A member is not at risk because they skipped a class. They are at risk when their own pattern breaks: the regular Tuesday gone quiet, visits halving, bookings drying up. Xyzios reads that pattern for every member every night.
Signals, from each member's own rhythm
- Days since the last visit
- Visits down month over month
- No-shows and late cancels
- A new member with low engagement
- No future bookings
- A failed payment
Scored every night
0–10
weights learn from your own cancellations once there are enough of them
Where the score puts them
- 3 or moreon the list
- 6 or moreHigh
- 7 or morename shown in red
Nothing here contacts a member. A save is drafted into Approvals, or someone calls from the list and marks it Handled.
What the score is for
It orders a list, it does not judge a person. The top of the list is who to call first; the signal beside each name says why, in plain words; the suggested move says what usually works.
What happens next
- Your AI team drafts a save into Approvals for members who slipped recently, members whose payment failed, and brand-new members who never came in. Everyone else stays on the list for a personal call.
- A call made from the list counts once someone taps Handled. The row remains, marked with who and when.
- The default view hides people more than 30 days out, because on a studio’s first day the list is mostly people who left long before Xyzios saw them. Include long-lapsed shows them.
Where you meet it
At Risk is the list itself. Data counts the members at risk by location, and the Overview’s priorities often start with them.