Enterprise mobile apps operate under demanding constraints that includes highly customizable data models, dynamic UIs, and low bandwidth environments. At Salesforce, building and scaling our mobile app required confronting these performance bottlenecks head-on. In this session, we’ll explore how we reinterpreted "offline-first" not as a fallback mode but as a performance accelerator.

Starting with a “no latency is better than low latency” mindset, we implemented intelligent priming: a system that proactively prefetches and caches the most pertinent data before the user needs it, hence lowering perceived latency to near-zero. This strategy enabled seamless offline support, but more importantly, delivered significant performance gains even when online.

I’ll cover our evolution from simple MRU (Most Recently Used) heuristics to sophisticated, machine-learning-based predictors that anticipate user behavior and pre-load important data. You’ll hear about the architectural decisions, instrumentation choices, experimentation with cache hit optimization, and how we adopted the now patented predictive caching mechanism. Expect practical insights on building smarter caches, facilitating seamless offline-first workflows, and creating enterprise apps that feel fast, no matter the network conditions.