Help centre
Answers about your visibility, your evidence, and your next step.
Short guides written for the person running the business, not the person running the servers. If a term on a screen is unfamiliar, start here.
Browse by topic
Getting started
Create a workspace, add your domain and competitors, and approve the buyer questions we monitor for you.
- Setting up your first workspace
- Choosing competitors that reflect your market
- Reviewing your Buyer Question Map
- Running your first scan
Reading your results
What each number counts, how the visibility map works, and how to read a result you disagree with.
- Brand named vs. site linked vs. recommended
- Using the visibility map
- Opening the evidence behind a result
- Why results move between runs
Sources and competitors
Understand which pages shape the answers in your category and where competitors are cited instead of you.
- Source archetypes explained
- Reading competitor source gaps
- When a source gap is real
- Adding or rotating competitors
Actions and verification
Take a recommendation from draft through review, publishing, and a live page check.
- The recommendation lifecycle
- Reviewing a content draft
- Reviewing a structured data change
- What verification does and does not prove
Technical signals
Crawler policy, page reachability, and citation observation are three separate layers.
- Reading AI access signals
- Fixing a page that cannot be fetched
- Structured data checks
- Where llms.txt fits in
Plans and usage
How monitoring capacity is counted and what happens when you reach your monthly allowance.
- How runs are counted
- Choosing a monitoring cadence
- Changing plans
- What happens at your limit
Common questions
The questions we hear most often in the first month.
- How long before I see results?
- Your first scan usually finishes within about ten minutes. You will see evidence immediately, but trends need at least two runs before they mean anything. Most workspaces have a useful picture after two to three weeks of weekly monitoring.
- Why does one of my surfaces say “not measurable”?
- Some AI surfaces do not offer a permitted, consistent way to collect answers and citations. Rather than guess, we mark those cells as not measurable and leave them out of every percentage. It is never counted as a zero.
- My result changed between runs. Is something broken?
- Probably not. AI answers vary by model version, wording, timing, and location. That is why every result carries a repeat sample count and a confidence level. Look at the trend across runs rather than any single answer.
- Can you guarantee my brand will be cited?
- No, and be careful with anyone who says otherwise. We show you what AI systems actually returned, where competitors won, and what you can change to improve your eligibility. The systems themselves decide what to cite.
- Do you publish changes to my website automatically?
- No. Every draft we generate has to be reviewed and approved by you, then published through your own site or store. After you publish, we rescan the page to confirm the change is live.
- What counts as one monitoring run?
- One question, checked on one AI surface, in one market, one time. If you monitor eighteen questions on one surface each week, that is eighteen runs per scan.
- Can I monitor a domain I do not own?
- You can analyse public pages for competitor research. Verifying ownership is only required for the parts of the product that act on your behalf, such as publishing workflows and integrations.
- How do I get help with something not covered here?
- Use the Need help link inside the app while you are on the screen in question. Including the workspace name and the run date helps us find the exact evidence you are looking at.
Terms you will see
Four states appear across the product. Everything else is built on them.
An AI answer linked to a page on your domain. The strongest state we can observe.
The answer mentioned your brand but sent the reader to someone else's page.
A competitor got the citation for a question you care about, and you did not appear.
No permitted way to collect this surface. Reported openly and left out of every rate.
Still stuck on a specific result?
Open the evidence drawer on the screen you have a question about. It carries the run date, the surface, the locale, and the exact answer we recorded.