Not every question requires the same level of AI reasoning.
Sometimes you need a quick answer:
“What meetings do I have today?”
“Summarize my last conversation with John.”
“Find the email where a client mentioned their vacation plans.”
Other times, you need Vega to slow down and work through something more complex:
“Analyze this client situation and identify planning opportunities.”
“Compare these two strategies and explain the pros and cons.”
“Review this information and help me prepare recommendations.”
Reasoning modes allow Vega to adjust based on what you need: faster answers when the task is simple, deeper analysis when the task requires more reasoning.
How Reasoning modes work
At the bottom left of every Vega chat thread, next to the (+) button, you will see the Reasoning mode selector.
You can choose between three options:
Auto
Instant
Thinking
Auto: Let Vega decide
Auto is the default mode for every new chat thread.
In Auto mode, Vega analyzes your request and decides which approach is best.
For simpler requests, Vega will prioritize speed.
For more complex requests that require deeper reasoning, Vega can spend more time analyzing the information before generating a response.
For most users and most workflows, Auto is the recommended option because Vega automatically adapts to what you are trying to accomplish.
Instant: Fast, no reasoning
Instant is designed for speed.
This mode works best for everyday interactions where you want Vega to quickly retrieve information or answer straightforward questions.
Examples:
“Who is John Smith?”
“Summarize my last meeting.”
“Find notes mentioning Roth conversions.”
“Draft a quick email response.”
Instant is especially useful when:
you already know exactly what you need
the request is simple
you are having a quick back-and-forth conversation with Vega
Thinking: Deeper reasoning
Thinking gives Vega more time to analyze before responding.
This mode is designed for more complex questions where the first answer is not always obvious.
Examples:
“Analyze this client’s situation and tell me what I may be missing.”
“Compare different planning strategies based on this information.”
“Review this meeting history and identify patterns across conversations.”
“Help me prepare for this important client meeting.”
Thinking is especially useful when your request requires:
deeper analysis
connecting multiple pieces of information together
reviewing larger amounts of context
multi-step reasoning
more comprehensive answers
How to choose to the right mode
A simple way to think about it:
Auto → Let Vega decide
Instant → I need a quick answer
Thinking → I want Vega to spend more time analyzing
Auto will handle most situations well, but you always remain in control.
If you know your question is simple, you can switch to Instant.
If you know your question requires more careful analysis, you can switch to Thinking.
A practical way to think about it
Think about how you would work with a human assistant.
Sometimes you ask a quick question and expect an immediate answer.
Sometimes you give them a more complex problem and want them to spend more time researching, analyzing, and thinking through the details.
Reasoning modes allow Vega to work the same way.
Together, they help Vega adapt to the type of work you are doing.