Skip to main content
The Action Center in Prefer turns tracking data into a prioritized playbook. It surfaces evidence-backed actions you can take to improve how often AI engines cite your brand. Every card shows an impact score, the affected prompts and engines, and a step-by-step plan with copy-ready assets. You move cards through a Kanban board from suggestion to completion, and Prefer measures citation lift on later tracking runs.

​
How the board works

The board is a Kanban view with columns that match the stages of a typical workflow:
  • Suggested — actions generated from your latest tracking run or audit.
  • To do — actions you have accepted and plan to work.
  • In progress — actions you are actively executing.
  • In review — actions waiting for a final check before completion.
  • Done — completed actions, with measured outcomes where available.
  • Dismissed — actions you chose to skip, with an optional reason.
You can also switch to a List view when you want to sort, filter, or scan many actions at once.
Suggested actions regenerate after each tracking run. If you previously accepted or dismissed an action, Prefer matches it by signature and keeps your stage intact.

​
The seven action categories

Parity, off-page, and community actions are generated live from tracking data. Technical, schema, entity, and analytics actions are sourced from your most recent audit when available, and supplemented with best-practice templates otherwise. The provenance is shown on each card.

​
Card anatomy

Each card on the board contains:
  • Impact score — a transparent calculation based on Reach, Winnability, and Confidence divided by Effort. Higher scores mean bigger expected citation gains for the time invested.
  • Category tag — one of the seven categories above, color-coded for quick scanning.
  • Owner label — agent, assisted, or human. This is a filter and classification label, not an automation trigger. In the current release, all actions are manual.
  • Effort — low, medium, or high, so you can batch quick wins or plan larger projects.
  • Evidence — the competitor, source, or crawler signal that triggered the action, plus the specific prompts and engines where the gap was observed.

​
Working an action

Follow this flow to move an action from suggestion to completion.
1

Accept or dismiss

In the Suggested column, click a card to open the detail modal. Review the claim and evidence. Click Accept to move it to To do, or Dismiss with a reason if it is not relevant.
2

Review the detail modal

The modal is organized into three sections:
  • Claim — what the action will achieve and why it matters.
  • Proof — the evidence behind it, such as a competitor citation count, a blocked crawler, or a missing source.
  • Plan — numbered steps with copy-ready assets where available.
3

Use copy-ready assets

Expand asset cards inside the plan to access:
  • Outreach email templates
  • robots.txt or schema markup diffs
  • Checklists
  • Fact sheets Click the copy button on any asset to paste it into your workflow.
4

Drag through stages

Return to the board and drag the card from To do to In progress when you start work. Move it to In review when the change is live, and to Done when you have verified the result.
5

Measure citation lift

After the next tracking run completes, Prefer compares your new citation count against the baseline captured when you first accepted the action. The outcome appears on the done card.
Filter the board by category, owner, or effort to batch similar work. For example, group all low-effort technical fixes into one sprint.

​
Measuring outcomes

When you accept an action, Prefer records a before-cited baseline from the current tracking run. After the change goes live and a later run completes, the card shows the delta:
  • Before — citations or mentions at the time of acceptance.
  • After — citations or mentions on the most recent run.
  • Lift — the absolute and relative change.
This measurement reuses the same citation counting logic you see in the Monitor > Sources and Citations console, so the numbers are consistent across Prefer.
Outcomes remain null until a later tracking run completes after the action is marked live. If you dismiss an action, no baseline is recorded.
Content Studio
Sources and Citations