Engage Users Using AI Agents With Long-Term Memory and Situational Awareness
A case study by AurumNova
Background
The client is developing a AAA game and has a team of AI engineers (A few of them with PhDs) developing the next generation of AI game characters with emotional intelligence and awareness of self and surroundings.
The Challenge
Silos
Siloed AI engineers working on different AI features. The team tested many AI models with inconsistent outcomes.Irrelevance
New AI features launched by tech giants threatened to make own-developed code obsolete.
Copyleft
The product used a modified open-source library whose license required us to share our own developed code as open-source.
Our Results
- The AI product quickly implemented innovative new features.
- The development team became more effective and productive.
How We Did It
Utilize Shared Knowledge
Host weekly workshops for knowledge sharing among AI engineers to create a shared vision where everyone on the team can contribute.
Rally Behind Best-in-class AI
Harness AI state-of-the-art models by utilizing their latest features. Understand the direction of major technological advancements and develop features that align with these trends.
Maximize IP
Develop proprietary intellectual property by auditing existing code and gradually integrating new features into the business's IP. One such enhancement involved streaming LLM output to a top-tier text-to-speech service.eam LLM output to a best-in-class text-to-speech service.