Optimize — Actions
Take action on your GEO score improvements.
The Optimize page is your Impact Queue — a prioritized list of specific fixes that improve your AI Visibility Score, generated automatically after each scan.
What the Impact Queue is
After each scan, Genlytic analyzes your results and generates optimize actions. Each action targets a specific gap between your current brand presence and where it could be. Actions are ranked by impact score and deduplicated — the same action type on the same target will not appear more than once per 7-day window.
Action types
Seven action types are generated based on scan results:
| Type | What triggers it | Placement | Visibility delta |
|---|---|---|---|
| Schema Gap | Your URL cited in 2+ prompts with no schema markup | On-page | +40 pts |
| FAQ Block | Question-type prompt where your brand is not Tier 1 on any model | On-page | +25 pts |
| Answer Block | Your brand is Tier 3 across ALL models for a prompt | On-page | +20 pts |
| Citation Gap | A source cited by AI doesn't mention your brand (impact >= 2) | Off-page | +35 pts |
| Narrative Fix | Negative sentiment + AI context contradicts your ground truth | On-page | +30 pts |
| Content Gap | Prompt with win rate = 0% and no brand domain in citations | On-page | +45 pts |
| Tier Upgrade | Brand at Tier 3 for 2+ consecutive scans on the same model | On-page | +25 pts |
Impact Score
Each action has an impact score of 1, 2, or 3:
impactRaw = (visibilityDelta/100 × modelCount/4 × citationWeight) / effortScore
citationWeight: 1.0 (avg tier <= 1.5) | 0.6 (avg tier <= 2.5) | 0.3 (avg tier > 2.5)
impactScore: 3 (high) if impactRaw >= 0.33
2 (medium) if impactRaw >= 0.15
1 (low) if impactRaw < 0.15
Actions affecting more AI models and those with lower effort scores receive higher impact scores.
Effort map
Effort scores (1=low effort, 4=high effort) by action type:
| Action | Effort |
|---|---|
| Schema Gap | 1 — add JSON-LD to an existing page |
| FAQ Block | 2 — add structured FAQ section to a page |
| Answer Block | 2 — restructure page answer format |
| Citation Gap | 3 — outreach to a third-party domain |
| Narrative Fix | 2 — update core content to correct AI narrative |
| Content Gap | 4 — create a new page targeting a prompt |
| Tier Upgrade | 3 — content and structural improvements |
How to use the queue
- Open the Optimize page — actions are sorted by impact score (highest first)
- Click an action to read the title, description, affected prompts, and affected models
- Click Generate Fix to produce a specific fix recommendation (streaming output)
- Copy the generated fix content and apply it to your site
- Mark the action as Applied when done, or Dismiss to remove it from the queue
Status workflow
| Status | Meaning |
|---|---|
| pending | Action is in the queue, not yet acted on |
| applied | Fix has been implemented |
| dismissed | Action was intentionally skipped |
Fix generation uses Claude to produce specific, actionable output — JSON-LD schema, FAQ HTML, answer restructure guidance, or outreach email copy — tailored to the specific prompt and action type.
Next steps
- Optimize — Briefs — longer-form content creation
- GEO Agent — automate action execution