A 30-day plan will not fix your AI visibility, but it gives you something more valuable: knowing exactly where you stand and what to move first.
Week 1 — Baseline
- Define the set of 15-25 customer prompts (discovery, comparison, price, objections, local).
- Run them across the main assistants and record: mention yes/no, position, competitors and cited sources.
- Compute your initial mention rate and share of voice against the three most-named competitors.
By the end of the week you have a number. That number is the whole project.
Week 2 — Diagnosis
- Cluster the sources AI cites: media, directories, comparison sites, forums? That is the map of your category.
- Mark the key sources where you are absent: those are your gaps.
- Audit your own house: entity consistency (name, category, description), structured data, visible pricing, answer-shaped content.
- Check robots.txt: make sure you are not accidentally blocking AI crawlers.
Week 3 — First actions
Prioritise by effort/impact, not taste:
- Fix the entity (fast, high impact).
- Publish or rewrite 2-3 answer-shaped pages on the questions where you are missing.
- Add Organization, Product and FAQPage where absent.
- Start outreach to the two most-cited third-party roundups in your category.
Week 4 — Measurement and cadence
- Re-run the exact same prompt set, same day of the week.
- Compare mention rate, average position and share of voice against the baseline.
- Document what changed and what did not: four weeks is early for new content, enough for entity fixes.
- Set the cadence: weekly scan, monthly review, quarterly report to leadership.
What to realistically expect
Entity fixes show up within weeks. New content takes one to two months. Third-party roundup presence is the slowest and most durable lever. The key is not speed but never stopping the measurement: without a time series, any improvement is indistinguishable from chance.