This guide will help you quickly get started with Semly and understand how to analyze your brand’s visibility in AI answers.
Semly shows whether your brand appears in the answers of AI models like ChatGPT, Gemini, Claude, Grok or Google AI. This lets you check for which questions you’re visible, how you compare to the competition, which sources AI models use, and what actions are worth taking to increase your chances of getting recommended.
The quick start consists of 4 steps:
Fill in your brand details
Check prompts and competition
Analyze the visibility report
Implement the recommendations
First, make sure that Semly has the basic information about your brand, website, and offering.
Project data is the foundation for later analysis. Thanks to this, Semly can better recognize your brand in AI answers, link it to the right domain, analyze your offer, and prepare more accurate recommendations.
Go to the Brand section and check if the most important information is correct.
First of all, check, correct, or complete:
brand name
logo URL
URL of the sitemap, that is sitemap.xml
industry
product or service category
target group
scope of operations
price level
detailed company description
business model
sales channels
delivery and returns policy
payment methods
The more precisely you describe the brand, the easier it will be to analyze its visibility and compare it with the competition.
Good practice: don’t describe the brand too broadly. Instead of writing “we sell home products”, write what products you offer, for whom, in which market, and what makes your offer stand out from the competition.
Example:
If you run an online store, go to the Products section and add a product file in Google Shopping XML format or use the ready-made integrations available in the menu in the TOOLS section under the Integrations tab.
The product feed lets Semly analyze not only the brand as a whole, but also specific products, categories, descriptions, prices, availability, and links to offers.
To add products from an XML file:
Paste the link to the Google Shopping XML file
Set the credentials if the file requires a login or password
Check the automatic update schedule
Click Import products
Good practice: make sure the file contains up-to-date names, descriptions, prices, categories, availability, and product URLs. Poor or outdated data can make later analysis and generating recommendations harder.
If your platform is available in the Integrations section, you can connect it directly to Semly. Available integrations can include, among others, RedCart, WooCommerce, Shopify, and Shoptet.
If you don’t see your platform, you can still use Semly via a Google Shopping XML file.
After filling in the project, check for which customer questions your brand should be monitored and who it will be compared with.
In Semly, prompts are crucial, meaning questions similar to those users ask AI models.
Go to the Prompts section and see the list of monitored questions.
In this section you can check, among other things:
prompt content
the average position of your brand
visibility
sentiment
assigned topic
prompt activity status
A prompt shouldn’t look like a classic SEO phrase. A prompt is a question or command from a user to AI that contains intent and context. Instead of a short phrase:
“living room furniture”
it’s better to use a natural question:
“I’m looking for a store where I can order modern, minimalist living room furniture with delivery and assembly included. Where’s the best place to look for it?”
Good prompts answer real customer questions and show their purchase intent.
Good practice: ask yourself: “For which customer questions should my brand be recommended by AI?” This is the best starting point for working with prompts.
If you’re monitoring a lot of prompts, you can group them into topics. This makes it easier to analyze brand visibility in specific areas, e.g. product categories, types of services, stages of the purchase decision, or customer segments.
Go to the Competition section and check which brands are being compared to yours.
In this view you can see:
number of mentions of your brand
number of mentions of the competitor
joint mentions
prompts where you both appear together
prompts where only your brand appears
prompts where only the competitor appears
This way you’ll quickly see in which questions the competition beats your brand.
Good practice: don’t just analyze the biggest competitors on the market. Also pay attention to brands that show up a lot in AI answers, even if they didn’t use to be an obvious competitor for you.
When the project, prompts, and competitors are ready, go to the Visibility report section.
This is the main place where you can check how your brand performs in AI answers.
In the report, you’ll see among other things:
the number of mentions of your brand
the number of mentions of all brands
the number of brands detected in AI answers
a chart of brand visibility over time
sources of AI answers
questions that customers ask in AI
brand position
answer sentiment
This data helps answer the most important questions:
Does AI know my brand?
