A first run is a snapshot, not a trend. With one generation across a handful of prompts and models, scores like Mention Rate, Share of Voice, or Sentiment are based on a small sample — they will move, sometimes a lot, as more data comes in over the next few weekly runs. Don’t over-read a single week of numbers.
What to check first
Open Cockpit > Overview and look for any mentions at all
Go to Cockpit > Overview (the brand root page) and check the 5 core KPIs: Mention Rate, Citation Rate, Share of Voice, Average Sentiment, and Coverage. At this stage, don’t fixate on the exact percentages — first confirm your brand is mentioned in at least some responses. Zero across the board is the signal worth investigating (see below), not a low number.
Check performance by AI provider
Scroll to Performance by AI Provider on the same dashboard. Since providers pull from different sources — some from live web search (Perplexity, AI Overview), others mostly from training data (ChatGPT, Claude) — it’s normal to see your brand show up on one provider and not another this early. Note which providers mention you and which show 0%, so you know where to focus once you have more data.
Check the Action Center for early recommendations
Head to Cockpit > Action Center (available on Starter and above). Even from a first run, the engine may already surface early suggestions — for example Setup actions (foundations like robots.txt or llms.txt) or the first Prompt Gaps. Review the Suggestions tab and accept anything that looks like a genuine quick win; decline what doesn’t apply. Don’t expect a full queue yet — the richer families (Competitor Gaps, Sources & PR) need more monitoring history to have enough signal.
Read the actual answers in Prompt Tracking
Go to Monitor > Prompts and open a few individual prompts. This is the most important sanity check: read the AI responses verbatim, not just the aggregate score. Ask yourself two things — does this response make sense for a customer to ask when researching your category, and does your brand’s absence (or presence) in the answer feel right given what you know about your market? This is also where you’ll spot if a generated prompt doesn’t actually fit your brand.
What’s normal vs. what’s a problem
Normal in week one:- Low or 0% on some individual providers — this often reflects how that provider sources answers, not a real gap.
- A Mention Rate or Coverage that looks unimpressive on 10-20 prompts — small samples swing widely.
- An Action Center with only a few Setup-type suggestions rather than a full prioritized queue.
- Zero mentions across every provider and every prompt. Before treating this as a visibility crisis, check whether it’s a targeting problem instead: go to Workspace > Topics and Workspace > Prompts and ask whether the auto-generated topics and prompts actually match your category and the way customers describe it. Overly narrow, off-topic, or oddly phrased prompts will produce zero mentions even for a strong brand. Adjust or add prompts, then wait for the next run.
- A provider stuck at 0% while your Coverage elsewhere looks healthy. That’s usually a real, provider-specific gap (see the FAQ on “Why does a provider show 0% visibility?”) rather than a setup issue — worth flagging for later content or source work, not an emergency today.
Where to go next
- Core Concepts — the vocabulary behind every metric you just saw (mentions, citations, Share of Voice, sentiment).
- Performance Dashboard — the full breakdown of each KPI, by provider, keyword, and source.
- What quick wins can I achieve to boost my GEO score this month? — once you’ve got a few runs behind you, use the Action Center to build a focused 30-day plan.