People often ask: “How do I get Weasley to really get me?” The answer is simpler than prompt-engineering Twitter would have you believe. You teach it the same way you’d onboard a sharp new hire — by working together and giving light feedback.

Start with how you like answers

The fastest win is telling Weasley your output preferences:

• “Keep responses under 200 words.”
• “Use bullet points, not paragraphs.”
• “Be direct — skip the preamble.”
• “Write in a casual tone, not corporate.”

Say it once in a conversation, or add it to your profile settings. Weasley treats these as standing instructions, not one-off requests.

Share context as you go

You don’t need a perfect onboarding doc on day one. Just mention what’s relevant:

“I’m preparing for a Series A pitch next month. The deck is rough — focus on narrative, not design.”

“My kid starts school in August. I need help planning the logistics.”

Weasley builds context incrementally. A throwaway detail today becomes useful background next week.

Correct, don’t restart

When an answer misses the mark, correct it in place:

• “Too formal — rewrite like I’d text a colleague.”
• “You missed the pricing section. That’s the main gap.”
• “I already decided on Option B. Help me execute, not re-debate.”

Corrections are gold. They teach Weasley more than a perfect first prompt ever could.

Connect your tools when you’re ready

Calendar, email, and notes give Weasley signal it can’t infer from chat alone. But don’t feel pressured to connect everything on day one. Many users start with chat only and add integrations once they trust the basics.

Give it a week

Personal AI gets meaningfully better after a handful of real tasks — not toy examples. Draft an email. Plan a trip. Summarize a doc. Push back on a bad answer. By day seven, you’ll notice the difference.

The goal isn’t a perfect profile on day one. It’s an assistant that learns your rhythms and stops wasting your time. That’s the whole point.