NPCx is a company building the next generation of Artificial Intelligence (AI) for Non-Player Characters (NPC) in Interactive Entertainment, including Video Games, Augmented Reality, and Virtual Reality. Pulse 2.0 interviewed NPCx CEO and co-founder Cameron Madani to learn more.
Cameron Madani’s Background
Madani started in a traditional financial services route, with his first job working at Merrill Lynch. Madani said:
“I was an institutional manager for one of their products and later worked overseas for a few years in investor relations. And that ties in because I hit a roadblock a few years later, being unhappy in my career. At that time, a good friend of mine was attending the University of Southern California MBA program, and he encouraged me to pursue an MBA as well, it was an opportunity to redesign myself and rediscover my passion, and that got me thinking: I really like video games!”
“I also liked the idea of business and entrepreneurship, so I enrolled in the accelerated full-time MBA at Pepperdine, taking it as an opportunity for a career sabbatical. During my program, I told the faculty that if a company wants to have a business plan written, please let me know. Luckily, somebody came in from a video game development company, and I wrote him a business plan. About a year after I graduated from my MBA program, in 2008, the company contacted me and said, ‘We’re not going to do the idea you came up with, but we want you to be our corporate development officer.’”
“So, I learned the video game business and worked under a gentleman who had worked at Warner Brothers Interactive. He had extensive experience in licensing and intellectual property (IP). So, for several years, I learned the entire product lifecycle for video game development, everything from licensing IP, development, art and animation, publishing and distribution. I call this period my second MBA.”
“Shortly after that, I launched my first company, and we developed a successful game for Microsoft Xbox 360, ‘Torchlight’ in 2011. This was my first opportunity as an entrepreneur, and I had some failures and successes, but I ended up pivoting into consulting for about two years afterward. However, in 2014, I started another company called ‘Motion Burner’ which specialized in motion capture and 3D animation for video games, film, and virtual reality (VR). It had some outstanding achievements, and we worked on over 30 titles with 12 companies. In fact, two VR products we worked on went on to win awards; one was an Academy Award for Special Achievement called ‘Carne Y Arena in 2017. The second was an Emmy for ‘Wolves in the Wall,’ in 2019.”
“As I was doing this, I always knew that animation and motion capture were manually intensive and could be tedious and time-consuming. So, around 2018, I started thinking about AI and machine learning (ML); I didn’t know anything about it and started asking all my peers, ‘Do you know any data scientists?’ Then I met Michael Puscar (a successful software engineer and entrepreneur, who later became a co-founder and chief technology officer (CTO) at NPCx). Michael and I discussed use cases, and we devised a hypothesis leveraging Michael’s AI and entrepreneurial expertise. We got our first customer, and we launched NPCx in early 2020. Soon after, we added another partner, Alberto Menache, a well-known pioneer in animation and motion capture, with decades of experience working on major video games and motion picture titles, as our chief product officer (CPO).”
Formation Of NPCx
How did the idea for the company come together? Madani shared:
“The idea came about after meeting with Micheal and establishing our hypothesis, as the manually intensive aspect of motion capture is an industry-wide problem—it’s not creative at all and burns out creative people quickly. We named the company NPCx (with NPC standing for non-player character).”
“And while there’s quite a bit of fear and disruption around AI in general, in my opinion, it’s allowing storytellers to be more creative, and you’ll always need an animator and animation director to be involved in any AI creation process. So, the genesis of NPCx was our goal to build tools that will help animation pipelines to be more efficient, less costly, and allow the turnaround time to be substantially reduced.”
Favorite Memory
What has been Madani’s favorite memory working for the company so far? Madani reflected:
“A favorite memory of mine was receiving our first client in late 2020 because that proves your concept, as you don’t have a company until you have somebody willing to pay for what you’re doing.”
“The second is getting the investment from Kakao Investment Corp (Kakao)— a lead investor in our seed round, which we’re still currently in. We’re raising $3 million, and Kakao put in the first $1.5 million. It was a massive endorsement from the investment community, especially as currently it’s a very challenging time for raising venture capital, as opposed to two years ago when it was much more accessible. And for us, it was an honor. Kakao is a vast, multi-billion dollar company in Korea, and they looked at our team, track record, and vision and decided to make a bet on us. So that was definitely a champagne cork-popping moment for us.”
Challenges Faced
What challenges have you faced in building the company? Madani acknowledged:
“So, in a nutshell, the main challenges have been raising capital, finding talent, and building minimal viable products (MVPs) that will benefit our clients and the greater marketplace.”
