01
Project
Choose or create your Anomaly Detection project.
A scope for data drafts and saved scoring history
- Scope
- Anomaly project
- History
- Saved scoring runs
- Selection
- Choose intentionally
MAG OptiAIOperational AIAnomaly DetectionEarly Access
Spot Unusual Patterns. Review Them in Context.
Find unusual points in your operational series and review them in context. Prepare the data, choose a detector, run scoring and inspect the flags and explanations before deciding what needs attention.
Prepare the series. Choose the detector. Review what the flagged points mean.
Prefer to explore first? Start free14-day trial · lower limits. Beta access requires approval.
Built for planning teams
Series data and unusual points
Cloud-based Pro Console
Guided evaluation
The anomaly challenge
An Unusual Point Needs More Than a Flag.
Data quality, the surrounding pattern and detector settings shape which observations deserve review.
Challenge 01 / 04
Data Issues Can Look Like Operational Issues
An unusual point deserves a look at the input as well as the outcome.
Dates, values and optional contextual inputs need to represent the series correctly. Missing or inconsistent records can change how a result should be interpreted.
Impact
Review data readiness before drawing conclusions from flagged points.
The challenge is more than highlighting a spike.
It is reviewing unusual points with their data, detector settings and operational context connected.
Why MAG OptiAI
A Clearer Way to Find and Review Unusual Points.
Keep the data, detector settings and scoring result connected throughout the review.
A Clear Data Foundation
Prepare or restore a named data draft with the series and applicable contextual inputs.
A Deliberate Detector Choice
Choose a supported method and inspect the settings used to identify unusual points.
Results with Review Context
Inspect flagged observations, scores, threshold and available explanation before deciding what to investigate.
Your scoring context stays connected.
Saved projects hold data drafts and scoring history. Review the data and detector settings behind the result you are investigating.
Process
How We Work
From a prepared series to unusual points you can review.
MAG OptiAI guides your team from preparing the data to choosing a detector and reviewing its scoring result.
02
Data
Prepare and validate the data draft.
Dates and values · applicable contextual inputs · upload or samples
- Series
- Dated observations
- Context
- Where supported
- Readiness
- Review input checks
03
Detector
Choose the detection method and scoring settings.
Supported detector · applicable seasonality and sensitivity settings
- Method
- Supported detector
- Settings
- Applicable to the method
- Policy
- Review scoring context
04
Score
Run scoring for the prepared series.
Generate and save scores, the returned threshold and flagged indices
- Run
- Score the prepared data
- Outcome
- Scores and flags
- History
- Persisted scoring result
05
Results
Review the unusual points and their context.
Series chart · flagged points · scores · exports · grounded explanation
- Review
- Flags and surrounding data
- History
- Open a saved run
- Use
- Investigate and export
Core capabilities
What Anomaly Can Do for Your Organization
A connected workflow for preparing series data, configuring detection and inspecting unusual points.
Prepare Operational Series
Upload or use samples, inspect readiness and restore saved data drafts.
Include Applicable Context
Use single-series or multivariate inputs where supported by the detector.
Choose a Detection Method
Select from Z-score, seasonal/contextual methods, Isolation Forest, LOF and One-Class SVM.
Unusual points.
Review with context.
Review Sensitivity Settings
Configure the settings supported by the detector and inspect the returned threshold.
Keep Saved Scoring History
Open a saved scoring run intentionally as the reference result.
Inspect and Export Outputs
Review series charts, flagged observations and scores, then export results.
Represent Seasonal Patterns
Use the applicable seasonal and contextual settings of the chosen method.
Use Grounded AI Explanation
Ask for interpretation tied to the successful anomaly result and its context.
Prepare Operational Series
Upload or use samples, inspect readiness and restore saved data drafts.
Include Applicable Context
Use single-series or multivariate inputs where supported by the detector.
Choose a Detection Method
Select from Z-score, seasonal/contextual methods, Isolation Forest, LOF and One-Class SVM.
Represent Seasonal Patterns
Use the applicable seasonal and contextual settings of the chosen method.
Review Sensitivity Settings
Configure the settings supported by the detector and inspect the returned threshold.
Keep Saved Scoring History
Open a saved scoring run intentionally as the reference result.
Inspect and Export Outputs
Review series charts, flagged observations and scores, then export results.
Use Grounded AI Explanation
Ask for interpretation tied to the successful anomaly result and its context.
