Welcome to the Personal AI product newsletter, where the future of personal digital experiences unfolds before you. As we step into a world where technology intimately understands you, our commitment is to make every interaction more insightful, efficient, and uniquely tailored to your tasks.
Personas are a powerful tool for understanding and catering to the diverse needs of your audience. At Personal AI, we leverage personas to create tailored AI experiences that seamlessly integrate into the unique contexts of our users’ lives. Whether it’s a lawyer seeking to streamline legal document analysis, a financial advisor aiming to provide personal guidance, or a family member looking to automate daily tasks, our multi-persona model ensures each individual receives a customized AI companion trained specifically for their role and requirements.
The core value proposition is creating AI companions finely tuned to each user’s unique context, persona, and requirements. This allows us to deliver hyper-personal and productive AI experiences across diverse professional and personal use cases.
Personas in the Personal AI system are designed to represent different aspects or roles of an individual. Each persona has its own identity, directives, and memory bank tailored to that specific persona.
For example, an individual could have separate personas for their professional role (e.g., CEO), personal life (e.g., family member), and specific interests (e.g., legal or financial advisor). Each persona would have its own tailored knowledge base and memory, while still being part of the overall multi-persona AI system.
This approach allows for a more organized and contextual presentation of information, as well as better data privacy and control over what information is shared with different personas.
The primary AI, which you will see at the top, is the core model that serves as the foundation for all personas. It is trained on all of the data you upload into your AI. Each persona underneath the primary AI, on the other hand, is a specialized version of the primary AI tailored for specific roles, contexts, or audiences. These personas inherit the foundational knowledge from the primary AI but have unique traits, language styles, and specialized knowledge relevant to their intended use.
If you are using multiple personas, we recommend training only those AI personas specifically instead of training the Primary AI. Just remember, when training your AI Personas, your Primary AI will obtain all of the same knowledge.
By clicking on the persona, you will see a gear icon pop up next to the persona’s name. You can navigate to the “Train” tab and select the “Memory Upload” option to upload multiple files at a time.
When inside the “Documents” section underneath the memory stack, you will see the dropdown menu for each of your documents. You can select which persona you want to assign the document to by checking the boxes.
When uploading documents, audio, or URL links, you will have the option to assign which persona you want the data to go to.
Using integrations is a fast and efficient way to upload many documents to a given persona.
Use this icon next to the message bar inside the chat page for one of your personas to access integrations.
Our upload button inside the document editor will also give you access to integrations.
After finishing the first step of the automation with your trigger, you can set up with the event “Upload Document”
After setting up the event, you can sign in to your account using your API key, which can be found on the integrations page underneath the memory stack. Just select “Show API Key” next to Zapier. When you’re on the action step, use the “DomainName” window to enter your persona’s domain name. Use all of the characters before “.personal.ai” only.
You can also use our new “Instruct AI” even step to access the cleaning tools in the document during an automation. By using this step before the “Upload Document” step, you can natively clean your documents through a single automation by selecting which tool you want to use in the “Message Command” section.