How to analyze prompt details

Prompt details show the analysis of one specific question asked to AI models. This is the place where you can see not only general metrics, but also the actual model responses, brand position, sentiment, sources, and visibility changes over time.

This makes it easier to understand what affects your brand’s result for a specific user question.

What are prompt details?

A prompt is a question or instruction that a user can give to an AI model.

Example:

Prompt details show how AI models respond to this one specific question.

In this view you can check:

• whether your brand was mentioned

• which position it appeared in

• what sentiment the answer has

• which AI models generated the answers

• which sources were used

• which competitors show up for the same question

• how visibility changes over time

Prompt details help you move from a general report to a specific situation: one user question and one purchase intent.

Where to find prompt details

To go to the prompt details:

1. Log in to the Semly panel

2. Select the right project

3. Go to the Prompts section or Visibility report

4. Click the selected prompt

5. Open the detailed analysis view

What to check first?

After opening the prompt details, start with the basic information.

Check:

• prompt content

• prompt topic

• prompt status

• brand visibility

• brand position

• sentiment

• AI models covered by the analysis

• latest answers

• answer sources

This lets you quickly see if a prompt works the way you want and if it’s important for your brand.

First check if the prompt really describes the question where your brand should be visible. If the prompt is too generic or badly matched, the results might be less useful.

How to read prompt visibility

Prompt visibility shows how often your brand appears in AI answers for a given question.

Example:

Visibility helps you assess whether AI associates the brand with a specific user need.

High visibility

High visibility means that AI often includes your brand in answers to a given question.

This can mean that:

• the brand is well connected with the given category

• the site responds well to this intent

• sources support the brand

• product data is well matched

• AI has enough information about the brand

Low visibility

Low visibility means that your brand appears rarely or doesn’t appear at all.

This may mean that:

• AI doesn’t associate the brand with this question

• the competition has stronger sources

• there’s not enough content answering this specific intent

• the brand description is too generic

• product data needs improvement

• the site may be harder for AI bots to access

Low visibility for an important prompt doesn’t always mean a technical error. It often means there’s a lack of clear enough content or sources that connect the brand with that question.

How to analyze the visibility chart?

The chart in the prompt details shows how the brand’s visibility has changed over time for a single question.

This way you can check whether the result:

• is growing

• is dropping

• is stable

• changes after implementing the recommendations

• differs between days

• reacts to new content or sources

What does an increase in visibility mean?

An increase can mean that the AI has started to connect your brand with a given question more often.

Possible reasons:

• implementing recommendations

• improving the content

• updating product data

• better website accessibility

• new sources appearing

• a change in the AI model’s answer

What does a drop in visibility mean?

A drop can mean that the brand shows up less often in answers for this prompt.

It’s worth checking:

• does the drop affect one model or multiple models

• did the competition gain visibility

• did the sources of answers change

• are the page and the product file available

• did new recommendations appear

In AI, result fluctuations are natural. What matters most is the trend over time, not a single day.

How to read brand position?

Position shows in which place your brand appeared in the AI answer if the model lists several brands, stores, products, or solutions.

Example answer:

1. Competitor A

2. Your brand

3. Competitor B

In this case, your brand has position #2.

How to interpret the position?

• #1 means the brand is the first recommendation

• #2-#3 means high visibility, but not full dominance

• #4 and lower means the brand is visible, but might be less noticeable

• no position means the brand wasn’t detected in the answer

The best result is a high position for prompts that are close to the purchase decision.

How to read prompt sentiment?

Sentiment shows the tone in which AI describes your brand in response to a given prompt.

It can be:

• positive

• neutral

• negative

Positive sentiment

AI describes the brand in a positive way, points out its advantages, or recommends it as a good solution.

Example:

Neutral sentiment

AI mentions the brand but without a clear evaluation or recommendation.

Example:

Negative sentiment

AI points out limitations, lack of data, problems, or a poor match to the question.

Example:

High visibility with neutral or negative sentiment may mean that the brand is known but isn’t described positively enough.

How to analyze AI model responses?

In the prompt details you can see real responses generated by AI models.

This is a very important part of the analysis, because the numbers show the result, but the answer shows the context.

In the model responses, check:

• whether the brand was mentioned

• whether the brand was recommended

• how the AI describes the brand

• whether the description is specific or general

• whether competitors show up

• whether the model points to sources

• whether the answer contains incorrect or outdated information

• whether the user could make a decision based on this answer

If AI mentions a brand but describes it in a very general way, it’s worth improving the brand profile, product data, or creating content that responds to that prompt.

How to compare AI models?

Different AI models can respond differently to the same prompt.

Your brand might be visible in one model but invisible in another. It can also have a different position or sentiment depending on the model.

It’s worth comparing:

• ChatGPT

• Gemini

• Claude

• Grok

• Google AI Overviews and Google AI Mode

Check:

• which model mentions your brand most often

• which model shows the competition

• does the sentiment differ between models

• are the sources similar or different

• does one model use information that the other one doesn’t use

Differences between models are normal. Each model can use different data, sources, and ways of generating answers.

How to analyze prompt sources

Sources show which websites and materials the AI uses when answering a given prompt.

In the prompt details, sources help you understand why the AI chose specific brands or recommendations.

Check:

• what domains show up in the answer

• is your website visible as a source

• do the sources support your competition

• are the sources up to date

• do the sources cover the right topic

• whether there are rankings, guides, or directories

• is your brand missing from important sources

Example:

Sources often show what needs fixing outside of the brand’s own site: presence in rankings, reviews, directories, guides, or industry media.

