Practical guides for the core resenIA workflow.
Create manual reviews, filter the inbox and keep track of replies from draft to publication.
Give resenIA the business context it needs before asking for consistent replies.
The analysis page turns enough diagnosed reviews into trends, risks and suggested actions.
Corrections, approvals and feedback become reusable guidance for future drafts.
A short path to understand resenIA without configuring everything at once.
Add business name, category, tone and preferred language so the first drafts already have a clear personality.
Use two or three recent reviews with different ratings to see how resenIA adapts the response strategy.
Treat AI output as a draft. Adjust facts, promises and tone before marking the reply as approved.
Once there is enough activity, use Analysis and Learning to understand patterns and improve future replies.
Use statuses to avoid losing track of what still needs a human decision.
Before publishing, confirm that the reply is useful, specific and safe for the brand.
Some reviews need extra care because they mention legal, health, safety or reputation-sensitive issues.
The profile is not a marketing brochure. It should give the AI practical rules, facts and boundaries it can reuse.
Opening hours, service area, usual tone, escalation policy, refund boundaries and frequent customer concerns.
Vague instructions like be nice, sound professional or answer quickly without explaining what that means for the business.
For negative reviews, apologize for the experience, mention the specific issue, invite the customer to contact the manager by email, and avoid offering compensation in public.
Analysis is most useful when it becomes part of weekly operations, not only a dashboard to read once.
The system learns best from clear human decisions repeated over real examples.
Add more concrete guidelines, business facts, things to avoid and examples of the tone you expect.
Change the default tone, add one or two style rules, and mark bad drafts with feedback so learning has a signal.
Add more reviews. Analysis needs enough diagnosed material before it can produce reliable patterns.
The interface only shows the flows active for your account. If you need help, check the FAQs or contact support.
Start with configuration, add two or three real reviews, then revisit analysis and learning when there is enough signal.