How to read Semly recommendations

Recommendations in Semly are ready-made action suggestions prepared based on the analysis of your brand, prompts, sources, competitors, AI model responses, and project configuration.

Their goal is to show what’s worth improving, adding, or implementing to increase the chances of your brand appearing in AI answers.

What’s important is that Semly doesn’t just show a general suggestion. In many cases you get a ready-made solution that you can implement automatically in Semly, or copy and paste into the right place on your website or in your technical configuration.

What are Semly recommendations?

Recommendations are specific tasks that help improve your brand’s visibility in AI.

They can refer to, among other things:

• technical configuration of the website or online store

• robots.txt file

• llms.txt file

• sitemap

• brand data

• product data

• content on the site

• knowledge base

• sources

• prompts

• competitors

Each recommendation points out a problem or opportunity and shows what action is worth taking.

Recommendations help you move from analysis to action. Instead of looking for the cause of low visibility on your own, you get a concrete task to do.

Why are recommendations important?

The report itself shows what’s happening with your brand’s visibility. The recommendations show what you can do about it.

Thanks to the recommendations, you can:

• improve your site’s accessibility for AI bots

• organize your brand data

• fill in missing information

• create content for important customer questions

• expand your knowledge base

• improve product data

• increase your chances of being recommended in AI

• reduce gaps compared to your competitors

• react faster to technical issues

Semly brings the most value when you regularly implement the recommendations and track how visibility changes over time.

Where can you find the recommendations?

To go to the recommendations:

1. Log in to the Semly panel

2. Choose the right project

3. Go to the Recommendations section

4. Check the task list

5. Open the details of the selected recommendation

6. Perform the action automatically or copy the ready-made solution

In the Recommendations section you can see among other things:

• recommendation title

• description of the problem or opportunity

• priority

• status

• action details

• ready-made solution

• action execution button if available

• manual implementation instructions

What types of recommendations can you see?

Semly can generate different types of recommendations, depending on the data detected in the project.

The most common types of recommendations are:

• technical

• content-related

• product-related

• brand-related

• source-related

• prompt-related

• competitors-related

• knowledge-base-related

Technical recommendations

These concern elements that can affect data availability for AI models.

Examples:

• fix the robots.txt file

• add or update the llms.txt file

• fix or add a sitemap

• add a link to the knowledge base

• make sure AI bots can read the important pages

• fix or add schema.org structured data

Content recommendations

These are about articles, descriptions, FAQs and content that answers customer questions.

Examples:

• create an article on the given topic

• expand the content for a specific prompt

• add an answer to a frequently asked question

• prepare a comparison guide

• describe the brand’s advantages in a given category

• create a knowledge base for selected topic gaps

Product recommendations

They concern the quality of product data and how well the offer matches customer questions.

Examples:

• improve product descriptions

• fill in missing categories

• add better product names

• check product availability

• update the Google Shopping XML file

Recommendations about sources

They concern the places that AI can use when creating answers.

Examples:

• check the source that supports the competitors

• make sure you’re present in an important directory

• prepare content similar to a source often used by AI

• analyze the ranking where the competitors appears

Not every recommendation requires a technical action. Some recommendations are about content, data, or brand presence in sources that affect AI answers.

Automatic and manual recommendations

In Semly, some recommendations can be done automatically, and some require manual implementation outside the panel.

Recommendations done automatically

You can do some actions directly in Semly.

This can apply, for example, to:

• generating an article

• preparing content in the Content Generator

• creating content in the knowledge base

• marking the recommendation as completed

• preparing a ready-made file or a piece of content

• running AI Agents actions available in the panel

If a recommendation can be done automatically, you’ll see the right button or action in its details.

Example:

Recommendations that require manual implementation

Some recommendations require copying the ready-made solution and pasting it in the right place outside Semly.

This may apply for example to:

• the robots.txt file

• the llms.txt file

• the sitemap

• DNS settings

• a link in the page footer

• a link in the menu

• changes in the CMS

• changes in the shop panel

• updating the content on the site

Example:

Semly can prepare the ready content of the llms.txt file, but you or a tech person have to paste it on the site in the right place.

If a recommendation needs to be implemented manually, read the instructions carefully and make sure the ready solution has been pasted in the right place.

How to read the recommendation priority

Priority helps you decide where to start.

In Semly you can see recommendations with different levels of importance.

Most often:

• critical

• high

• medium

• low

Critical priority

Means an action that can block or significantly limit the analysis and visibility of the brand in AI.

Examples:

• AI bots have limited access to the site

• the robots.txt file blocks important resources

• a key sitemap is missing

• important data isn’t available

• the setup makes the site hard to read

It’s worth doing these recommendations first.

High priority

Means an action that can have a big impact on the brand’s visibility.

Examples:

• creating content for an important prompt

• improving product data

• expanding the knowledge base

• improving the brand description

• filling in important sources

Medium priority

Means an action that is important, but usually doesn’t block the basic analysis.

Examples:

• polishing the content

• expanding the FAQ

• improving a less important category

• filling in additional data

Low priority

Means a supporting or optimization task.

Examples:

• small content tweaks

• additional data expansion

• tidying up less important elements

If you don’t know where to start, follow the recommendations from the highest priority: critical, high, medium, low.

What’s in the recommendation details?

When you open a recommendation, you’ll see more info about why Semly suggests a given action.

The details can include:

• a description of the problem

• an explanation of the impact on visibility

• a ready-made solution to copy

• a button to run an automatic action

• instructions for manual implementation

Details of the recommendation show not only what to do, but also why a given action can improve your brand’s visibility in AI.

How to use the ready-made solution?

In many recommendations, Semly prepares a ready-made solution that you can use right away.

