AI Platform Segment Analysis

AI Visibility Analysis turns the monitored answers in the AI Visibility dashboard into practical findings and prioritised actions.

You can analyse one audience configuration in detail or analyse all active configurations to identify patterns shared across personas.

The analysis does not run new monitoring questions. It reviews the processed monitoring evidence already available for the selected period.

What the analysis can help you understand

Use Analyse with AI to answer questions such as:

• Why is the brand visible for some personas but weak for others?
• Which AI platforms represent the largest visibility gaps?
• Is the brand mentioned but rarely cited?
• Do citations point mainly to your website or to third-party sources?
• Are the monitored questions balanced across Awareness, Consideration and Decision?
• Which competitors repeatedly win recommendations?
• Which positioning or content initiatives should be prioritised?
• Which improvements can Wheremore measure after implementation?

The report is an evidence-led interpretation. It is not a guarantee that a recommended change will alter an AI platform's answer.

Before you begin

You need:

1. An AI Visibility project in the active workspace.
2. At least one active monitoring configuration.
3. Saved monitoring questions.
4. Completed, processed responses for the selected period.
5. Your brand aliases and monitored domain configured correctly.
6. Sufficient credits for the requested analysis.

For useful cross-persona findings, the project should have at least two active configurations with comparable monitoring evidence.

Run an analysis

1. Open the AI Visibility dashboard.
2. Select the Project.
3. Select the Date Range.
4. Optionally select a Journey Phase and AI Platform.
5. Choose one configuration or All configurations.
6. Select Analyse with AI.
7. Keep the page open while Wheremore processes the report.
8. Review the priority initiatives before opening the detailed analysis.
9. Add suitable initiatives to the Action Board or discuss the stored report with Ask Wheremore.

The browser can wait for up to 20 minutes. A slow single-configuration request remains marked as processing rather than automatically starting a duplicate provider request.

Analysis scope

The analysis uses these dashboard selections:

• Project
• Configuration
• Journey Phase
• AI Platform
• Date Range

Custom date boundaries are inclusive.

Preset periods use UTC calendar dates. For example, Last 30 Days means today and the preceding 29 UTC dates.

Provider Position

The dashboard's Provider Position control is not currently passed into AI Visibility Analysis. It can change dashboard charts and tables, but the analysis uses the broader qualifying monitoring evidence for its selected configuration, phase, platform and period.

Do not assume the report is limited to Top 3 or #1 recommendations merely because that dashboard control was selected.

When All Platforms is selected, eligible Google AI Mode evidence is included alongside the other monitored platforms. Select Google AI Mode to analyse google-ai-mode  responses independently.

The analysis distinguishes:

• structured provider evidence used for visibility, rank, win-rate and sentiment findings; and
• answer-level sources used for citation and evidence findings.

A Google source is not assigned to Your Brands or a competitor unless the stored data contains a genuine relationship. This avoids turning a source used for a general statement into a false provider endorsement.

Single-configuration analysis

Selecting one configuration creates an AI Visibility Segment Insight.

The report analyses the selected persona or audience configuration using its monitored questions and processed responses.

Report header

The report identifies the configuration and selected scope. It can show:

• overall visibility;
• win rate;
• average provider rank;
• citation rate;
• platform and phase coverage; and
• the monitoring period.

Visibility Initiatives

These are the recommended actions derived from the selected evidence.

Each initiative can include:

• priority;
• the measurable problem;
• the recommended action;
• the rationale;
• affected platforms, phases or configurations;
• the primary success signal; and
• Add to Board.

Priorities are:

• Immediate
• Near-term
• Can wait

Treat every initiative as a hypothesis to validate. AI platforms decide how they retrieve, rank and present information.

Summary

The Summary explains the main visibility situation and highlights the strongest evidence.

Platform-Specific Gaps

This section identifies platforms where visibility is absent, weak or below the segment average.

A platform gap is based on the monitoring evidence in scope. It does not prove why the platform produced that result.

Citation Source Ownership

This compares resolved brand citations pointing to:

• your associated website; and
• third-party websites.

Only citations with a usable resolved destination can contribute to the ownership ratio. An unresolved citation is not silently classified as third-party.

High third-party ownership can indicate that AI platforms rely on other publishers when discussing the brand. It does not mean those citations are inaccurate or harmful.

Test Design Assessment

This reviews the monitoring-question set itself, including:

• question routing quality;
• Awareness, Consideration and Decision coverage;
• questions that may need revision; and
• priority improvements to the test design.

A weakly designed question can distort the monitoring picture. Improving the test set may therefore be more appropriate than changing website content.

Systemic Issues

Systemic Issues group recurring weaknesses across multiple questions, phases or platforms.

Review the affected counts and scope. A repeated pattern is more strategically important than a one-off response, but it still does not prove causation.

Competitive Insight

This section identifies the strongest observed competitor threat and possible positioning opportunities.

Competitive findings reflect the monitored question set. They are not a complete market-share analysis.

All-configurations and cross-persona analysis

Selecting All configurations starts a two-stage process:

1. Wheremore runs the segment analysis for each accessible active configuration.
2. It combines the completed results into a Cross-Persona Strategic Analysis.

If the project has only one eligible configuration, Wheremore falls back to the single-configuration report.

Processing progress

The page shows progress as configurations complete. A failed configuration can prevent the strategic synthesis from being complete.