All in one place.
The tools to prepare, score and review in one connected workflow.
Anomaly Detection in action
See the Observations Behind the Flags.
Review the unusual points alongside the surrounding series. This illustration explains the concept; it is not a measured detector result.
A flag deserves context before a conclusion.
Inspect the point's value, score, returned threshold and surrounding observations. Expected values can provide additional context where the detector returns them.
Use flags to focus investigation. An unusual value does not by itself establish a fault or root cause.
Series and points for reviewIllustrative
Observed seriesIllustrative flags
| Observation | Value | Review |
|---|---|---|
| Point 4 | 130 | Flagged for review |
| Point 8 | 38 | Flagged for review |
Who it’s for
Is Anomaly Detection a Good Fit for Your Team?
For teams reviewing unusual observations in a defined operational series.
Operations Review Teams
Prepare a relevant series and investigate changes, spikes or drops against its operational context.
Analysts Choosing Detectors
Select supported methods and evaluate applicable seasonal, contextual and sensitivity settings.
Teams Investigating Flags
Inspect scoring outputs and surrounding observations before deciding what needs attention.
Every series and review question is different. Evaluate whether the selected detector and inputs provide useful signals for your investigation.
The Anomaly Detection Evaluation
Real Problem. Real Evaluation. Real Impact.
Bring your anomaly review question and evaluate the fit with your organization’s information.
What You Get
Access for a Larger Evaluation
Apply through the existing beta process to evaluate beyond the smaller free-trial limits.
A Guided Data Setup
Frame the operational question and prepare a series with the applicable context.
A Detector-Based Workflow
Choose supported methods and inspect the settings used for scoring.
A Direct Feedback Channel
Discuss setup difficulties, unclear flags and requirements that need a better fit.
Results You Can Review
Inspect flagged observations and available interpretation to judge usefulness for investigation.
What We Ask From You
A Defined Review Question
Bring the operational series and the unusual patterns your team wants to investigate.
Usable Historical Data
Provide dates, values and applicable contextual inputs. Anonymized data should retain the patterns and required relationships.
Practical Feedback
Tell us where preparation, detector configuration or interpretation needs improvement.
A Review of the Outcome
Investigate the flagged points in your operational context before drawing a conclusion.
Real Problem
Real Evaluation
Better planning decisions.
A real-world evaluation with your organization.
Bring your anomaly detection problem and real or anonymized historical data. We will review the fit and help you get started.
Anomaly Beta
Frequently Asked Questions
What to know before evaluating your anomaly requirements.
What data do I need?
Start with a dated series and a clear review question. Prepare the values and contextual inputs required by your chosen method, then review the product's requirements and readiness checks.
Which detectors are available?
The current workspace supports Z-score, STL Z-score, contextual Z-score, Isolation Forest, Local Outlier Factor and One-Class SVM. Input and setting requirements depend on the method.
Do I train and stage saved model versions here?
This product's workflow is Data, Detector and Score. Choose the detector and scoring settings, run scoring and review saved results. It does not use Forecasting's model-training and staging workflow.
Can I use seasonal or contextual information?
Yes, where supported by the selected detector and dataset. Applicable settings and input requirements differ between methods.
Does every detector's score mean the same thing?
No. Review the method, returned threshold and flagged points together. Scores should not automatically be interpreted as probabilities or compared as if they share one scale.
Does a flag prove an operational fault?
No. A flag identifies a point for review under the selected method. Check the surrounding observations and operational context before deciding whether there is an issue.
How does Early Access relate to the trial and beta?
Early Access describes product maturity. Start Free uses the existing 14-day trial with lower limits. Larger evaluations use the existing beta application and team approval; there is no separate access-request system.
What can AI explanation help with?
Explanation can interpret the successful scoring result and available context. It does not prove root cause, establish causality or replace an operational investigation.
How do I apply for a larger evaluation?
Use Apply for the Anomaly Detection Beta to reach the existing product-specific application. Our team reviews your requirements and confirms the fit for an evaluation.
Still have a question?
Talk to us about your anomaly evaluation.
Ready to get started?
Ready to Evaluate Anomaly with Your Organization’s Data?
Tell us about your operational series and anomaly review requirements. We will review your application and confirm whether the current Anomaly Detection Beta is a good fit for your organization.
No payment requiredApplication reviewed by our team