JetPlay wants to offer you the game ideas machine.
That’s the promising prospect from the startup behind Ludo AI – the world’s first true AI platform for games concepting and ideation.
Ludo promises to make the games concepting process simple, by allowing the user to search by keywords through its database of almost one million games, to deliver a brand new game concept – complete with text description, images and comparisons of similar games.
Ludo, Latin for ‘I play,’ uses machine learning and natural language processing to deliver a series of brand new game ideas each time. The platform’s capabilities are within the reach of studios of any size, with Ludo arguably being most useful for the quick turnaround of the hyper-casual space.
To find out a little more about Ludo, and its potential for the games industry, we sat down with Tom Pigott, CEO of JetPlay
Jetplay has a history of game development itself, with a background in creating games in the virtual reality space, and more recently has experience with hypercasual titles.
“A commonly used phrase is ‘necessity is the mother of invention’,” says Pigott. “And as a hypercasual studio, we’re in a very competitive segment, right? There’s this constant need to come up with new game concepts to test, particularly in that space. And most of them really don’t make it past the initial metrics.
“And so, as we looked at it, in this very crowded field with many, many competing studios worldwide… Was there an opportunity to maybe have a tool, given some advances that were going on in machine learning, that could aid us with sorting game concepts?”
The platform’s library of almost one million games definitely skews heavily towards the hypercasual market – mostly because, as Pigott points out, there’s just more of those out on the market.
So how does Ludo work, from a user perspective?
“Ludo is built upon open source machine learning models,” says Pigott. “It has eight million web pages, that is really its base, in terms of vocabulary. And then what we did was to focus it on games and gaming. And so we have this enormous library of games, most of which are skewed towards, hyper-casual indie style games, because obviously, there’s just a lot more of those.
“So as a game creator, you go to Ludo, and maybe you have a thought of a concept in your head that you want to explore more. You can enter in just keywords: it might be a mechanic like swerving or tapping or stacking, or it might be, ‘hey, I want to focus on a zombie game.’
“You can type as little as you want into Ludo, and what happens is that the machine learning is based on natural language processing. And it comes back to you with a complete concept along with accompanying game images of some similar type of games.
“The exciting thing about it is that it doesn’t just do that once. You can do it scores and scores of times, and it’ll be different every time. So that’s where you get the advantage – it’s pretty difficult as a small team or an individual to come up with constant new sources of ideas. So this gives you a great tool, it turbocharges your creativity for what we call the whole ‘game-storming’ process.”
Ludo’s enormous library of titles to pull ideas from is certainly impressive, but putting all that together was far from the most difficult part of creating the platform, as Pigott explains.
“The actual building of the game database is not is not the difficult part. The challenge is, once you have the data – it doesn’t matter if it’s 100,000, 500,000 or a million games – it’s what you do with those games. So what we have spent the past year on, is training that data. And that’s where the machine learning comes in. So you’re weeding out the non-relevant terms or descriptions, so that you’re ideally getting more and more relevant types of output.”
Any sort of AI automation, in any business, often comes with the expectation that it’s going to be a ‘job killer.’ How often these technological advancements actually impact jobs can vary from sector to sector, but Pigott is keen to stress that this is not the case here.
Ludo’s main selling point is not that it seeks to replace anyone, but that it’s a tool intended to make the game development cycle easier.
“This is absolutely not a job killer,” says Pigott. “This is honestly what we consider a job enhancer. Whether you’re an individual developer, or working for small studios as a game creator, you’ve got this great resource that you can go to from the beginning and come up with a new concept.
“What this doesn’t do, to be clear, is build the game for you – you still have to develop the game, but it jumpstarts your whole creative game-storming process. And that’s what we feel is a huge value add, because we have lots and lots of studios testing Ludo now. And the constant issue that we face is ‘okay, we have to come up with a game idea.’
