Personal Assistants

Personal AI Assistants That Actually Remember

Open Context Layer can enable personal AI assistants that actually remember. Today’s assistants forget your goals, tone, and preferences between sessions. If a user creates an “assistant” context using the OCL, this assistant can remember things like:

• User’s communication tone (formal, casual)

• Preferred language, humor style, etc.

• Preferred tools (Notion, Asana, Google Docs)

• Ongoing tasks and goals

As they move from mobile to desktop or from one LLM (e.g., Claude) to another (e.g., GPT), the assistant retains memory, without repeating instructions. This opens a whole new class of Personal Assistants that can navigate the internet based on your intents while knowing deeply what exactly it is you need.

Converting Existing Agents Into Personalized Assistants

With OCL’s user-owned contexts, the user can fine tune their agents to fit their specific needs. They can do it by:

• Running private LLMs or fine-tuning local agents using custom contexts powered by OCL.

• Moving between agents with their own context to select the best agent for the job e.g. Gemini for research, Claude for code, ChatGPT for writing etc.

Hyper-Personalized E-commerce Experience

E-commerce recommendations are shallow, based only on recent clicks. But your context can include:

• Your style preferences, brands you avoid, ethical considerations (e.g., no leather), budget limits

• Saved preferences and prior purchases across different shops

When you visit a new store or use an AI shopping agent, it tailors results instantly even if you’ve never used that site before reducing cold starts.

Recommendation Engines for Entertainment

Users can create contexts for their specific entertainment taste e.g. artists they love, music genres they like, shows they watch, etc. These contexts can be used for asking your Personal Assistant about finding you recommended content from the internet or can be input into external apps and platforms to show you content relevant to your specific taste.