SecretSauce Labs Exits Game Development to Focus on AI Brand Consistency

Wearemighty is leaving game development behind and rebranding as SecretSauce Labs, shifting its focus entirely to artificial intelligence content technology. The company spent three years developing tools to solve a production problem that originated inside games. Previously known as Mighty Bear Games before adopting the wearemighty name in January 2026, the studio developed titles including Disney Melee Mania, Mighty Action Heroes, and Butter Royale. The newly branded company now focuses on helping brands produce content at scale while maintaining consistency across advertising, social media, and product content.

For CEO and co-founder Simon Davis, the move stems from a problem that first emerged while the studio was still building games. In late 2022, Davis and co-founder Ben were working on a game that required around 25 million distinct, on-brand playable avatars. Creating that volume of content manually was unrealistic, and the early generative AI tools available at the time could not reliably produce consistent results. Davis stated that asking for the same character twice frequently resulted in a different face, a different outfit, or extra fingers. That experience drove the team to spend three years building the infrastructure needed to make AI-generated content usable in a real creative production pipeline.

Davis noted that generation itself is no longer the biggest challenge, with memory, judgment, and quality control becoming far more important. The studio learned that looking good and remaining on-brand are two separate challenges, and off-the-shelf tools were only solving the visual quality challenge. Maintaining brand consistency is difficult because storing brand material is straightforward, but checking every new generation against that material to catch errors before a human sees it remains largely unsolved. SecretSauce Labs built its infrastructure around generative models to determine what a request requires and which existing brand elements should be applied.

Addressing wider industry discussions around AI productivity, Davis believes that arguments regarding time savings can be misleading if they only consider the time taken to generate an asset. Generating ten assets and only being able to use two of them is not faster once editing and brand fixes are factored in. However, he noted that work previously requiring an art team of six people over three months can now be completed by one person in a day under the right circumstances, provided the output is usable immediately without extensive rework.

Davis also addressed concerns regarding creative control and workforce apprehension. He described much of the control currently marketed in generative AI game tooling as theatre, pointing to complex interfaces that add unnecessary steps between the user and the output. Furthermore, he acknowledged that non-technical barriers exist, such as creative teams worrying about their roles. He emphasized that the technology is meant for building team capability rather than cutting headcount, allowing staff to maintain workflows while utilizing new tools.