Why Static Routing Becomes Inefficient
When merchants rely only on fixed routing logic:
- Approval rates fluctuate as issuer behaviour changes
- Processor performance varies by BIN and country
- Brand-level routing (e.g. “all Visa to processor X”) becomes overly simplistic
- Cascading logic does not adapt to real approval probabilities
Modern payment optimisation requires dynamic, data-driven decision-making.
How AI Routing Maximiser Improves Results
For every transaction, AI/ML evaluates a range of contextual signals — card data, transaction context, historical approval performance, previous routing outcomes, and merchant-specific payment patterns — to select the processor with the highest predicted approval probability from those defined within the routing rule.
There is no universal "best processor" — routing decisions are made per transaction, based on real context.
Continuous Learning with Built-In Exploration
Processor performance evolves over time. To remain accurate, AI Routing Maximiser includes controlled exploration:
- A small percentage of traffic is intentionally routed to alternative processors
- Performance changes are detected early
- New processors can be evaluated safely under real transaction traffic
Meaningful optimisation typically begins after ~1,000 transactions
Before that threshold, AI/ML relies on pooled data across merchants and controlled exploration to stay effective.
Full Control Through Routing Rules
AI Routing Maximiser operates strictly within merchant-defined routing logic.
- Enable AI/ML within specific routing rules
- Define which processors AI/ML can select from
- Control the percentage of traffic routed through AI/ML
- Position AI/ML within the rule priority hierarchy
AI/ML always respects rule priority and selects only among the processors you've defined. The rest of your traffic keeps following static routing logic.
Intelligent Retries for Decline Recovery
If the initial attempt was AI/ML-driven, cascading stays AI/ML-driven — even with "AI/ML Cascading" disabled. If the initial routing was static, AI/ML cascading applies only when:
For all retries, AI/ML selects only from processors defined within the routing rule.
End-to-End Merchant Benefits
Higher Approval Rates
More transactions approved through context-aware routing decisions.
Improved Stability
Performance automatically adapts to issuer and processor behaviour changes.
Smarter Cascading
Retries follow predicted approval probability, not fixed processor order.
Processor-Agnostic Optimisation
AI/ML operates across all processors configured within UpGate.
Transparent Governance
Merchants activate AI/ML within existing routing rules. No additional integration is required.
Reduced Operational Overhead
The routing engine adapts automatically, saving analysts from manual rule updates.
How to Enable AI Routing Maximiser in UpGate
AI Routing Maximiser can be enabled and configured directly in the UpGate back office. Merchants can manage AI/ML settings within their routing configuration, including:

Tell us about your company
Get a custom setup designed for your business model and scale.





