Worst Performing AI Conversations
Worst Performing Conversations surfaces the tracked conversation prompts where brand visibility in AI responses is lowest—identifying where the brand is being overlooked and where improvement will have the greatest impact.
Quick start
1. Open the AI Visibility section and allow the page to load
2. Set Provider Position to Primary in the global filter bar—this focuses the analysis on whether the brand earns the leading recommendation
3. Set the date range and persona to match the analysis focus
4. Review the list—conversations with the lowest average visibility appear first
5. Check the platform bars to identify which AI platforms show the weakest brand visibility for each conversation
Understanding the display
Each row represents a tracked conversation prompt where brand visibility has been measured across one or more AI platforms. The conversations with the lowest average visibility across all platforms appear at the top of the list—these are the prompts where the brand is least likely to be recommended.
Visibility percentage is the core metric. It measures how often the brand appeared as a recommended provider in AI responses to that conversation prompt, expressed as a percentage of total runs. A visibility of 0% means the brand was not recommended in any run. A visibility of 100% means it appeared in every run.
The meta tags on each row show the platform with the primary data, the associated persona, journey phase, and average visibility across all platforms. The platform bars highlight the lowest-performing platform in each row.
Visibility percentage
Visibility is calculated per conversation per platform: the number of runs where the brand appeared as a provider, divided by the total number of runs for that conversation and platform combination, multiplied by 100. The average visibility shown in the meta tag is the mean across all platforms for that conversation.
Platform bars
Each bar shows the brand's visibility percentage on a specific platform for that conversation. The lowest-performing platform bar is highlighted. Where multiple platforms show 0%, the brand is absent from responses on all of them for that prompt.
Ranking
Conversations are sorted by average visibility, lowest first. Where two conversations have the same average visibility, the one with the lower minimum platform visibility ranks higher—surfacing the worst absolute gaps first.
What should I do next?
Use visibility levels to prioritise content and strategy work:
| What you see | Action |
| 0% visibility across all platforms | The brand is absent from AI responses for this prompt. Review whether brand content addresses the topic directly and consider whether additional content is needed. |
| 0% on one platform, higher on others | Platform-specific gap. The brand's content or authority signals may not be reaching that platform's data sources. Investigate what competitors are recommending for this prompt on that platform. |
| Low visibility (1–20%) across platforms | The brand appears occasionally but inconsistently. Review the quality and relevance of brand content for the query topic and assess whether it directly answers the prompt. |
| Visibility improving after content changes | The changes are working. Document what was updated and apply the same approach to other low-performing conversations. |
| No conversations shown | No run data exists for the selected filters. Expand the date range or adjust the persona and phase filters. |
Using the filters
Worst Performing Conversations has no per-component filters. All filtering is controlled by the global filter bar, which applies across the entire AI Visibility section.
Global filter bar
| Filter | What it controls |
| Date range | The period for visibility data. Options: 7 days, 30 days, 90 days, 180 days, or a custom range. |
| Persona | Limits results to conversations run under a specific buyer persona configuration. |
| Platform | Limits results to a single AI platform or shows all platforms combined. |
| Phase | Filters by journey phase (such as awareness or consideration). |
| Provider Position | The most impactful filter for this component. Controls which provider recommendations are counted. Primary limits to responses where the brand appears as the leading recommendation. Top 3 includes the top three. Any includes all positions. |
Choosing a Provider Position
Primary is the recommended setting for identifying the most significant gaps—conversations where the brand does not earn the top recommendation in AI responses. This is the default position when the component loads.
Top 3 broadens the analysis to include conversations where the brand appears in the top three recommendations. Lower visibility under Top 3 is harder to recover and indicates a deeper visibility gap.
Any includes all recommendation positions. Use this setting for a complete picture of brand presence across AI responses, including lower-ranked mentions.
Troubleshooting
No conversations shown
No run data exists for the selected filters, or the filters are too restrictive.
• Expand the date range
• Set Platform to all
• Set Phase and Persona to all
• Verify that conversation prompts have been configured and run
All conversations show 0% visibility
This indicates the brand is not appearing in AI responses for the tracked prompts under the current Provider Position setting.
• Check Provider Position—if set to Primary, the brand may appear in lower positions; switch to Any to confirm
• Verify that brand configuration is complete and the company name is set correctly
• Review whether brand content addresses the tracked conversation topics
Visibility seems lower than expected
Provider Position has a significant effect on visibility scores. Primary position is the strictest filter—only runs where the brand is the top recommendation count.
• Switch to Top 3 or Any to see how visibility changes
• Check the date range—a short range may capture a period of lower performance
• Filter by a specific platform to isolate whether the gap is platform-wide or concentrated
Expected conversations are missing
Conversations only appear when run data exists for that prompt within the selected period.
• Expand the date range
• Check that the relevant persona and phase are included in the filters
• Verify that conversation prompts are active and have been run recently
FAQ
How is visibility different from citation count?
Visibility measures how often the brand appears as a recommended provider in AI responses—whether the AI recommends the brand when answering a query. Citation count, shown in other components, measures how often the brand's pages are linked as sources. A brand can be recommended without being cited, or cited without being the top recommendation.
Why does Provider Position matter so much for this component?
Visibility is calculated as a proportion of runs where the brand meets the position threshold. Setting Primary means only runs where the brand is the leading AI recommendation count—a much stricter measure than Any. The same conversation can show 5% visibility at Primary and 60% at Any, reflecting the difference between earning the top slot versus appearing anywhere in the response.
What does it mean when one platform shows 0% and another shows high visibility?
Platform-specific gaps indicate that the brand's content or signals are reaching some AI platforms but not others. Different AI platforms use different data sources, crawl schedules, and weighting criteria. A gap on one platform is a targeted opportunity rather than a general content problem.
How many conversations are shown?
Up to ten conversations are shown—the ten with the lowest average visibility under the selected filters. Adjusting filters changes which conversations appear.
Can the data be exported?
Worst Performing Conversations does not include a built-in export. Contact support for data export options.
Getting help
6. Check Provider Position first—it has the largest effect on visibility scores
7. Switch to Any position to establish a baseline before narrowing
8. Contact support for brand configuration questions or unexpected behaviour