Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.shinecx.com/llms.txt

Use this file to discover all available pages before exploring further.

When you mark a job as complete, ShineCRM automatically drafts personalized review request SMS messages for each platform the customer has not yet reviewed on. The drafts are scheduled as pending follow-ups, and you review them before they send.

​
When it runs

Two triggers:
  1. Job completion. When a job’s status flips to complete, the AI drafts review requests for any of Google, Facebook, or Yelp that you have a review link for, in your configured platform priority order, and that the customer has not already reviewed on.
  2. Customer leaves a review. When you flip the reviewed_google, reviewed_facebook, or reviewed_yelp flag to true on a contact, the AI proposes cancelling any scheduled review requests on that platform so you don’t pester them after they’ve already left a review.
Both paths only run when review requests are enabled under Settings → Reviews.

​
What the AI sees

For each completed job, the AI is given:
  • The customer’s name
  • The job address
  • The list of services performed on the job
  • The technician’s notes from the job
  • Your configured tone, max attempts per platform, and gap days between attempts
  • The review links for each unreviewed platform

​
What it produces

A JSON array of drafts, one per platform per attempt. Each draft becomes a row in your scheduled messages list with:
  • Channel: SMS
  • Status: pending (waiting for your approval)
  • Body: a short, warm SMS under 320 characters that includes the review link
  • Scheduled time: starting tomorrow, with your configured gap between attempts
The AI is instructed to reference something specific from the job when possible (a service performed, a note from the tech) instead of generic boilerplate. It will not invent details.

​
What you control

Under Settings → Reviews you can configure:
  • Whether review requests are enabled at all
  • Review links for Google, Facebook, and Yelp
  • Platform priority order (which platform the AI tries first)
  • Max attempts per platform
  • Gap days between attempts
  • Tone (e.g. “friendly, concise, grateful”)
  • The system prompt the AI uses for drafting (advanced)

​
How review-completion cleanup works

When you mark a customer as having reviewed on a platform, the AI looks at their pending review request follow-ups for that platform and proposes cancellation. The cancellation proposal lands in Settings → Command Centre as a pending suggestion. You approve or reject it. Cancellations are not automatic.

​
What it cannot do

  • Send the review request itself (your normal scheduled-messages send cycle does that)
  • Send a review request for a platform you have not configured a link for
  • Send a review request to a contact who has no phone number
  • Re-request a review on a platform the customer has already reviewed on