The Knowledge section allows you to teach your assistant how you work, what you care about, and how you want tasks performed.
To access Knowledge, navigate to: app.skej.com/knowledge
Knowledge helps your assistant make better decisions, provide more personalized responses, and complete tasks in a way that reflects your preferences.
How Knowledge Works
Your assistant learns in two ways:
Assistant-Learned Memories
As you interact with your assistant, it automatically creates memories about your preferences, habits, and working style. For example, if you tell your assistant:
"I don't like meetings on Friday afternoons."
"Keep emails short and direct."
"Always include action items in meeting summaries."
"I prefer window seats when traveling."
Your assistant may remember those preferences and apply them in future tasks.
Over time, your assistant builds a better understanding of how you like to work.
User-Provided Instructions
You can also proactively provide instructions and preferences. This is useful when there are things you already know you want your assistant to remember.
Rather than waiting for the assistant to learn through experience, you can define preferences up front.
General Preferences
General Preferences contain instructions that apply broadly across many tasks and interactions. Think of these as your assistant's understanding of how you prefer to work. Examples:
"Keep communications concise unless I ask for detail."
"When making recommendations, include pros and cons."
"Flag scheduling conflicts aggressively."
"Always identify action items and next steps."
General preferences help your assistant make better decisions across all areas of the platform.
Task-Specific Instructions
In addition to general preferences, you can provide instructions for specific types of tasks. These instructions only apply when your assistant is performing that category of work.
Task-specific instructions allow you to customize how your assistant performs different types of work.
Scheduling Instructions
Teach your assistant how you want meetings and calendar coordination handled. Examples:
"Avoid Friday afternoons when possible."
"Group meetings together to preserve focus time."
"Always prioritize customer meetings over internal meetings."
"Protect lunch from 12:00 PM to 1:00 PM."
These instructions help your assistant make scheduling decisions that reflect your preferences.
Follow-Up Instructions
Define how your assistant should handle follow-ups and communication. Examples:
"Always follow up after customer meetings."
"Keep follow-up emails brief and action-oriented."
"Escalate unanswered customer requests after one week."
Research Instructions
Customize how your assistant performs research and analysis. Examples:
"Focus on recent developments from the last 12 months."
"Include competitor analysis."
"Prioritize primary sources whenever possible."
"Summarize findings in bullet points."
These instructions can be applied to company research, people research, market research, and other investigative tasks.
Document Creation Instructions
Teach your assistant how you prefer reports, spreadsheets, presentations, and written content to be structured. Examples:
"Start reports with an executive summary."
"Keep presentations concise."
"Use bullet points whenever possible."
Building a Better Assistant Over Time
You do not need to capture every preference immediately. Many users start with a handful of instructions and allow the assistant to learn over time. As you work together, you can:
Add new instructions
Update existing instructions
Remove outdated preferences
Refine task-specific guidance
The combination of your instructions and the assistant's learned memories creates a personalized assistant that becomes more useful over time.
Best Practices
To get the most value from Knowledge:
Add instructions that reflect how you actually work.
Focus on recurring preferences and workflows.
Create task-specific guidance for work you do frequently.
Update instructions when your preferences change.
Think of Knowledge as onboarding for your assistant. The more context you provide, the better it can represent your preferences, make decisions, and complete work the way you would.