AI Review Assistant helps customers write reviews when they do not know what to say. Instead of dropping them onto Google or Facebook and hoping they can think of the right words, the assistant guides them through a short set of questions and then drafts a review from their answers.
That usually leads to more completed reviews, more specific reviews, and less abandonment during the review-writing step.
What the AI Review Assistant is
The assistant lives inside Feedback Forms. After the customer has already entered the review flow, it can open a friendly, chat-like sequence of questions and then generate a review draft they can edit before posting.
This is not about inventing experiences. It is about helping a real customer express a real experience more easily.
What you need before setting it up
- A working AI provider, either OpenAI or OpenRouter
- The AI Assistant Model configured on that provider
- At least one feedback form for the organization
- The customer plan must allow AI Review Assistant access
The AI Assistant Model is different from the Review Response Model. Review Response is for replying to existing reviews. AI Review Assistant is for helping customers write new ones.
Plan access matters
The feature is controlled through custom plans. In the plan editor, the AI Review Assistant sits under the feedback form tools area. If feedback forms are hidden, the assistant is effectively hidden too.
- Enabled gives full access
- Upgrade keeps it visible but locked behind an upsell
- Hidden removes it from view
Enable it on the feedback form
Go to `Grow -> Review Tools -> Feedback Forms`, open the form builder, find the AI Review Assistant option and turn it on. From there you configure the experience that the customer will actually move through.
This is a customer-side screen. Because you are in your agency account, you can reach it two ways: go to `Business -> Customers -> Manage Customers` and click `Open dashboard` on the customer — you go into their dashboard with nothing locked, even features their plan hides, and a bar at the top shows whose dashboard you are in (handy for done-for-you setups or granting access beyond a plan); or choose `View as customer` on the same page to see their dashboard exactly as they see it, with their plan's limits. If your sidebar shows the `Agency` / `Customer` switcher, the `Customer` tab plus the location switcher is a shortcut to the same place.
Write strong business context
Business context is one of the biggest quality levers. The assistant uses that description to understand what the business actually does, what kind of customer experience it delivers, and what details belong naturally in a review.
Vague business context produces vague reviews. Specific business context produces reviews that sound more believable and more useful.
- Describe what the business does
- Include specialties or signature services
- Add useful setting or audience context
- Avoid one-word or generic descriptions
Design the questions carefully
The question set is the heart of the feature. The answers become the raw material the AI uses to draft the review, so the question design matters more than most teams expect.
The practical sweet spot is usually 3 to 5 questions. Enough detail to guide the AI, but not so many that the customer loses patience.
- Use the auto-generate button when you want the system to propose niche-aware questions
- Use text or textarea questions when you want richer customer language
- Use radio buttons and checkboxes when you want faster completion and more structured answers
- Mix question types instead of making everything text-heavy
Use images and suggestion chips when they help
Image options and suggestion chips can dramatically improve completion rate, especially on mobile. They reduce thinking effort and make the form feel more like guided input than work.
This is especially powerful in service businesses where the customer may recognise a staff member or a product visually faster than they can type it.
- Use team photos for staff-identification questions
- Use product or service imagery where that helps the customer choose faster
- Use suggestion chips generously on mobile-heavy forms
Match the tone to the business
The writing tone should feel natural for the business type and customer base. A mismatch here makes the review feel off even if the content is technically correct.
| Tone | Best fit |
|---|---|
| Casual (recommended) | Conversational, contractions, reads most natural |
| Neutral | Balanced everyday language |
| Enthusiastic | Higher energy |
Use optional settings with restraint
Keywords are used sparingly on purpose — roughly one generated review in three includes them, and only one or two at a time; emojis appear in about one in five when allowed. That is deliberate anti-pattern protection, so do not add keywords expecting them in every review.
- Allow emojis only if that fits the brand voice
- Add keywords only when they should appear naturally
- Do not force jargon or service terms into every review
What the customer sees
The customer experience usually unfolds in three stages: answer a few guided questions, review the AI-generated draft, then copy or edit it before publishing to the review platform you configured.
This is why the feature works. The blank-page problem disappears because the customer is no longer starting from nothing.
- Chat-style question flow
- Editable AI-generated draft
- Publish step with the selected review platforms
Language support
The assistant follows the language configured for the form and adapts the generated writing accordingly. This is especially useful in multilingual markets where different forms may need different language behavior.
Best practices
- Write detailed business context
- Keep the question count tight
- Mix question types instead of relying on one style
- Use suggestion chips generously
- Match tone to audience
- Use staff photos when the service is person-led
- Encourage light customer editing before posting