How to analyze competition in a prompt?

Prompt details help you check which brands show up for the same question.

Pay attention to:

• which brands show up most often

• which brand is in first position

• does the competition have better sentiment

• is the competition described more concretely

• what advantages of the competition does AI point out

• what sources does AI use for the competition

• does your brand appear together with the competition or is it skipped

If the competition shows up often and your brand doesn’t, it may mean a visibility gap.

The most important prompts are the ones where the competition appears often and your brand doesn’t show up at all or appears much lower in the ranking.

How do you spot a valuable prompt?

Not every prompt has the same business weight.

A valuable prompt is one that can influence the user's decision.

Such a prompt often includes:

• a specific need

• a product or service category

• a comparison of brands

• an intention to choose a provider

• a question asking for a recommendation

• information about the budget

• location

• the expected outcome

Examples of valuable prompts:

First analyze the prompts that are closest to the purchase decision. Those usually have the highest business value.

What to do if the brand doesn’t show up for a prompt?

If your brand doesn’t show up for an important prompt, it doesn’t automatically mean there’s an error. It’s a signal that the AI model doesn’t have a strong enough reason to mention the brand in that context.

Check:

• is the prompt well matched to the offer

• does the brand profile describe this category or service

• are the products imported and well described (online stores only)

• does the site contain content that addresses this topic

• is the AI using sources where your brand doesn’t appear

• does the competition have better sources

• does Semly show recommendations for this area

Possible actions:

• improve the brand description

• complete the product data (online stores only)

• add or improve the content on the site

• generate and publish articles in the Content Generator

• expand the knowledge base in Semly

• check the robots.txt file

• make sure you’re present in important sources

• add similar prompts for further analysis

Don’t fix everything at once. Start with the most important prompts where the brand should be visible but doesn’t show up in AI answers.

What to do if the brand appears in a low position?

If the brand appears in the answer but further down the list, it means the AI knows it but doesn’t treat it as the main recommendation.

Check:

• who shows up higher

• what advantages does AI point out for the competition

• does your brand have a concrete enough description

• do the contents show the differences compared to the competition

• do the sources support your brand

• are the product data complete

• does the AI answer contain outdated information

Possible actions:

• clarify the brand’s advantages

• expand comparison content

• improve product or service descriptions

• generate and publish 2 - 5 pieces of content for the selected prompt using the Semly Content Generator

• increase presence in sources

• implement Semly’s prompt-related recommendations

A lower position is often a good starting point for improvement. AI already associates the brand with the topic, but it needs stronger signals to show it higher.

What to do if the sentiment is neutral or negative?

If AI describes the brand neutrally, it may mean it has too little data about its advantages.

If the sentiment is negative, it’s worth checking whether the model is using outdated or unfavorable sources.

Check:

• how exactly AI describes the brand

• whether any accusations or limitations show up

• whether the answer contains outdated information

• what sources support the answer

• whether you clearly describe the brand’s advantages on the site

• whether reviews and external content are consistent with the brand positioning

Possible actions:

• improve the brand description

• add specific advantages

• create content that explains the offer

• update outdated information

• make sure you have better external sources

• prepare an FAQ that answers customers’ doubts

Sentiment shows not only whether the brand is visible, but also how it’s perceived in the AI’s answer.

How to analyze prompt details step by step?

If you don’t know where to start, use a simple framework.

Step 1: Read the prompt content

Check whether the question fits your offer well.

Step 2: Check visibility

See how often the brand shows up in the answers.

Step 3: Check the position

Assess whether the brand is one of the first recommendations.

Step 4: Check sentiment

See whether the AI describes the brand in a positive, neutral, or negative way.

Step 5: Read the models’ answers

Check the real context of the answers and the way the brand is described.

Step 6: Check the competition

See which brands appear instead of yours or next to it.

Step 7: Check the sources

See which domains the AI uses.

Step 8: Go to recommendations

Check if Semly suggests actions to take for this area.

The most common mistakes when analyzing a prompt

Avoid jumping to conclusions too quickly.

The most common mistakes are:

• analyzing only one AI model

• judging the result based on too short a period

• ignoring the content of the answer

• looking only at visibility without positions

• skipping sentiment

• not analyzing sources

• not comparing results with competitors

• focusing on prompts without business value

• not taking action after detecting a gap

The details of the prompt are most valuable when, after the analysis, you move on to concrete actions: content, data, sources, or recommendations.

What to do after analyzing the prompt?

After going through the details of the prompt, pick your next step.

You can:

• improve your brand profile

• improve product data or service descriptions

• create new content in the Semly Content Generator

• expand the knowledge base in Semly

• check competitors’ sources

• add similar prompts

• implement Semly’s recommendations

• monitor changes in visibility over time

Related articles

Summary

Prompt details show how your brand performs for one specific user question. This is the best place to analyze the reasons: why the brand is visible, why it’s skipped, why the competition wins, or why the sentiment isn’t good enough.

The most important rules:

• analyze the prompt in the context of the user’s intent

• check visibility, position, and sentiment together

• read real responses from AI models

• compare results between models

• analyze the sources of the responses

• check the competition

• look for prompts with high purchase intent

• implement recommendations that come from the analysis

Prompt details help you find specific spots where the brand can improve visibility in AI.