Feedback & Learning

Live demo · example data. Read-only preview — buttons, templates and AI only work in your own copy. → www.neurocanvas.ai

</aside>

<aside>

Make the AI better over time.

</aside>

Agent Feedback

For ideas that affect several skills or Master AI. Most daily feedback is handled inline in chat.

Make sure the skill is enabled (the Enabled tab on Agent Skills).
Open a new AI chat and paste the test prompt.
Check: did it use the right skill, read the right database, and give the output described in "What good looks like"?
Off? Type Feedback: … in that same chat.

Skill	Test prompt	What good looks like
‣	"I have 4 hours today and too much on my list. What should I do?"	Max 3 priorities from your Tasks, one thing to drop or defer, done in under 10 lines.
‣	"Review my week and plan next week."	Wins, misses, open loops from Tasks/Projects, 3 priorities for next week, offer to save it as a row.
‣	"Should I raise my prices now or after the next launch?"	Options side by side, trade-offs, one clear recommendation, a reversible next step.
‣	"Turn 'relaunch my newsletter' into a plan."	Outcome, milestones, first 5 tasks — proposed as Project + Task rows in the right canvas.
‣	"Write an SOP for how I onboard a new client."	Purpose, numbered verb-first steps, tools, owner, frequency — saved as an SOPs row.
‣	"Here are my messy call notes: … Turn them into a summary."	Decisions, action items with owners, follow-up — linked to the meeting and tasks.
‣	"Draft a short follow-up email to a client who went quiet."	A ready-to-send draft in your voice, one alternative subject line, no filler.
‣	"I'm tired all week. Help me find one small change."	One question at a time, one small experiment, offer to log it in Try a Change.
‣	"Set me up"	A short interview (max 8 questions), then filled Knowledge Base sections for your approval.