All Collections
Customer Case Studies
The “Merge Park” Achieved a 240% Increase in Productivity Using Layer AI
The “Merge Park” Achieved a 240% Increase in Productivity Using Layer AI

Read how Games United was able to expedite the delivery of backgrounds and UI elements in their live title.

Updated over a week ago
Merge park and Layer AI

"Who are the team behind Merge Park?"

"Games United is an independent Digital Entertainment Content Factory with a particular emphasis on mobile gaming. Within the realm of Games United, universes are crafted, and narratives are unfolded through mobile games. The guiding principle here is that gaming is beyond gaming. Anchored by this philosophy, the emblematic motto, and the demonstrable data from their products, an investment was secured at a substantial valuation of 60 million dollars in January 2023. Many games have been produced over time; while some were discontinued, others have undergone iterative improvements. Presently, the Games United's "flagship" game is Merge Park, boasting an impressive Day 1 retention rate exceeding 40%."

Help Emma and her friends and bring the amusement park back to life.

"What made Games United choose Layer AI? What value did you see in using Layer AI?"

"Games United: Layer provides customized asset generation for game developers or artists. If you have a specific style or theme for a game, Layer can learn your specific style, enabling the creation of similar or compatible assets. This ensures that your game has a consistent aesthetic and user experience. We all picked our high-quality assets, grouped them per use case, and performed style training. Having variation in the inputs and uploading images respecting guidelines brought us the best results. In some cases, like terrain, we didn’t have enough input assets, and we used our limited set of 6 assets to produce more assets to feed them back to training. It’s a cool hack to multiply assets."

"Can you show examples of your style trainings?"

"We created independent styles for each use case. For example, we made different training on Layer AI for bags, currency, objects, and terrain:

It means that each style works perfectly for its specific use case."

"Can you tell us more about how you use these ‘styles’ to produce new production-ready assets?"

"Once our style is trained, we use the canvas to create variations out of our trained styles. We train multiple styles per different use cases in the same game. While AI training is a big word, the process is very straightforward once assets are formatted according to the guidelines. Here are some examples showing which use cases we used Layer for:

1 - Terrains

We prepare our references in order to make the preliminary preparation of contents such as road, railing or landscaping in our existing ground assets. Our artists drew 6 lands, 6 maps, and I also had some existing assets in the same style we trained one for terrains and created new map items within a week.

Terrains example inputs

Terrains example outputs

2 - Bags

Bags example inputs

Bags example outputs

3 - UI: Offers, Specials Offers, UI Asset, Panels

UI example inputs

UI example outputs

Using prompts like Wooden frame, nature details with flowers, sharp corner, soft colors, gold frame with column, we created all of the artwork and screen designs for our exclusive content, in-game offers and shop screens:

4 - Components

Components example outputs

Created using the following prompts: Wood chest, iron and diamond chest, gold and wood chest, royal chest, iron chest.

Created using the following prompts: Bronze shield icon, Gold shield icon, Silver shield icon, Diamond shield icon, Platinum shield icon.

Example in use:

During the bulk asset creation stage, Layer streamlines the process by generating numerous assets that adhere to the established style guide. This is particularly useful in large-scale projects that require a high volume of stylistically consistent assets."

"What would be your advice for new studios exploring AI?"

"We have a few thoughts on that:

  • Game Artists require minimal new skills to integrate Layer into their workflow. You can find efficiency if you invest a tiny bit of time learning to use the tool, and the team is very willing to support you.

  • AI training is something everybody can perform in the team; actually, anyone can use Layer and make ‘styles’ ready for the team. However, if input images passed the Art director's approval, it would be great as they are critical for your success.

  • Train a few styles for different use cases you have; focused training is best for output accuracy.

In our case, we use artist time actually to use the styles and polish the assets to deliver. We did some math and realized that the team's productivity was 240% of what it was before, Layer has become a backbone of our demanding art pipeline."

"For a company like ours, which specializes in Merge-3 games such as Merge Park, Layer has significantly expanded our art libraries and production capabilities.”


Closing thoughts from Layer AI to help you get started:

  • Training multiple models in Layer words “styles" was key in getting the best quality outputs required when making a game. Check out this article.

  • Games United used Layer AI for both concepts and production assets

  • All monetization offers assets (backgrounds, chests and UI elements) in Merge Park were created with Layer.

Did this answer your question?