Customer service that knows the cost.

Xolus turns every after-sales conversation into a controlled decision: policy, cost, approval, and outcome.

Xolus is an operations layer for commerce service teams, designed to turn repetitive conversations into consistent decisions, measurable outcomes, and calm support.

Built first for Douyin merchants, Xolus connects customer intent with product rules, order context, supplier policies, and refund limits.

The result is not an AI chatbot. It is a controlled service workflow.

01

Understand

Classify intent, issue, and risk from the live conversation and order context.

02

Calculate

Apply product rules, supplier policies, and refund limits to set a financial boundary.

03

Approve

Your team reviews the recommendation and approves the final action.

04

Learn

Every decision refines the rules, so the next case is faster and more consistent.

Solutions

A service desk with financial memory.

Every recommendation is attached to a rule, a cost, and an approval state.

Xolus · Decision ConsoleLive

Case Queue

Decision Panel

Customer message“The corner arrived scratched.”
IssueSurface damage
RiskLow
Evidence requiredPhoto of damage
¥15Offer¥23Ceiling¥46Return loss

Suggested reply

Apologize, request evidence, and offer partial compensation within the approved range.
The model does not send refunds.Rules calculate.People approve.
Results

Measured in seconds, adoption, and yuan.

0sAverage decision time
0%Recommendation adoption rate
¥0Monthly leakage reduced
0%Return pressure down
Security & Control

Your team keeps control of every refund, every message, and every exception before anything reaches the customer.

Xolus is designed for assisted operations. Models understand context. Rules calculate boundaries. Humans approve the final action.

AI

Understands context

Rules

Calculate boundaries

Human Approval

Final action authorized

Customer

Receives one clear answer

Pilot Program

100 real cases. Real decisions. No guesswork.

We model your products, costs, supplier policies, and support rules, then test Xolus against real after-sales conversations from your store.

01

Import historical support conversations

Recent cases become the ground truth for the pilot.

02

Map products

Each SKU is linked to its cost, margin, and known failure modes.

03

Configure supplier policies

Compensation ceilings and evidence requirements are encoded as rules.

04

Run simulations

Xolus processes every case as if it were live, without touching a customer.

05

Compare decisions

Your team reviews Xolus recommendations side by side with actual outcomes.

Pilot Application

Tell us about your operation.

FAQ

Questions, answered plainly.

Orbit complete

Ready to make every service decision measurable?