For example, this could be:

• ready article content

• ready file llms.txt

• ready sitemap

• ready robots.txt file

• ready brand description

• ready list of recommendations

• ready content to paste on the website

• ready FAQ structure

• ready suggestions for product descriptions

If the recommendation contains a ready-made solution:

1. Open the recommendation details

2. Read the description of how it works

3. Check the ready-made content or configuration

4. Copy the prepared solution

5. Paste it in the right place

6. Save the changes in your system

7. Go back to Semly and mark the recommendation as completed if that option is available

Ready-made solutions save time and reduce the risk of errors, because you don’t have to prepare all the content or configuration from scratch yourself.

Example: robots.txt file recommendation

If Semly detects that the robots.txt configuration may make it harder for AI bots to access important website resources, it can prepare a technical recommendation.

In such a recommendation you might see:

• problem description

• an explanation of why robots.txt matters

• the entire proposed file content ready to use

• instructions on where to paste the change

• info on what to check after deployment

Example action:

1. Copy the ready-made content prepared by Semly

2. Open the robots.txt file of your site

3. Paste the copied content

4. Save the file

5. Check if the file is available at https://yourdomain.com/robots.txt

6. Mark the recommendation as done or run the test by clicking Check again

Before changing the robots.txt file, make sure you’re not removing important rules needed for Google or other search engines.

Example: llms.txt file recommendation

The llms.txt file is a simple text file placed in the root directory of the domain that helps AI systems understand what the site does and which resources are the most important. It can include a short brand description, links to key sections, sitemaps, knowledge bases, and other materials that are worth sharing with AI models.

If Semly generates a recommendation regarding llms.txt, you can get a ready-made file content.

In such a recommendation you might see:

• a ready description of the site

• links to the most important sections

• information about the language and market

• instructions for publishing the file

Example workflow:

1. Copy the ready-made llms.txt content

2. Create a file llms.txt in the main directory of the domain

3. Paste the content prepared by Semly

4. Save the file

5. Check if it works at https://yourdomain.com/llms.txt

6. Mark the recommendation as done

If you don't have access to the site files, pass the ready-made content to a technical person or the store administrator.

Example: content recommendation

If Semly detects a prompt where the competitors is visible and your brand doesn't appear in the answers, it can suggest creating content.

The recommendation can indicate:

• article topic

• prompt the gap refers to

• competitor that appears more often

• reason for preparing the content

• potential impact on visibility

• option to generate an article in Semly

Example workflow:

1. Open the recommendation

2. Check the prompt and the reasoning

3. Click the option to generate an article if it’s available

4. Review the generated content

5. Edit it if needed

6. Publish the content in the knowledge base or on the website

7. Monitor visibility changes in the next reports

It’s best to create content for specific customer questions, not just for general keywords.

How to mark recommendations as completed?

After implementing a recommendation, it’s a good idea to mark it as completed if the panel offers such an option.

This helps organize your work and separate finished tasks from those that still need action.

Marking a recommendation as completed helps you:

• track progress

• avoid repeating the same actions

• keep your task list tidy

• come back later to analyze the results

• manage your team’s work better

If a recommendation needs to be implemented outside of Semly, first make the change in your system, and only then mark it as completed.

How to measure the impact of implementing recommendations?

The impact of a recommendation doesn’t always show up right away. AI models and systems that use sources may need some time to take new data into account.

After implementing the recommendations, check:

• brand visibility

• position for the related prompt

• sentiment of the responses

• sources of the AI response

• presence of competitors

• change over time

• new recommendations

It’s best to compare the results before and after implementation in the same area, e.g. for the same prompt, topic, or source.

Don’t judge the impact of a recommendation after just one day. In AI, the trend over time matters more than a single result.

What to do if the recommendation is technical?

Some recommendations require access to site files, DNS, CMS, or the store panel.

If you don’t have technical access, pass the ready-made solution to the person responsible for the site.

It’s worth passing on:

• content of the recommendation

• ready-made solution from Semly

• implementation location

• short description of the goal of the change

• link to the help article

• screenshot of the recommendation

Sample message to a technical person:

If you’re not sure how to implement a technical recommendation, get in touch with the Semly support team or the person managing your website.

What to avoid when implementing recommendations?

The most common mistakes are:

• ignoring critical recommendations

• doing only easy tasks without analyzing priority

• copying content to the wrong place or to several places (Duplicate content)

• marking a recommendation as done before implementation

• not checking how it works after implementation

• removing existing robots.txt rules without analysis

• publishing content without reading and editing it

• implementing many changes at once without monitoring the effect

• skipping recommendations about sources

• not comparing the results before and after implementation

Don’t implement technical changes blindly. If a recommendation concerns files like robots.txt, sitemap or llms.txt, make sure you understand where to add it or pass it to a technical person.

How to work with recommendations regularly?

It’s best to treat recommendations as a current list of actions that improve your brand’s visibility in AI.

A good work rhythm can look like this:

• once a week, check new recommendations

• start with critical and high priority

• carry out the most important technical actions

• pick 1 to 3 content-related recommendations

• generate or prepare the content

• mark completed tasks

• check the results after a few days or weeks

• compare the change in visibility in the report

Working with recommendations regularly is more effective than implementing many changes at once without later analysis.

Related articles

Summary

Semly recommendations show specific actions that can help improve your brand’s visibility in AI answers.

The most important rules:

• start with the highest-priority recommendations

• check the details and justification of the recommendations

• use ready-made solutions prepared by Semly

• carry out some actions automatically in the panel

• copy some actions and implement them manually on the site or in the configuration

• once implemented, mark the task as done

• monitor the results in the visibility report

• don’t ignore technical recommendations

Semly not only points out the problem, but often prepares a ready-made solution. Your job is to implement it automatically in the panel or copy it and add it in the right place.