Knowing you have a problem is only half the battle. If your dashboard shows a 0% visibility score on Gemini, or tells you that only 8 prompts have been tested so far, the natural next question is: okay, so what do I actually do about it?
That’s where most AI visibility tools stop and where Recommendations picks up.
Instead of leaving you to interpret raw numbers on your own, Recommendations translates your visibility data into a clear, prioritized action plan. Think of it as the difference between a doctor handing you a lab report full of numbers, versus a doctor sitting down and saying, “Here’s what these numbers mean, and here’s what we should do about it.”
Here’s a step-by-step guide to actually using it.
Step 1: Start with what “needs attention”
Every audit begins with triage. Recommendations surfaces the issues that are actively hurting your visibility first the equivalent of a “Critical” or “Needs Attention” flag. This might look like:
- “Critical: Gemini visibility is only 0.0% you’re nearly invisible on this platform.”
- “Only 8 prompt executions found run more tests for reliable insights.”
These aren’t vague warnings. Each one points to a specific, fixable gap. Before touching anything else, work through these flagged items first they’re the issues most directly dragging down your overall score.
Step 2: Expand your prompt coverage
If your visibility data is based on a small number of prompts, the insights underneath it are inherently shaky a handful of test queries can’t reliably represent how your brand performs across the full range of ways real customers ask AI for recommendations.
Action: Go back to Prompt Management and add more realistic prompts mixing categories, geographies, and comparison-style queries (like “best GEO company for startups“ or “top GEO company in Italy“). More prompts tested means more reliable recommendations going forward, since the system has more real data to base its suggestions on.
Step 3: Target your weakest platform first
Recommendations doesn’t treat all AI platforms the same and neither should you. If your data shows strong visibility on ChatGPT but near-zero visibility on Gemini, that’s not a coincidence; it’s a signal that something about how you’re represented (or not represented) differs meaningfully between platforms.
Action: Prioritize the platform with the lowest score first. Trying to improve everywhere at once spreads your effort too thin fixing your weakest link tends to move your overall visibility score the most.
Here’s what this looks like in practice inside the Recommendations panel a real snapshot showing 0% visibility on Gemini, alongside two other panels we’ll get to next:
The Recommendations page: overall AI visibility broken down by platform, the external sources AI is citing instead of your own domain, and the competitors currently dominating the conversation.
Step 4: Strengthen the content behind your citations
Citations are how AI models justify their answers they’re the sources being pulled from to build a response. If your total citation count is low, Recommendations will often point you toward the underlying cause: your website or public content may not contain the kind of clear, structured, authoritative information that AI models tend to cite.
Look closely at the “Sources AI Trusts” panel above. Notice that every listed source punku.ai, aigl. blog, tredence.com, and others is marked “Not listed,” meaning none of them are your own domain. In other words, AI is actively citing sources for this topic, just none of them are you. That’s a direct, actionable signal.
Action: Review the pages AI models are pulling from instead of yours (as shown in the “Sources AI Trusts” panel). Make sure key facts about your product, pricing, use cases, and differentiators are clearly stated on your own site not buried in marketing language that’s hard for an AI system to extract and cite confidently.
Step 5: Study Your Competitive Gaps
Sometimes a low share of voice isn’t about you being absent it’s about competitors dominating the same conversation. Recommendations often highlights where a competitor is consistently outperforming you in prompts you’d expect to win.
The “Competitors in AI Responses” panel above makes this concrete: Accenture, SenseTime, Tiger Analytics, Google Cloud, Microsoft, and IBM are each capturing an even 19% share of the AI conversation for these prompts while your own brand doesn’t appear in that list at all.
Action: Look at the prompts where competitors are cited more often than you. Ask what their content, positioning, or public information might be doing differently then close that specific gap rather than guessing broadly.
Step 6: Fix, then re-test
Improving AI visibility isn’t a one-time fix it’s a loop. Once you’ve acted on a recommendation (added prompts, improved a page, addressed a platform gap), the next step is simply re-running your tracked prompts to see if the change moved the needle.
Action: After making changes, re-check your Visibility Trend chart over the following weeks. AI models don’t update instantly, so give it time but consistent tracking is how you separate real improvement from noise.
Step 7: Repeat the cycle
AI visibility isn’t something you solve once and forget. New competitors enter the conversation, AI models update, and customer language shifts over time. Treat Recommendations as an ongoing checklist you return to regularly, not a one-time audit.
Action: Set a recurring cadence monthly or quarterly to revisit your Recommendations panel, review new flagged issues, and confirm previous fixes are holding.
Why this step-by-step approach works
The biggest trap in AI visibility work is trying to fix everything at once based on a vague sense that “we need to be more visible on AI.” Recommendations avoids that trap by doing the prioritization for you pointing to the specific platform, the specific prompt gap, or the specific citation issue that’s costing you the most visibility right now.
That specificity is what turns a dashboard full of numbers into an actual plan you can execute this week.
Final thought
Data without direction is just noise. The real value of tracking AI visibility isn’t the score itself it’s knowing exactly what to do to improve it. Recommendations closes that loop, turning “why is our visibility score 0%” into a clear, prioritized checklist you can start working through today.
Ready to turn your AI visibility data into action? Check your Recommendations panel on Lunocode and start closing your biggest visibility gaps.
SEO & GEO specialist.
