Academy
4. Detecting Anomalies with the Anomaly Narrator
Use the Anomaly Narrator to automatically spot unusual spending patterns and understand the AI-generated explanations.
The Anomaly Narrator is an AI-powered feature in OUPI Finance that automatically detects unusual spending patterns in your transaction history. Rather than simply flagging numbers, it generates natural-language explanations so you can quickly understand why a particular expense stands out. For example, it might tell you that your dining spending this month is 3× higher than your recent average, and suggest a possible reason based on your transaction details.
To work effectively, the Anomaly Narrator compares your current spending against at least the last 3 months of transaction data. This rolling baseline lets the AI distinguish a genuine anomaly from normal seasonal variation. If you have fewer than 3 months of history, the feature won't have enough context to produce reliable results, so make sure your transactions are up to date before running an analysis.
To access the Anomaly Narrator, navigate to Finance > Anomalies from the sidebar. The page displays a list of detected anomalies, each accompanied by an AI-generated narrative. Each narrative includes:
• The flagged category or transaction – what the AI considers unusual. • A comparison metric – how the current amount differs from your 3-month baseline. • A plain-language explanation – the AI's interpretation of why this anomaly may have occurred.
You can review these explanations to decide whether the spike is expected (e.g., an annual insurance premium) or something that needs attention (e.g., a subscription you forgot to cancel).
Each Anomaly Narrator analysis costs 1 credit. This is separate from other AI features: the Smart Financial Advisor costs ~2 credits per question, and the Health Score costs 1 credit per calculation. Basic market data browsing remains free.
Keep credit usage in mind if you plan to run anomaly analyses frequently. A good practice is to run the Anomaly Narrator once a week or after importing a batch of new transactions, rather than after every single entry. This balances cost-efficiency with timely detection.
Make sure you have at least 3 months of categorized transactions before running the Anomaly Narrator. The AI auto-categorization feature can help speed this up — when you import transactions via Finance > Transactions, categories are suggested automatically, giving the Narrator a richer baseline to compare against.
Pair the Anomaly Narrator with the Spending DNA feature (Finance > Spending DNA) to get a fuller picture. While the Narrator flags one-off spikes, Spending DNA reveals your long-term habit profile. Together, they help you distinguish a temporary anomaly from a shifting trend.
Open Finance > Anomalies in OUPI now. If you have 3+ months of transaction data, trigger an analysis and read through the AI-generated explanations. For each flagged anomaly, decide: is this expected or does it need action? Note any surprises — you can follow up by asking the Smart Financial Advisor for deeper advice.
The Anomaly Narrator (Finance > Anomalies) automatically spots unusual spending by comparing your current transactions against a 3-month rolling baseline. It produces plain-language explanations — not just numbers — so you understand the why behind each flag. Each analysis costs 1 credit. For best results, keep your transactions categorized and up to date, and combine the Narrator with Spending DNA for a complete view of your financial behavior.