Instructions · Working with your results
One report. One clear next task.
Work through your Transparion report step by step: interpret the metrics, review answers and sources, and turn a finding into a concrete improvement.
01 · Put in context
The score is your starting point.
The values describe the answers and content examined. They do not represent overall market share or predict enquiries or revenue.
- AI Visibility / Visibility Score · Percent
- The percentage of evaluated prompt responses that mention your brand. Repeated mentions within the same response count only once. In the Pro report, this metric is called Visibility Score.
- Weighted Visibility · Percent
- This additional metric weights category questions at 50%, use cases at 30% and specialisations at 20%. Read it alongside the unweighted value and the size of each question group.
- Citation Share · Percent
- The percentage of evaluated responses that include a source from your domain. This is not your domain’s share of all individual source links. If the response set contains no visible citations, the report may use documented search sources as a proxy; check the metric explanation in your results.
- Brand Sentiment · Points
- Separate brand research combines positive and negative evidence into a score from 0 to 100 points. Read the evidence in context: the score is not the percentage of positive prompt responses.
- AEO Content Quality · Points
- Feedback on the clarity, structure and supporting evidence of a submitted page, scored from 0 to 100. It does not replace a full technical website audit or guarantee a future mention or citation.
- Competitors · The same question set
- Compare the vendors appearing in the same evaluated responses. Open the answers to understand their reasoning, limitations and sources.
02 · Transparion Insight
Take a closer look at one buyer question.
Insight examines automatically derived questions with ChatGPT. You receive your free online results through an email link, without creating an account. You can optionally submit a content page for review.
Start a free analysis- Read the answer
- Choose a question your actual target audience might ask. Is your brand mentioned? Does the description match your offering? Which alternative is recommended?
- Open the sources
- Review the listed URLs yourself. Is the information current, relevant and correct? A source link alone does not establish why the model mentions a vendor.
- Review the first recommendation
- Match the recommendation to a specific product, use-case or comparison page. If you submitted a content URL, also use the feedback for that page.
03 · Transparion Pro Scan
Identify patterns and continue in your account.
With the Pro Scan, you choose 10–30 questions, a market and target audience, up to four competitors and up to five content URLs. The analysis covers ChatGPT, Claude and Gemini.
See the complete process- Focus the comparison
- Review the same buyer question across models and competitors. Group similar gaps by use case, requirement or target audience instead of comparing only overall scores.
- Define work packages
- Use the detailed content briefs and configured technical SEO checks. Assess the findings and effort before handing tasks to your content or development team.
- Work with your results
- In your account, you can save prompt sets, schedule further scans and export reports as PDF or editable PowerPoint files. Each website keeps its own history.
04 · From observation to action
Turn a finding into a brief.
Illustrative example: a SaaS vendor does not appear in an answer about a specific integration. This identifies something to investigate; it does not establish a cause.
- Document the observation
- Record the buyer question, model, measurement date and answer. The brand is absent; another vendor is described as offering the requested integration.
- Test the hypothesis
- Check whether your integration page clearly explains its functionality, requirements and limitations. Consult your product team and current documentation for evidence.
- Assign the change
- Define the target URL, owner and deadline. Add missing information and link to the page from the relevant product section. Acceptance criteria: claims are correct and supported by current documentation.
- Plan the follow-up review
- Document publication and repeat the measurement later with a comparable configuration. Record changes in the answers separately from the work completed.
05 · Comparable measurements
Changes need context.
A trend appears in your account after at least two completed Pro Scans of the same website. Where possible, compare the same questions, models, markets, target audiences and competitors.
Compare Insight and Pro Scan- Flag changes
- Record new prompts, configuration changes and published content. A changed question set can shift the overall score even when brand perception has not changed.
- Review the answers
- Review the specific answers and sources behind notable changes. Model responses can vary. A single comparison does not prove that a particular action caused the change.
Common questions
Useful to know before you continue.
When should I measure again?
Plan your follow-up measurement after a documented change, using a configuration as close to the original as possible. A fixed time period does not guarantee a different answer. Comparable conditions and a review of individual responses matter most.
Why do the models differ?
The systems examined may answer the same question differently and cite different sources. Read each model’s responses individually; Insight provides a starting point with ChatGPT, while Pro Scan adds Claude and Gemini.
Does Transparion implement recommendations automatically?
No. Insight and Pro Scan provide analysis and recommendations. Your team reviews, prioritises and publishes changes. Advisory and Managed are planned as additional services.
For your next step