For what kind of questions does my brand show up?
How often is it recommended?
In which position does it appear?
Does AI describe the brand positively, neutrally, or negatively?
What sources affect AI’s answers?
How does my brand compare to the competition?
Good practice: don’t judge visibility based on a single number. It’s best to analyze visibility, position, sentiment, sources, and competition together.
After clicking a specific prompt, you’ll see a detailed analysis of AI’s answers for that question.
In the prompt details you can check:
brand visibility over time
AI answer sources
latest model answers
information on whether the brand was mentioned
brand position in the answer
sentiment
answer date
AI model that generated the answer
This lets you see not only numerical data, but also the actual answer a user might get.
Example: if the model lists your brand in 2nd place but describes it in a very generic way, it might mean it’s worth filling in the brand data, expanding your content, or improving product information.
Sources show which domains, sites, and materials AI models use when generating answers.
In the Sources section, you’ll see among other things:
the total number of source occurrences
the main source in your industry
the biggest source gap
source share, i.e. SOV (Share of Voice)
the number of answers
the number of occurrences
your brand’s appearances in a given source
Sources help you understand why AI recommends certain brands and where it gets its information from.
Good practice: look for sources where your competitors show up often and your brand is absent or has a low share. These can be good places for further content, PR, or SEO activities.
After analyzing the report, go to the Recommendations section. This is where Semly shows specific actions that can improve your brand’s visibility in AI.
Recommendations can include things like:
improving the robots.txt file
creating an article on the suggested topic using our Content Generator
expanding your knowledge base
improving brand data
adding missing product information
increasing your presence in important sources
preparing content for specific prompts
First, complete the tasks marked as critical or high. They can have the biggest impact on further analysis and your brand’s visibility.
robots.txt file, it means it’s worth checking whether AI models can properly access important resources on the site.When you open a specific recommendation, you’ll see why Semly suggests a given action.
In the details you can check:
which prompt the recommendation refers to
for which question your brand is losing visibility
which competitor shows up more often
how many times your brand appeared
how many times the competitor appeared
what impact the recommendation might have on visibility
This way you know why a given task matters and what problem it helps solve.
Part of the recommendations may be about creating an article or other content that answers a specific customer question.
In that case, you can use the Content Generator.
The generator lets you:
create articles on suggested topics
edit the generated content
copy the material
prepare content for publication
build the brand’s knowledge base
Content created in the generator should answer real customer questions and support topics where the brand has low visibility or loses to the competition.
Good practice: don’t create content just for keywords. Create it for specific questions, intentions, and user problems that show up in prompts.
If you want to publish content in a separate space prepared for AI, you can set up a knowledge base on a subdomain.
Example:
knowledge.yourdomain.comTo launch it:
Enter the subdomain name, e.g. knowledge
Add an A record in the domain’s DNS panel
Point it to the IP address 46.248.186.77 provided in Semly
Save the changes with your domain provider
Click Verify and activate.
@ record if it handles the main website. Add a separate record only for the selected subdomain, e.g. knowledge.example.comAfter doing the basic steps, you can also check out additional sections.
In the Integrations section, you’ll find available connections with e-commerce platforms like RedCart, Shopify, or Shoptet.
If your platform isn’t on the list, you can still use import via Google Shopping XML.
Semly Club includes extra offers and benefits from partners. They can support running your store, handling payments, finances, or growing your online business.
In Settings you can check the active plan, limits, payments, invoices, platform language, and billing details.
After completing these 4 steps, you have the basic project setup and you can start working with Semly on a regular basis.
The best next actions:
check if the brand data is complete
make sure the products have been imported correctly
analyze the most important prompts
add or approve competitors
check the sources that support the competition
implement the highest-priority recommendations
generate and publish the first content for an important visibility gap.
AI visibility doesn’t change instantly. The first data shows your starting point, and steady improvement needs regular analysis, up-to-date data, good content, and actually implementing the recommendations.