“It’s a hits driven business, particularly in the hyper casual side of things. So even if you have a successful game that gets to the top of the charts, typically it won’t last there that long. And so you have to come up with the next concept and the next concept. We feel that Ludo really meets that type of demand in terms of helping creators create better concepts, and certainly more concepts at scale.”
Still, with a machine learning platform searching through a library of existing games, is there not a risk of it delivering derivative ideas, games that chase a trend rather than create a new one?
“That’s very possible,” admits Pigott, “but that’s where you keep iterating, so that you can ultimately create something that is unique. There’s another feature that we have and we call it the ‘game blender,’ and effectively it allows you to take existing games and blend them together.
“For instance, you could take something like Among Us, which is super popular, and then you could add two totally different titles from the trending charts, and have Ludo blend all those concepts together. And maybe you’ll get a mechanic from one, a description of the other and images of the third. And that’s a way where it becomes non-derivative that you can create something kind of interesting. You can take all that and put it into the concept that you’ve already been working on. There’s a lot of ways you can make it unique.
“But needless to say, with the millions of games out there, it’s pretty difficult to have a completely original new concept. Someone, somewhere out there, has already thought of something similar to your idea. And that’s alright.”
He’s not wrong there – after all, game journalists are often notorious for likening new titles to familiar touchstones. ‘It’s the Dark Souls of dating sims! It’s the Super Mario of simulator games!’ Hell, this month’s When We Made deals with a game that could definitely have been born out of a game blender process – a mashup of Breath of the Wild and Animal Crossing, and the game feels no less unique because of that.
All games, no matter how unique, pull ideas from somewhere. Developers are always inspiring one another to find new ways of expressing their creativity. Ludo, Pigott argues, just provides a helping hand – reminding developers of titles they might well have forgotten about.
“People have been in the industry for decades, and it’s tough to have all of that in your head. I think what’s been interesting about so many games that are in there, there’s games from the 90s and the early 2000s, games you forgot about. And that’s what a lot of it is about, it’s being inspired by the past hits and making a modern version of it.”
THE AI FUTURE
Of course, machine learning is being used across the industry already. And Pigott believes that these AI development tools are only going to become a
larger force in not just our industry, but across the creative industries.
“We just fundamentally believe that AI development tools are going to become part of the industry. I mean, there’s already some here, there’s a lot of AI tools in the back end in terms of monetisation, optimisation, things like that.
“But I think that creative tools that use machine learning are going to come to many, many industries. I mean, look at the music industry – just think of how many songs you have to think of and create. Or it could be in advertising, or movies – you’re just taking the particular database from that industry, and you’re applying it to the AI platform. And so I’m sure that’ll be the case down the road.”
PINNING IT UP
Ludo launched in open beta in January, with developers already getting to grips with the platform. So how has the feedback been so far?
“It’s been great,” says Pigott. “We’ve been very, very happy, both in how many individuals and studios have signed up, and in the feedback we’ve been getting.
“People are finding that they can use it in a collaborative sense. Because when you’re game-storming, it’s not always an individual effort, it’s a collaborative effort. And this lets you do that within the Ludo environment.
“One thing that’s actually become a really well used feature, that we maybe didn’t expect, is that we have an image database of close to two million game images. And the feedback we’re getting from a lot of users is that for artists, it’s this huge resource to get inspired by if you’re looking for a type of game. People have said, ‘this is like Pinterest for game designers,’ because it’s such a massive resource. You can just keep picking your favourites and building mood boards from that.”
The JetPlay team is currently at work on feature requests from their users – such as integrating the platform with other creative tools. Those interested in finding out more can sign up to the beta at askludo.ai.
“We’re still very much in beta,” says Pigott. “So there’s going to be a lot of great improvements and changes by the time we release commercially. And we hope this becomes a real game-storming tool for small studios, and helps them to create better game concepts. That’s really all it is – it’s helping turbocharge that creative process. So we’re very excited about it, and look forward to getting it released commercially.”