The Review Generation Agent helps you consistently drive more reviews with minimal manual effort. Instead of running one-time campaigns, the agent works automatically in the background—choosing the right review sites, optimizing send times, and personalizing invitations at scale.

With this agent, you can:

• Target the best review sites for your business.
• Send requests at the optimal time based on customer events.
• Test multiple message variations and auto-switch to the best performer.
• Personalize content using tokens like customer name, location, or service.
• Automate reminder scheduling for higher completion rates.

The result: more reviews, better distribution across platforms, and stronger visibility in search.

Navigation

To access the Review Generation Agent:

Click ‘Settings.’ Go to ‘AI Agents’ and select ‘Review generation agent.’

You will be taken to the agent’s overview page, where you can see its ‘Goals,’ configure ‘Triggers,’ set up ‘Tasks,’ and ‘Outcomes.’

Goals

Here are the key goals of the agent:

• Save time and cost every week by automating review requests from customers.
• Get more reviews by improving request templates and optimizing delivery times.
• Improve your reputation on underperforming review sites with low ratings or fewer reviews.

Trigger

Go to the ‘Trigger’ tab. Triggers define how often the Review Generation Agent should optimize. Once configured, the agent will dynamically adjust review request timing, sites, and templates whenever triggered.

You can set triggers to run monthly, choosing between the 1st and 15th of the month.

Alternatively, you can set triggers to run weekly, choosing any weekday (Monday through Sunday).

This ensures your agent stays updated regularly, optimizing delivery times, templates, and site selection based on performance data.

Tasks

Configure the set of tasks that help optimize review collection and improve performance.

Choose the most relevant review sites

Click the ‘Review site selector’ to pick the platforms most relevant for your business.

Select the sites you want the agent to include. You can pick up to four sources.

Decide how sites should be prioritized:

• Manual – Adjust the weightage for each site manually. Make sure the total adds up to 100%.

Auto – Let AI automatically select and prioritize review sites based on past performance. The agent identifies underperforming platforms and prioritizes them per request based on reviews, ratings, recency, and click-through rates, and shows up to four in each request. Click ‘Save’ to apply your preferences. Ensure to click 'Save' after you make changes in each task.

Run an A/B test on templates

This task allows you to improve review request performance through automated A/B testing of templates. The agent creates a test template inspired by the main template that you select and identifies which one converts better. If the test template performs better, it is automatically set as the new main template, and the agent continues testing with fresh variations.

How it works:

Set up templates

• Choose a main template to create a test template to run performance tests.

Automatic comparison

• The agent compares both templates to see which converts better.
• If the test template shows higher conversion, it is promoted as the new main template.

Continuous optimization

• New test templates are generated to ensure ongoing optimization.
• If results are unclear or insufficient, no changes are made.

Configuring the review request optimizer

Click ‘Review request optimizer.’

Channel selection:

Click the dropdown button and choose the channel to send the review requests.

Template selection:

Click the dropdown button to select the email template and then the text template that the agent should optimize.

Personalization tokens:

Click on the tokens and add tokens like business name, average rating, review count, etc.

Tone of voice:

Pick a tone of voice for the test template to be generated by the agent.

Sample size for A/B testing:

Select the percentage of review requests to be used for testing (10%, 20%, 30%).

Example: If you set 10%, then 90% of review requests are sent using your original template, and 10% are sent using the agent’s test template. The agent then analyzes conversion results to decide if the new template should replace the original.

You can also add prompts to provide personalized instructions to the agent by clicking ‘Add prompt.’

Enter the prompt and click ‘Add.’

Find the best time to send review requests

The agent uses past performance data to determine when review requests are most likely to convert. By automatically testing and adjusting send times, the agent ensures your messages reach customers at the right moment for maximum response.

Click ‘Optimal send time.’ A configuration panel opens with multiple options:

• Auto: Lets the agent choose the best send time automatically, based on the last 30 days of performance data.

• Send immediately: Sends the review request right after the trigger event.

• Delay by: Allows you to delay requests by a set time — choose from hours, days, weeks, or months.

If data is limited, the agent will automatically test different time slots to identify which delivers the best results. The system also prioritizes top-performing windows while experimenting with low-traffic times to uncover new opportunities.

Select recipients for the review requests

The agent allows you to choose the right audience for sending review requests, ensuring requests are targeted to the most relevant recipients.

Click ‘Target audience’ to configure your recipients.

You’ll see two main configuration options:

Sources of recipient data:

• Choose ‘All’ or from available options like Bulk Import, Check-in, SFTP, Integration, and API.

When should the agent send the review request?

Choose a trigger for when requests will be sent. Scroll down in the dropdown box to view more triggers.

Identify the best time to send reminders

The agent also helps optimize reminder emails so that customers who haven’t responded to the first review request can be nudged at the right time.

Click ‘Reminder manager’ to configure reminder settings.

In the Reminder manager window, you’ll see two modes:

• Auto: Lets the agent decide when to send reminders, how many reminders to send, and which subject line to use—based on performance data.

• Manual: Allows you to define your own reminder settings, including subject line, number of reminders, and interval between reminders.

If you select Manual, you can:

• Enter a custom subject line (e.g., Reminder: Review us!).
• Choose the number of reminders to send (e.g., 1, 2, or more).
• Set the interval between reminders (e.g., 3 days after the initial request).

Run Test

After you’ve configured the tasks, click ‘Run test’ in the preview section to see how the agent will function in real-time.

The results shown here are directly aligned with the tasks you’ve configured. Since there are 5 tasks in total, you would normally see 5 results.

Choose the most relevant review sites

The agent analyzes review volume and performance of the last 30 days from the review sites that were selected.

Run an A/B test on templates

The agent selects the main template and generates another variant to test.

• Main Email template: Review request email – Experience Matters (90% usage). Click ‘Text’ to view the text main and test template.

Here’s how the main email template appears.

Scroll down in the preview section to view the test email template generated by the agent.

Difference between Main and Test Template

• Split usage: The main template gets 90% traffic, the test template 10% for comparison.
• Tone: Main uses a formal tone while the test uses a value-based tone.
• Site priority: Test template moves Apartments and Birdeye above Google to boost reviews on underperforming sites.

Find the best time to send review requests

The agent recommends an optimized send time.

Example:

• Send time: 6th Sep, 10:00 AM
• Channel: Email and text
• Status: Optimization applied

Select recipients for the review requests

Shows the audience who will receive review requests.

Identify the best time to send reminders

The follow-up reminder was set manually, hence it shows:

• Reminder date: 9th Sep, 10:00 AM
• Status: Reminder pending
• Email subject: Reminder: Review us!

After reviewing the results, pause the agent, and then you can publish the agent.

Outcomes

Outcomes measure how well the agent is performing against the goals.

Contacts reached – Total number of unique contacts who received a review request. Total unique contacts who received at least one review request via channel. A contact is counted once, even if they received multiple requests.

Click-through rate – Percentage of unique contacts who clicked at least once on a review request received across email and text.

Review sites clicked – Total number of clicks on review platforms from review requests.

View Agent Performance In Campaigns

To track your agent’s performance, navigate to ‘Campaigns’ and open the ‘Split automations’ tab. Select the campaign to view details.

Since the campaign is still running, performance metrics may not yet be available. However, you can:
View details of the selected campaign.

• View recipients who received review requests.
• View reports once data is available
• Use the dropdown menu to switch between the main template and the test template for comparison.