AI import

Patsnap’s AI Import tool uses natural‑language models to interpret and map information from your Excel spreadsheets to the appropriate patent fields and custom fields. This automated mapping reduces manual editing and speeds up bulk patent import. The import feature supports uploading lists of patents (based on publication numbers) or full data tables. It matches publication numbers in your file to the Patsnap database, ensuring the correct records are imported

Access the AI Import feature

You can start AI Import from two places within the Workspace module:

1. Workspace home page – Import menu. On the workspace homepage, click the Import button in the upper‑right corner. From the drop‑down menu choose Import patent records. A red arrow in the figure below shows where to find this option:

   

2. AI import button directly within the workspace. In In the toolbar choose AI Import from the Import drop‑down.

   

Select an import method

Once the AI Import window opens, choose how you want to provide your data. Two options are available:

• Excel Import – Patsnap AI. Use this option when you have an Excel file containing patent information. The AI engine will read column headers and intelligently map them to patent or custom fields.

• Publication number import. Use this option when you only have a list of publication numbers. The system will match them to patents in the database

The import method selector displays both options. Choose Excel Import to continue with AI import:

Upload your Excel file

On the Upload file step, drag‑and‑drop your Excel file or click Upload file to browse. A help panel summarises the upload requirements and includes a link to a template:

• Your file must contain a “Publication Number” column. Patsnap matches this field to its database

• You can upload up to 100 000 patent records in one import.

• Files must be in .xls (up to 8 MB) or .xlsx (up to 20 MB) format and contain 300 columns or fewer.

• Downloading and following the template improves mapping accuracy.

The upload page is illustrated here, with the upload panel on the left and helpful tips on the right:

After uploading, the system will validate the file and display the next step.

Preview and adjust field mapping

The Preview patent records step shows a sample of your data and how AI has mapped each column. Recognised patent fields appear in green; custom fields appear in blue. A panel on the right summarises which fields were identified. If the AI maps a column incorrectly (for example, a custom field misclassified as a patent field), you can click the field name and choose the correct type. You can also adjust tag‑mapping options to determine whether new values should be applied to all patents, only untagged patents, or added to existing values. The preview screen looks like this:

During this step you should:

• Confirm patent fields: Ensure that columns such as Title, Application Date and Publication Date are correctly recognised as patent fields. Change any misidentified fields to the appropriate type.

• Confirm custom fields: Check that your custom fields (drop‑down menus or drop‑down tree fields) are recognised properly. Adjust their field type if necessary and decide how new tag values should be applied.

• Review tagging options: If records in the folder already use the same custom fields, choose whether to apply tags to all patents or only untagged ones.

Once satisfied with the mapping, click Import records at the bottom of the page.

Import patent records

When you click Import records, Patsnap will import the patents into your selected workspace. The progress page shows the status of the import and reminds you that large imports can take 5–20 minutes, depending on the number of records. You are free to navigate away from the page; the import continues in the background. An example of the progress page is shown below:

After the import completes, a confirmation page appears indicating how many patents were successfully imported. You can choose Import another file or Take me to Workspace to view the results:

Frequently asked questions and best practices

How can I improve AI Import accuracy?

Prepare your spreadsheet using the recommended fields. At a minimum, include a Publication Number column since the system matches patents by this field. Including other key fields such as Title, Application Date and Application Number improves matching accuracy. Use the official template provided on the upload page to ensure correct formatting. Consistent column names (matching those in your workspace) help Patsnap identify custom fields

What happens if some patents fail to import?

If certain records cannot be matched or imported, Patsnap generates a matching log after the import. The log lists any failed records, explains the error (for example, invalid publication numbers), and provides recommendations. Correct the issues in your file and repeat the import. Patsnap ignores duplicate records that already exist in your workspace

What should I do if a field is recognised incorrectly?

AI mapping occasionally misclassifies columns, especially when data is inconsistent or ambiguous. During the preview step, click on the misidentified field and select the correct type (patent field, drop‑down menu, drop‑down tree or text field). After correcting the mapping, continue with the import. If persistent recognition errors occur, contact Patsnap support with details so they can investigate.