalising support, anticipating problems, and keeping players engaged.
The transmedia opportunity is real and accelerating. As game IP expands into film, TV, and merchandise, Sam sees Google’s connectivity play as the thing that could finally make those experiences feel like one unified whole.
For small studios, Sam’s practical advice is simply to start somewhere and iterate fast. AI’s value reveals itself differently at each stage of development, and the studios that benefit most are those willing to get their hands dirty.
AI Gamechangers: You’ve just joined Google to establish an AI practice for gaming, media, and entertainment. What does that mean day-to-day?
Sam Barberie: Specifically, I’m a specialist on the product suite we call Gemini Enterprise for Customer Experience. That is a recently renamed product line which Google Cloud has had for a number of years, specifically to improve the experience across the entire life cycle of a customer (from pre-sales and helping them think through purchasing decisions, all the way through ongoing support).
What excites me a lot about that, specifically for gaming, media and entertainment, is that these consumer journeys are the most complex typically, because they involve a lot of different ways of engaging with an IP, particularly as companies get increasingly more multimedia and transmedia. But we’re also talking about a consumer base that is by far the most knowledgeable, passionate and sensitive about what their experience looks like.
And so there are things that happen in that journey where having knowledge becomes crucial; as an audience member having a specific challenge, being able to access information or support in a way that is finely tuned to you as a player, is important. In gaming, that’s something we’ve left up to “game design” by itself – we expect the game design to be intrinsically good at monetising, and getting me to think about game items or upgrades or anything else that I might need.
It can always be better, and it can always be more personalised to a specific player. If I’m running into a problem with something in the game, or a purchase, or my account has been banned, for instance, they can be fine-tuned to address that. It can know that it’s dealing with someone who’s already spent money, and somebody with a fervent passion to have this gaming experience.
What attracted you to Google as the next chapter in your career? Why now?
It comes down to the opportunity Google represents – not just in tackling AI generally, but specifically in gaming and entertainment.
My most recent company was a full-stack development platform for web3-enabled applications, including games. One of the core challenges we saw in that nascent market was that developers were building in a really fragmented way – grabbing this wallet, that indexer, jerry-rigging a stack just to get things to work. And in the meantime, you’re accumulating technical debt, expanding your attack surface for vulnerabilities, and ultimately not building the kind of comprehensive player experience you’re actually after.
I kept coming back to the constant refrain: you can’t build half a game on Unity and half on Unreal and expect it to work. It just doesn’t. And I think the same is true of AI transformation for companies, whether that’s early in the content creation stage, all the way through to purchasing, customer support, and everything in between.
“I like to point out that the first real application of AI was a game! Gaming and AI have always gone together, and of course, every NPC is some flavour of AI. So as much as people have concerns about AI in gaming, it’s already there”
Sam Barberie
What attracted me to Google, and specifically the way they’re solving for the customer experience journey, is that this is already where companies across the globe have their data. They already use GKE [Google Kubernetes Engine] to run their applications, they do content creation here. So what does it look like when that connects through to a player account – to everything someone is doing inside a game?
And that puts Google in the strongest position to help game developers across the full constellation of challenges they’re facing right now: rising development costs, stagnating monetisation, and (the part I think people often overlook, beyond UA and pure monetisation) retention. How do we keep players happy and engaged? Especially in a live service model, we tend to think about feeding players more content, but not necessarily about what else needs to come with that: improving the experience, making sure it’s fulfilling and satisfying, building the kind of sentiment that keeps people around longer.
Google has a modular but complete stack for everything a game does, and that’s a genuinely strong position to be in. For me personally, the decision came from having watched developers make choices that didn’t quite fit and creating more work for themselves. And right now, no game developer has time for that.
You wrote in your own newsletter that AI itself can’t save gaming. Can you unpack that a little bit for us?
What I was getting at in that piece was this idea that AI is some kind of magic wand that’s going to solve everything. And if you’ve ever worked with gamers or built a game, you know that’s completely impossible! Gamers are hyper-vigilant. They’re the hardest consumer set to please, and they have literally millions of options.
Given all that, the human creators behind games are always going to matter enormously. As Jack [Buser] often says, AI is really about giving those creators superpowers – not replacing any component of the creative journey, the gameplay experience, or the consumer relationship. It’s about enhancing that.
And I think that’s where a lot of developers get genuinely concerned. We see this with every technological shift, every monetisation change, every distribution change in gaming – it’s always the same reaction: “This is a scam, it’s a cash grab, it’s going to kill creativity!” And that’s never actually been true. New platforms unlock new player bases, drive incremental monetisation, and ultimately bring new benefits to the industry. We’re at one of those moments again.
