AI: Under Construction
There’s a time in the journey of every successful business when it needs to embrace the AI revolution in order to continue to grow, or stagnate and die out.
Gaming is still on that wild frontier, where you can start building a game out of your garage, invest yourself in a wild idea and reach success. It’s a classic story of an underdog reaching for the stars, that is in good part a reason why so many flock to gaming, hoping to hit that golden vein.
But, what happens once you make a successful game? Bliss? A happy end? Nope, that’s when you need to build a successful business. And grow it. And keep growing. And to make that possible, at one point you need to embrace the AI revolution. That’s where Two Desperados is. And here I am, too, to help.
More AI Hype
As a matter of fact, I am quite conservative when it comes to AI, so it’s not my intention to add to the hype of AI (if it’s even possible). However, there is something profound in the AI transformation that I want to talk about.
When Data Science appeared, it opened the door for something truly novel, and, no, I don’t mean any of the algorithms. Data Science introduced scientific methodology to business, a new way of thinking. Formulate a hypothesis. Build data collection. Run an experiment. Learn. Repeat. Improve.
That’s the true power of Data Science. AI is the tip of the iceberg, so to speak. Of course, it’s easy to counter this, just find a business that introduced X (some fancy AI solution) and bam! Their revenue went up. Although true, and possible, many attempts in this approach end in disaster. More so than in success. Our first attempt was difficult here as well. Dissatisfying. Any result that came back felt overworked and underwhelming.
AI is the People
And here I stand, after 10 years in the business and even more learning about and working on this tech, here I stand and say: AI is the people. It’s a way of thinking. When a company truly goes through AI transformation it’s not because it implemented X (insert fancy AI solution), it’s because the people adopted another perspective. That’s the true magic, tech is easy. Really. Well, mostly. Real magic comes when everyone in the organization gains a new lens, a new window into the world.
AI at Two Desperados
We are now 4 months in on a project that revolutionizes the organization from the ground up. Our first task was to allow the organization to own its data. You own the data when you are accountable for its quality and when you can build upon it. As I said, tech is easy. There are many tools out there. We built the pipeline on GCP using open source Snowplow. It took us only 3 months to set up the proper data pipeline, from event triggers to the tables with ready for use crunched data. We have three great Data Engineers working on that, setting up everything, owning the quality.
But, AI is people and quality is not only in the hands of Data Engineers. Same as for any fancy machine learning: garbage in-garbage out, if garbage gets into the pipeline, there’s no magic that will turn it into the perfect data you need to build really useful stuff. Developers need to set up triggers and the product team needs to learn how the data they want to track fits into the big picture. How the data will help measure, validate, hypothesise, learn or automate through AI. The product team needs to adopt this way of thinking to even do something as trivial as setting up new triggers for data tracking in a game. Once the perspective changes, it impacts everything, even the way people think about problems. We are not there yet, but we have taken important steps.
Fight on Two Fronts
Like for any battle, good allies increase chances of success. It’s probably near impossible to implement such profound changes in an organization without the support of company leadership. That’s one, huge problem that we won’t have on our journey. But, the journey can be long and it’s not always easy to see the value at the end of the road. That’s why we are fighting on two fronts, trying to straighten the path and set up the right foundations for the future, while also trying to implement that fancy AI solution. It’s hard, and it would be so much easier if all the prior work was already done.
We’ll start the new year with that shiny AI solution that will adapt our game(s) to ensure a better user experience. It’s a huge boost to be able to showcase the benefits, but it’s all for the benefit of the long(er) game. My initial estimate was that we would need at least 9 months to get to the point where we are now, and here we are. Times are turbulent, but in the year 2022 one thing is for sure—a focused organization can go through the AI transformation quicker than ever before.
Marko Jevremović Lead Data Scientist