Interview: Mike Walsh

Keen to hear more about AI middleware giant xaitment’s plans for the future, and its vision for the evolution of AI, Will Freeman caught up with the outfit’s new CEO and managing director Mike Walsh.

You’ve relatively recently joined xaitment.
I’d left Havok in September or so in 2010. I am a big fan of the games industry community, and I’ve worked in and around it since 1995, and I think that AI is sort of the next biggest thing.

In the early part of the current generation of consoles, and in the early part of the development of things like the Xbox 360 and PS3, people were really after graphics and physics. The past five years have been about full graphics engines and lots of work on physics, but I don’t think people have given enough thought to pathfinding.

I really think that AI is the next thing that needs to be addressed. I guess I’ve said this a number of times; you’ve got these Xbox 360 and PS3 games that you play on your 60-inch plasma TV, and you’ve got tremendous surround sound, and the graphics look amazing, with everything running in HD, and the physics are amazing, and you’ve still got stupid characters.

I really think that tackling that is the next greatest thing, and I think people need to think not just about getting a character from point-A to point-B, but about how to make them more intelligent and get them to stop doing things like running into barrels or walls.

So I’d always thought AI was the next big thing to conquer, I looked around and talked to several companies, and it just so happened that xaitment was looking for someone to help with their business plan. They have great talent here and great engineers, and not a lot of practical business experience.

The company was founded out of Saarland University, and it just seemed like it was a really perfect match, where I have all this experience in sales and putting together contacts and so forth, and they just didn’t have anybody like that in the company. So they brought me on to help expand their reach on all of their regions, whether it’s Europe or China, or Korea or North America. They’ve just asked me to help them improve sales all over the place.

So are you noticing that AI is at least starting to get the attention – or investment – in the wider industry it deserves?
I think it seems to be on the forefront of all the games developers’ minds. They’re buying engines – whether its Unity or Unreal or Gamebryo or Vision – and they’re saying that the pathfinding is OK, but asking if there’s anything they can do to improve upon that. So many of those customers are coming to us asking ‘what can you do to help us?’. They want to push AI in different directions, and need additional things provided to them that their current solutions don’t provide.

So this is a time of opportunities aplenty?
Opportunities are aplenty, and we’re certainly talking to a lot of customers, and making positive moves in the right directions, always selling new licenses and getting new customers we haven’t really talked to before. I do think that it’s a very exciting time for the industry, and specifically around AI middleware. And that means that there’s a lot of room for improvement in that area.

If we’re seeing this change in the AI specialty, how is xaitment changing as a company?
First and foremost, we’ve been in business for five years, and we’re on the third version of our product which we’ve been developing over those last five years. We’ve got a very rich feature set and a very stable product, and it works across four different platforms, and now we’re just out there trying to increase the number of customers we have.

We’ve got over 30 customers worldwide, Tencent being our largest customer, and they are really growing. With offices in China and the US they are really starting to attack that massively multiplayer online section of the market. And there’s also Shanda and Kingsoft and after giants in China that are doing the same thing. They’re attacking and really they are changing it.

So, for xaitment it’s about two things. It’s about getting more customers, and it’s about looking at what else we can do in AI. Right now it’s about pathfinding, and we’re tackling character behaviours and game logic with our xaitControl project, and we’re working on a product that we’re going to announce at GDC 2012. That product is going to be about making characters more intelligent in games and giving developers the tools and the runtime to create more intelligent characters.

So your evolving the present concept of AI? How far can that go and what do you have planned?
Let me be a little bit bold here. The sky is really the limit. With AI right now people are interested in pathfinding, but I think there are so many other applications for AI.

What we’re talking about is something that, to my knowledge nobody else in the industry is doing, and that’s character behaviours.

We’re also approaching from a tools perspective. There’s the old paradigm in games development is that designers want something and then engineers have to code it. We’ve taken that out of the loop and put tools in the hands of designers. We’re offering the tools they need to design the game relative to AI, so they don’t need to wait up in the engineering team.

That’s a big part of it; really giving designer the ability to do it. So they don’t have to code in C++ or even write scripts. Our product is really about user interface and allowing designers to do what they want with the AI in the game. That’s one thing. The second thing for us is about making those characters really intelligent. If you’re giving a character a sense of what’s in a room, for example, and a database of information about the contents, and giving the characters a logic tree that lets them make decisions about how they might use the objects in that room in something like a fire fight, then its about far more than just going to point-A to point-B. That’s what we’re working on right now.

In terms of the future, there’s lots of different stuff to do with AI, like the potential of facial recognition. There’s socialisation things where characters react with AI relative to society such as groups or packs. There seems to be so much there can be done with AI to make games more challenging.

So is this about democratising AI technology within a development team?
I think that’s a fair statement. The we’ve developed this is by sitting down with a whole bunch of companies and talking to them about what they want form an AI solution.

We not only talked with engineers; we also talked with a lot of designers, and the bottom line was that engineering tends to be the bottleneck in game development. Now, I don’t mean that in a bad way, because the engineers just have so many things that they are responsible for.

So what we heard over and over again was that there was a need for a tool that would enable designers so that they didn’t have to be dependent on C++ or the scripting skills that most of them don’t have. It’s been about taking some of the burden off of engineers and spreading it around to designers so that there’s the option to create efficiencies and spur creativity.

And what of the challenges in the AI sector that xaitment are having to address?
Firstly, we’re kind of a first mover in terms of character behaviour stuff, and being a fist mover is always a little bit difficult. I would say that about 60-to-70 per cent of the time, people really get it, and understand where we are taking not just our tools, but the way we are making characters a little bit more intelligent. That’s the first thing.

The other thing is, that with the rise of mobile, social and casual games, the industry doesn’t really have AI yet down at that level that could be helpful. The question is ‘how do we get there?’. We’re spending a lot of time trying to figure out how do we improve AI in those games; let alone your typical triple-A and double-A titles?