“The intersection of AI, UGC and web3 is the future of gaming. Web3 can be adopted to whatever degree different companies are comfortable with. But at the very least, we’re moving towards a world where not accepting crypto will be like not accepting Visa or Mastercard”
Sam Barberie
I also like to point out that the first real application of AI was a game! I think it was around 1956. Gaming and AI have always gone together, and of course every NPC is some flavour of AI. So as much as people have concerns about AI in gaming, it’s already there. With generative AI, the worry becomes, “If anyone can create anything based on anything, what does that mean for IP protection? For content quality?”
And I think that’s fair. But it was already playing out before generative AI arrived. Every year on Steam has been a record for the number of games published. It’s getting easier to make a game, and most of those games, just like every year before, won’t succeed.
Where AI genuinely helps is in accelerating time to market. Gaming is an iterative process, so being able to get something into Early Access or closed beta, gather feedback and fine-tune quickly: that matters. It also just means more shots on goal.
Could we take that to some theoretical extreme, where the market is suddenly flooded with gibberish gaming noise, brainrot, and things like that? Maybe. But the structure of the gaming market has always winnowed out experiences that aren’t sustainable. There will be ecosystems that thrive on scrollable, disposable gaming content – but even within those, each game still needs to be inherently good for a period of time. And that comes down to all the things we’ve always associated with success in gaming: a bit of luck, maybe getting picked up by the right streamer, having the right monetisation strategy and UA plan - and ultimately, be a game that’s creative, well-designed, engaging, and built to stick around.
We spoke with 10six Games recently, who are making You vs Zombies using Gemini. The game generates assets on the fly, as needed to personalise the game. Is that the kind of new experience use case you’re championing?
Where I see AI fitting in is primarily around solving universal developer problems: monetisation, user acquisition, retention, content creation. But like all new technologies, it also creates the opportunity for genuinely new gameplay. And that’s something we’re going to see studios experimenting with, which is really exciting.
Take personalisation: how do you dial it in so that an experience feels tailored to you, but still retains enough of a shared quality that players can connect over it? If you and I are both playing the same game but seeing completely different experiences, you lose the social element: the chatter, the shared reference points, all the stuff that comes with engaging in content together. That balance is an interesting design challenge.
But the opportunities around net-new experiences are honestly one of the things I find most exciting right now. If you’re an entirely new studio, you have a full suite of AI capabilities to help you get to market faster, and by doing so, maybe you feel free to think more creatively, to take on the risk of trying something genuinely different, because you can iterate quickly if it doesn’t land.
I think we’re going to see more and more of that from new studios. They might call themselves AI-first, or something like that, but really, they’re using the tools available to them to create things players haven’t engaged with before. And that opens up an entirely new paradigm of play experiences.
You mentioned transmedia. What’s your observation about where that’s going? And how Google can make it all feel joined up?
Over the past few years, the revenue opportunities in transmedia have become pretty hard to ignore. If you rewind 20 years, there were some very wonky film adaptations of games, and game narratives shoehorned into films, and they largely didn’t work.
But we’re finally at a point where the mass audiences available for both game IP and film or TV make this genuinely commercially viable.
People aren’t thinking of gaming as a niche thing with a guaranteed core audience anymore; they’re recognising the opportunity to appeal at real scale. Whether it’s the Mario movie, The Last of Us, or any number of other game IP adaptations, the results over the past few years have been enormous.
“Gamers are the hardest consumer set to please, and they have literally millions of options. Given all that, the human creators behind games are always going to matter enormously. AI is really about giving those creators superpowers”
Sam Barberie
And we’re seeing the reverse too: film and TV studios, Netflix included, taking their own IP and reach and building serious gaming strategies around it. It’s gone well beyond the old licensing play of “take my IP, hand it to someone else, put it on a lunchbox, collect the royalties.” Now the thinking is: how does a gaming strategy become an integral part of expanding the surface area through which I interact with my customers – whether that’s through advertising, streaming subscriptions, physical merch, or something else entirely?
The Google opportunity here is, again, about connectivity, making all of these experiences feel like one unified whole. Disney is the perfect example of this. They’re a famously siloed company. The parks division doesn’t talk to the merch team; Disney+ sits apart from everything else. Historically, it’s been deeply fragmented. But they’ve started to consolidate, recognising that passive IP fandom alone won’t unlock the greatest opportunities. What will is actively leveraging every touchpoint they have with customers to reward them, engage them, and draw them across multiple experiences.
In Google terms, that could look something like a shopping agent. Say I tell it I’m taking my family to Disney World and we’re huge Star Wars fans. It can put together a tailored package based on when I want to go and what kind of stay I’m after, and actually complete the purchase. And by the way, did you know there are exclusive Star Wars items you might want before you travel?