“Developing MVPs has been an extreme challenge. Our first commercial product, TrackerX, can accomplish what a manual team of motion capture cleaners can do about 50 times faster. We had a hypothesis of training neural networks to achieve what humans would do, but at this point, we didn’t have funding from Kakao. So, the founders put money in, and then we did a pre-seed Crowd SAFE raise and received over $540,000 with Republic. We used these funds, along with revenue earned in 2020 to 2021, to cover the development of this challenging problem of developing a totally original and novel technology. And it’s taken nearly two years and at least $800,000 in research and development to finally create a product TackerX. So, building something from nothing is quite challenging.”
“Furthermore, because the customer base that we’re going after is the same as Alberto and I have been working with for the last 12 or 15 years, we’ve been very cautious about not burning our relationships by over-promising and under-delivering. We want to make sure that when we go back to the same customers that we’ve previously worked with, we can assure them that we’re providing a better and less expensive solution.”
Core Products
What are the company’s core products and features? Madani explained:
“NPCx primarily focuses on two innovative aspects of video gaming: character movement and character intelligence. For character movement, we’re creating a suite of products that vastly reduces the time and cost of building and deploying motion capture and key-framed animations in video games, film, extended reality (XR), and the metaverse. Our first product, TrackerX, launched in March of 2023 and utilizes neural networks and biomechanical models to substantially reduce the processing time of motion capture data, which was done manually until now.”
“Traditionally, motion capture performances rely on optical or sensor-based hardware systems requiring painstaking manual “cleaning” to prepare them for the final product. This process, done with a mouse and keyboard, corrects issues such as feet going through the floor mats or limbs penetrating other characters and objects. TrackerX transforms this by combining biomechanical modeling and neural networks to automate this cleaning process. Currently, TrackerX speeds up manual cleaning by nearly 50 times, with ongoing neural network training for sustained improvements. This efficiency saves time and money for studios and enables more extensive motion capture content creation within the same budget.”
“While TrackerX is our main feature and product, we have more exciting developments in the pipeline. Our next release will likely be RetargetX, which focuses on retargeting within the motion capture process. When you retarget, you have to take that skeleton that was established from the motion capture cleaning and put it on characters with different geometries, which is currently done manually. And there’s been no solution we’ve seen from big companies and studios to automate this. At the moment, no one’s been able to crack this other technically challenging and costly issue. So, with TrackerX and RetargetX, we’re trying to capture the entire motion capture pipeline from the motion capture stage into the game or film.”
“Next, we’re going to do a similar system specializing in facial motion capture that we call FacialX. So, we have RetargetX and FacialX in development. And for character intelligence, NPCx is developing a novel way to model human behavior. We are specifically NOT creating god-like AI NPCs, like most of our competitors. Instead, we aim to virtually ‘clone’ players and place them into video games, XR, and the metaverse to such an extent that distinguishing between an NPC and a human player becomes nearly impossible.”
Evolution Of NPCx’s Technology
How has the company’s technology evolved since launching? Madani noted:
“Being able to pivot has been essential as an entrepreneur—you can’t be married to any idea. So, to start with, TrackerX was only going to be a neural network-based system, but that’s been quite challenging. Along the way, one of our developers said, ‘Why don’t we create a biomechanical model as a first step?’ So, we decided to launch TrackerX with the biomechanical model, which is still a fantastic product, significantly improving efficiency, cost, and time.”
“This decision has given us more time to continue to develop the neural networks. So, within this industry, technology is constantly evolving, and being able to pivot with technological decisions is a must.”
Significant Milestones
What have been some of the company’s most significant milestones? Madani cited:
“So, this is probably the same as my favorite memories, but the success with raising funds, especially from Kakao. Next, gaining our first paying client was very significant, and then deciding to launch TrackerX with the biomechanical model was also noteworthy.”
Customer Success Stories
After asking Madani about customer success stories, he highlighted:
“This has been a bit of a challenge because we’re limited in sharing about our client projects; many of our clients have titles that haven’t even been announced, let alone released—so we’re not at liberty to say—but hopefully, we can share more on this soon after these titles are released.”
Funding
After asking Madani about the company’s funding, Madani revealed:
“To date, we have raised over $2 million in equity, with approximately $540,000 through Republic.com’s Crowd SAFE, which closed in November 2022, and an additional $1.5 million in a Y Combinator style SAFE, which closed earlier this summer. We are still in the middle of our seed round and want to raise an additional $1.5 million to complete this round.”
“We’re currently in discussions with some other investment companies, mostly gaming funds, and we’re confident we’ll be able to close the remaining $1.5 million by early 2024.”
Total Addressable Market
What total addressable market (TAM) size is the company pursuing? Madani assessed:
“So, by using my knowledge from my previous company, Motion Burner, and using publicly available data, the total addressable market for both TrackerX and RetargetX is easily over $350 million worldwide annually. Plus, with the development and release of FacialX, we’d be looking at another $150 to 200 million TAM.”