The cross-persona stage only uses the trusted completed batch held by the server. It does not accept browser-supplied analysis results.

Priority Actions

Priority Actions identify improvements likely to benefit more than one persona.

Read:

• the problem being addressed;
• expected impact;
• personas benefited;
• applicable platforms and phases;
• priority; and
• the selected measurement signal.

Platform-Specific Gaps

This shows shared platform weaknesses and identifies which configurations are affected.

Citation Source Ownership

This aggregates the owned-domain and third-party citation patterns available across configurations.

Configurations without resolved citation ownership data are not treated as zero.

Visibility Locks

Locks describe recurring patterns that restrict visibility, such as:

• geographic locks;
• category or competitor locks;
• stage or persona locks;
• platform locks; and
• citation or authority locks.

These are analytical descriptions of the observed pattern, not confirmed details of an AI platform's internal systems.

Primary Positioning Recommendation

This proposes a narrative that can work across personas, together with its rationale and implementation direction.

Jobs-To-Be-Done Patterns

This groups recurring customer needs and indicates whether they align with the core category, an adjacent category or both.

Adjacency Risks

Adjacency Risks identify question themes that route attention towards another category or competitor.

Universal Issues

These are problems observed across a meaningful share of configurations.

Cross-Persona Competitor Analysis

This summarises the most consistent competitive threat and recommends a cross-persona response.

Understanding the metrics

Visibility

Visibility is the percentage of qualifying monitored responses in which one of your configured brand aliases appeared as a provider.

Win rate

Win rate measures how often your brand appeared in the leading provider position under the analysis evidence.

Average rank

Average rank is calculated from qualifying ranked brand appearances. It should be read with visibility because a good average from very few mentions can be misleading.

Citation rate

Citation rate measures the proportion of qualifying brand mentions that included at least one citation.

It does not measure whether the citation pointed to your website. That is covered by Citation Source Ownership.

Sentiment

Sentiment summarises the stored positive, neutral, mixed and negative classifications attached to qualifying brand appearances.

Mixed means the brand was presented with caveats, trade-offs or limited suitability. It is not the same as negative.

Add to Board and outcome measurement

Recommendations displayed by the report are bound to the completed server-side analysis.

Selecting Add to Board creates a diagnostic action with:

• its source analysis;
• the original evidence scope;
• priority;
• rationale;
• selected success signal; and
• outcome-measurement context where supported.

Aggregate signals can support later measurement, including:

• visibility;
• brand mention rate;
• win rate;
• citation rate;
• citation ownership; and
• platform visibility gaps.

Granular recommendations about a particular message, topic, feature or narrative may be retained without automated outcome measurement when the supplied evidence cannot remeasure that exact dimension.

An improved metric after implementation does not prove the recommendation caused the improvement.

Ask Wheremore

Completed reports can be stored as analysis context for Ask Wheremore.

Use it to:

• explain a metric or finding;
• compare the priorities in the report;
• turn a finding into a test plan; or
• relate the report to other supported dashboard evidence.

Ask Wheremore uses the stored completed analysis. It cannot complete a failed job or recover evidence that was not present when the report was generated.

Limitations

• The analysis uses processed monitoring evidence, not live AI answers.
• Results depend on the selected questions, personas, platforms and dates.
• A platform gap does not reveal the platform's private retrieval or ranking mechanics.
• Citation ownership requires resolved destination URLs.
• Small samples are directional.
• Cross-persona averages should not hide important configuration-specific differences.
• Provider Position is not currently part of the analysis request.
• Monitoring responses can change independently of work completed in Wheremore.
• Recommendations do not guarantee visibility, ranking, sentiment or citations.

Troubleshooting

Select a project first

Choose a project available in the active workspace.

No configurations are available

• Confirm the project has an active AI Visibility configuration.
• Confirm the configuration belongs to the selected project.
• Confirm workspace and account access.

No data is available for the selected filters

• Expand the date range.
• Select All Phases or All Platforms.
• Confirm monitoring jobs completed successfully.
• Confirm raw responses were processed into runs.
• Check that the selected configuration had questions during the period.

The analysis is taking longer than expected

Provider calls can take several minutes. Keep the page open.

After 12 minutes, the interface can report that processing is taking longer than usual. A single-configuration request is not automatically duplicated.

The browser stops polling after 20 minutes. If the report still does not appear, check background processing and provider logs before starting another analysis.

A batch or cross-persona analysis is stale

Batch and strategic jobs use progress-based stale detection and limited retries. Allow the current retry to finish. If the job ultimately fails, rerun the analysis after checking the failed configuration or provider request.

The citation ownership section is missing

• Confirm citations were extracted.
• Confirm citation destinations were resolved.
• Confirm the project has an associated monitored domain.
• Confirm your brand aliases identify the intended provider.

Unresolved citations cannot be assigned to owned or third-party destinations.

Add to Board is unavailable

The recommendation may no longer have a valid source signature, or it may not be bound to a stored completed analysis. Rerun the analysis and use the new recommendation.

Ask Wheremore has no report context

Wait for the analysis to complete. Only stored completed analyses are available as report context.

The report conflicts with a dashboard chart

Check:

• project;
• configuration;
• phase;
• platform;
• UTC date range;
• processing time; and
• metric definition.

Also remember that Provider Position is not currently passed into the analysis.