Summary
The AI may give incorrect or incomplete answers if the underlying data is missing, outdated, unclear, or incorrectly categorized.
Why This Happens in Nance
The AI depends entirely on your available financial data and how it is structured. Any gaps or inconsistencies can affect the output.
1. Data Not Yet Synced
- Transactions may not have been fetched yet
- Nance updates data through its automatic refresh system (no manual refresh)
- If a transaction hasn’t synced, the AI cannot include it
2. Incorrect or Missing Categorization
- If a transaction is:
- Misclassified
- Left in Miscellaneous
- The AI may:
- Place it under the wrong category
- Miss it in category-specific queries
3. Incomplete Transaction Data
- Some bank descriptions may be:
- Vague
- Inconsistent
- This affects:
- Category detection
- Contact identification
- Payment channel classification
4. Limited Context in Query
- If the question is:
- Too broad
- Ambiguous
- The AI may return:
- Partial results
- Generalized answers
Example:
- “Show my expenses” vs
- “Show my food expenses this month”
5. Cash Transactions Not Added
- Cash is manual only
- If not added:
- It won’t be included in analysis
What You Can Do
- Ensure transactions are properly categorized
- Add notes for unclear transactions
- Use clear, specific queries
- Wait for the next sync cycle if data seems missing
Important to Know
- The AI has read-only access
- It cannot infer beyond available data
- Accuracy improves as your data becomes more complete and consistent
Good data in, good answers out.