“One common entrepreneurial mistake when making a business plan is that you should never give a percentage of how much of the TAM you’re aiming to target. To get the correct numbers, you need to extrapolate based on each client and how much they spend per title and how many titles are released worldwide per annum. By way of example, an average AAA video game title spends roughly anywhere from $250,000 to $1 million on motion capture processing, and live-action animation films spend easily three to four times this amount per title, some even as much as $10 million.”
“But, if we can be the first to market with Retarget X, as we are with TrackerX, we could easily get one-quarter to one-third of the market. And, of course, the more products we develop, the larger the TAM we’ll pursue.”
Differentiation From The Competition
What differentiates the company from its competition? Madani affirmed:
“All the partners at NPCx are Gen Xs, so we’ve had a lot of experience with both successes and failures. Many young and ambitious entrepreneurs are coming out of university or quitting school and starting companies with unique ideas. But I would argue that a massive advantage of our founding team is our combined understanding of the industry. Alberto and I have over 45 years of combined experience in video games, film, and visual effects, and we’ve learned many lessons over these years. Our clients can trust that we’re not here to disrupt the creative process but to improve and enhance it, much like the combustion engine did to the transportation industry that was dependent for centuries on the horse and buggy. This development shifted jobs and skill sets and expanded the personal transportation industry well into the 20th century. We believe we are in the same type of transformation for digital storytellers.”
Future Company Goals
What are some of the company’s future company goals? Madani pointed out:
Firstly, we plan on closing our current seed round of $3 million over the next 6 months.
“Another one of our plans is to continue developing BehaviorX—our big ‘billion dollar’ idea. And I’m not sharing anything confidential, as we’ve shared this publicly before. Ultimately, this is what made Kakao interested. With BehaviorX, we are creating non-player characters that act and “think” like humans, that make mistakes, and do surprising and exciting things such as we do as a species. We believe that this is what players truly want, as opposed to god-like, super NPCs, with 99.9% lethality and precision the rest of the industry is pursuing.”
“To further this last point, most companies are developing super impressive and high-achieving NPCs for video games that can accomplish every single task and kill every human player. But we don’t think that’s very exciting. Rather, we want to build NPCs that act just like humans, and we are doing this by observing how human players actually play. By watching and analyzing human players’ decisions based on environmental inputs, such as what weapons you choose, are you stealthy or bold? How do you act with low armor and weapons? How long do you stay on a level, and so forth?”
“BehaviorX can then observe and predict how a human player would respond in certain situations. With this data, we can create a clone of any human player that can be used to play offline, whether as a single player or online in co-op mode. But the big idea is the monetization model. Human players who are elite at certain games, such as the top 500 players in the latest multiplayer first-person shooter games, could opt-in to create clones of themselves with BehaviorX technology. Then, they could rent out their BehaviorX clones to other players to use on co-op missions or to practice against, allowing super players to monetize their skills and talents, undoubtedly making their parents very happy with their career choice.”
“As far as rough dates for our future releases. Ideally, RetargetX Beta will be launched in Q2 of 2024, followed by FacialX in Q4 of 2024. Then, we’ll have our MVP for BehaviorX, which should be ready in early 2025.”
Additional Thoughts
Any other topics to discuss? Madani concluded:
“I think the stars have aligned for NPCx, with the maturity of the management, knowing the right people, understanding what the customer needs, and the excitement around our utilization of AI as an accelerator. We’re in a unique time, and in my opinion, it’s our game to lose at this point.”
Introduction
AAA games typically take 3-5 years to develop, with character animation significantly impacting the timeline. The manually intensive aspect of motion capture in character animation in gaming is an industry-wide problem—it’s not creative at all and burns out people at studios quickly. Traditionally, motion capture performances rely on optical or sensor-based hardware systems requiring painstaking manual data “cleaning” to prepare them for the final product. This process, done with a mouse and keyboard, corrects issues such as feet going through the floor mats or limbs penetrating other characters and objects.”
This article would explore how leveraging AI for character creation can cut the ‘cleaning’ process from days to minutes, which will allow AAA games to be finished in record times. Diving Deeper into Neural Networks and Biomechanical Models and explaining how this technology works and how it can be deployed by studios to make their animation pipelines and processes more efficient and cost-effective. Lastly, it will reflect on the potential of AI in the gaming industry to bring a new era of creativity and efficiency from creating animations faster to crafting intelligent and hyper- realistic characters.”
Fully autonomous and realistic NPCs: Using AI to craft intelligent and realistic NPCs to populate virtual worlds – from baristas to pedestrians, cashiers, and even friends.
The future of Gaming Realism: How AI will blur the distinction between reality and virtual gaming experiences.
