Skip to main content

​
AI-Powered Upload

Senzo’s AI upload analyzer removes the need to reformat your data before uploading. It detects your file’s structure, maps columns to Senzo’s schema automatically, and handles common structural issues — all before a single row is written to the database.

​
How it works

When you upload a file, Senzo:
  1. Reads the first 50 rows and all column headers
  2. Sends this sample to Claude (Anthropic’s AI) for analysis
  3. Receives a structured mapping proposal with confidence scores
  4. Displays the mapping for your review and confirmation
  5. Applies any transformations before import
The AI has no access to the rest of your data — only the headers and sample rows sent for analysis.

​
The mapping table

After analysis completes, you’ll see a table with one row per column in your file:
ColumnMeaning
Source ColumnThe column name from your file
Mapped ToThe Senzo field this column maps to, or “Do not import”
ConfidenceHow confident the AI is in this mapping (0–100%)
TransformationAny automatic conversion applied (e.g. date format, unit stripping)

​
Confidence scores

The AI is highly confident this mapping is correct. Common for standard column names like “Headcount”, “FTE”, “Unit”, “Date”.
The AI has a reasonable guess but isn’t certain. Review these mappings before confirming. Common for abbreviations like “HC” (could be Headcount or Health Care) or “OT” (could be Overtime or Occupational Therapist).
The AI could not reliably identify this column’s purpose. Use the dropdown to select the correct target field, or select “Do not import” if the column isn’t needed.

​
Overriding a mapping

Click the dropdown in the Mapped To column to change any mapping. Available options are all target schema fields plus “Do not import”.

​
Required fields warning

If any required fields (date, unit_name, provider_type_name, headcount) are unmapped, they appear as a red warning list:
⚠️ These required fields were not detected — please map them manually: unit_name, headcount
The Confirm Mapping and Continue button is disabled until all required fields are mapped.

​
Structural issue detection

The AI analyzer also detects common file structure problems and resolves them automatically:
IssueWhat Senzo does
Blank rows at file startRemoved automatically
Summary/Total rowsDetected and excluded
Wide format (periods as columns)Transposed to long format
Date format variantsNormalized to YYYY-MM-01
Currency symbols in numeric fieldsStripped ($, £, €)
Percentage signs in rate fieldsConverted to decimal
Each resolved issue appears as an amber banner above the mapping table explaining what was detected and what was done.

​
Saved mappings

After a successful upload, Senzo saves the mapping configuration for your organization. On subsequent uploads from the same source system (detected by matching column header patterns), the saved mapping is pre-loaded automatically. You’ll see a banner:
“We recognized this file format — using your previous mapping. Review and confirm.”
Review the pre-loaded mapping and confirm or adjust as needed.

​
Skipping AI analysis

If your file already matches the standard template exactly, click “Skip AI Analysis — use standard template” to bypass the AI step and proceed directly to Review. This is faster for organizations that export data in a consistent format every month.

​
Privacy and data handling

  • Only the first 50 rows and column headers are sent to the AI for analysis
  • The analysis request is processed server-side — your browser does not send data directly to Anthropic
  • No data is retained by Anthropic beyond the analysis call
  • The full file is processed within Senzo’s infrastructure only

​
See also