1. Using AI to Build a Custom Field
Go to Employees → Custom Fields and create or open a custom field.
The logic canvas opens — click the AI Assistant button (sparkle icon) on the right side to open the chat panel.
You'll see quick action prompts to get started, like:
"Flag employees below target"
"Create a segment"
"Calculate a score"
Pick one, or type a plain-English description of what you want (e.g., "Create a field that flags anyone whose compa-ratio is below 0.9").
The AI will build the logic graph for you — you'll see nodes appear on the canvas in real time.
It will also test the logic against sample employee data and show you results so you can validate.
If you want changes, just ask in the chat (e.g., "Change the threshold to 0.85" or "Also factor in tenure").
When you're happy, the AI submits the final result and you can save.
2. Using AI to Build Logic Within Comp Cycle Canvases (Salary, Bonus, Equity)
Open a comp cycle and go to the Salary, Bonus, or Equity canvas.
Click the AI Assistant button to open the chat panel.
You'll see suggested quick actions tailored to each surface:
Salary: "Merit / market / promotion split", "Move to range midpoint", "Merit matrix (perf × compa-ratio)"
Bonus: "Target × performance multiplier", "Weighted multi-component bonus", "Custom field passthrough"
Equity: "Grant to fill equity gap", "Refresh from performance", "Location-adjusted equity"
Select a quick action or describe what you want in plain English.
The AI builds the node graph on the canvas — you can watch nodes and connections appear.
It tests the logic against real proposals in your cycle so you can see actual computed values.
Iterate by chatting — ask it to tweak thresholds, add conditions, etc.
Once finalized, the AI submits and you save.
3. Asking AI to Analyze and Explain an Existing Logic Graph
On any logic canvas (custom field, salary, bonus, equity, outliers), look for the Explain button.
Click it to open the Explain Logic modal.
The AI will automatically analyze the full graph and give you a plain-English summary of what the logic does — no code or technical jargon.
You can then ask follow-up questions in the same conversation, like:
"What happens if an employee has no performance rating?"
"Which employees would get the highest increase?"
"How does this handle people in different pay zones?"
You can copy the explanation to share with stakeholders.