INTELLIGENCE & AUTOMATION
AI for commerce decisions and workflows
Detect unusual changes, run configured workflows, and review suggested next steps.
Intelligence for commerce workflows
Detect, configure, and review. Use available commerce data to guide your next step.
Anomaly detection
Spot unusual changes in commerce metrics using available analytics data and historical baselines.
Metric comparisons•Severity indicators•Confidence scores
Investigate unusual changes
Workflow automation
Run configured actions when events occur, with retries for failed workflow jobs.
Event triggers•Retry handling
Connect events and actions
Suggested next steps
Review explanations and suggested actions for anomaly alerts. AI-generated guidance falls back to template guidance when the AI service is unavailable.
Alert context•Action suggestions
Review before acting
Explore illustrative AI workflows
These animated examples use fictional data and scripted responses. They explain possible workflows, not live store activity, product availability, or measured results.
AI assistant · illustrative example
Show me products with low inventory across all locations
Example: 12 products are below their reorder points. SKU-4521 has 3 units, SKU-2890 has 5, and SKU-7713 has 7.
Example: welcome a new customer
Simulation
Example trigger: an order is created
Example check: first-time customer
Example action: send a welcome email
Example action: apply a customer tag
4
Example steps
1
Sample workflow
Demo
Execution mode
Why Tanqory Intelligence
Commerce context for AI
Use commerce metrics and anomaly context to generate explanations and suggested next steps.
Signals to help you investigate
Severity and confidence scores help prioritize unusual changes. Review the underlying data before acting.
Retry and error handling
Workflow jobs retry after failures and can trigger error workflows when retries are exhausted.
Frequently asked questions
What does the AI do?
Detect unusual changes in commerce metrics, run configured workflows, and review explanations and suggested actions for anomaly alerts.
Do I need ML expertise?
You do not need to train an ML model. Detection depends on available analytics inputs and sufficient historical data.
What do the alert scores mean?
Severity and confidence scores reflect statistical deviations from historical metrics. They help prioritize investigation; they are not guarantees of forecast accuracy.
How do I respond to suggestions?
Review suggested actions before deciding how to respond. Alerts can be acknowledged or marked resolved.
What data is used for explanations?
Anomaly explanations use commerce metrics and historical comparison context sent to the configured AI service.