That kind of joined-up experience has never really been possible before. Previously, that would have been entirely fragmented – I’d have had to think to go to the Disney Store, remember to buy the kids some t-shirts, work it all out myself. Now a company can proactively harness that, and do it in a way that’s specific to that fan. Not a generic “we have more products” push, but: you love the Mandalorian, so here’s something tailored to you.
And it extends beyond the purchase. Say you get to Disney World and there’s a problem with your park tickets, or you’re locked out of your hotel room. An AI agent that already knows you’re a Mandalorian superfan can respond accordingly – maybe it sorts you a lightning pass to the new Mandalorian ride as a goodwill gesture. A human agent could do that too, but even with access to the same data, it would take them significantly longer to join the dots. AI can do it instantly, and in a way that feels like it’s responding to who you actually are as a fan.
That, to me, is something we genuinely haven’t been able to deliver in the customer experience journey before.
Your background means you understand the web3 space as well. You’ve spent time thinking about very interesting technological spaces: blockchain, AI, transmedia, UGC. Do you see technologies intersecting in a way that creates opportunities for dev studios that weren’t possible before?
I genuinely believe the intersection of AI, UGC and web3 is the future of gaming.
Web3 specifically can be adopted to whatever degree different companies are comfortable with. But at the very least, we’re moving towards a world where not accepting stablecoin or crypto payments will be like not accepting Visa or Mastercard. You’re essentially saying you don’t want money.
And once you reach that point of accepting it as a currency, the other aspects of the technology follow naturally: IP provenance, universal public data, easier access to secondary trading opportunities. These stop being abstract concepts and start having real impact not just on game and media design, but on incentive alignment across the entire transmedia landscape in a way that hasn’t really existed before.
So if we take all those components together: in a world with AI coupled with UGC, anyone can create anything based on anything. That’s enormously exciting for creator-led economies, but also for individuals who suddenly have access to ways of creating and distributing content that simply weren’t available to them before. Add in web3’s incentive alignment, and everyone gets a fair cut: the IP originator, the creator, the licensor, all the way down the chain.
“What attracted me to Google is that this is already where companies across the globe have their data. And that puts Google in the strongest position to help game developers across the full constellation of challenges they’re facing right now”
Sam Barberie
Then you use a combination of AI and web3 to unify data and the consumer journey across as many fragmented silos as possible. AI is genuinely good at building one comprehensive picture of a user. But it can only work with the data it has access to. Web3 changes that equation, because that information can theoretically be accessed by any third party. Which means I can take my identity as a player, as a fan, to any application on the planet, and it can start learning specifically what I want and tailoring the experience accordingly, dynamically changing the items I see, the narrative I’m in, anything else.
So yes, AI, UGC, web3: I think that’s what I’d call the trifecta that’s going to unlock the next generation of gaming!
What’s the most practical thing a small indie studio should be doing with AI right now to stay ahead of the game?
The number one thing is to test an area. Iterate on assumptions as quickly as you can.
That’s a broad statement, and I think it’s true for playing around with marketing and concept art and actual game design and in-game dialogue and everything else! It’s a huge surface area for indie developers to start with different things. The most important thing is to just get your hands dirty and try it.
From there, you’re going to fine-tune and discover the ways in which AI works best for where you are in your development cycle. When you know the kind of game you’re making and the audience you’re trying to reach, at that point, you’re already ahead of the game, and you can leverage this technology in more and more sophisticated ways.
Further down the rabbit hole
Some useful news, views and links to keep you going until next time:
With GTA VI the most anticipated games release of the year, GamesIndustry.biz spoke with Take Two CEO Strauss Zelnick: “Technology is not going to take people’s jobs. The big tech companies who laid off thousands of people and said it was because of AI were not telling the truth.”
The Digital Futures Institute Festival of Storytelling will take place from 2–4 June in London. It features some free sessions about AI and game narratives, as well as wider conversations about the future of media (and the annual unveiling of the Arthur C Clarke Award shortlist).
Appcharge has launched an MCP server that enables mobile game publishers to configure their direct-to-consumer APIs directly from the development environment, using AI clients including Claude and Cursor. The company says it now processes over $1 billion in annual DTC transactions across over 150 games.
The AI-generated animated film Critterz missed its planned Cannes premiere after OpenAI shut down Sora, the video-generation platform at the heart of its production pipeline. The London/LA co-production is now hunting for a new tech partner, with Cannes 2027 the likely next window.
FirstLook has launched an AI agent that helps game studios analyse player sentiment, segment audiences, and deploy targeted campaigns across fragmented channels — from Discord and Steam to TikTok and Twitch — all via text prompts. Clients already include Team17, Krafton and Remedy Entertainment.
You can watch a neural net learn how to play Snake, with this in-browser Proximal Policy Optimisation training demo.
It’s oddly hypnotic watching a neural network play Snake over and over, with a PPO algorithm tweaking it after each game to gradually improve its decisions.
