Capturing values from invoices and cheques is the simplest and most straightforward way to use DigiScan for financial document processing.
Setup: Define the fields you want to capture and assign each field the correct data type (e.g., text, number, date, amount).
Setup complexity: 2/10
Recognition accuracy: >98%
Processing recurring document types such as delivery notes, contracts, or service reports is an efficient way to automate structured data capture with DigiScan. Once a template is configured, it can be reused indefinitely for similar document layouts.
Setup: Define and position all required fields on the template, assigning each one the correct data type (e.g., text, number, date).
Setup complexity: 1/10
Recognition accuracy: ~98%
Processing recurring document types that contain line items (these can be any documents that you use internally in your business: material reports, quality reports, inspection reports or any other) – enables full automation of tabular data capture with DigiScan. The system reads both header fields and structured line-item tables, ensuring totals are validated automatically.
Setup: Requires prompt in general instruction. Might require additional tuning and adjusting general instruction and ‘field instructions’. Configure validation rules for totals if needed.
Setup complexity: 3/10
Recognition accuracy: ~95%
Processing complex cheques that include multiple product items, packaging deposits, or discounts requires reconciliation logic to be added in ģeneral instructions’ and ‘field instructions’. Need to be tested and adjusted to comply with the standard of cheques in your country.
Setup: Define the required header fields and line-item columns, assign proper data types (text, number, amount), and enable automatic total verification.
Setup complexity: 5/10 (as cheques differ in how they look, how they are printed and how many line items contain).
Recognition accuracy: >90%
Processing complex invoices that contain detailed line items is a tough task.
Setup: Define all header fields and line-item columns, assign correct data types (text, number, amount, date), and enable total verification between line-item sums, VAT, and grand totals.
Require proper prompt in ‘general instructions’ and ‘field instructions’. Require extensive testing and adjustment before template can provide acceptable results.
Setup complexity: 8/10
Recognition accuracy: ~80% (very much depend on quality of custom prompt, that includes details of you most received invoice types).
Processing recurring document templates that combine printed structure with handwritten input—such as delivery notes, service reports, or approval forms—allows DigiScan to extract both printed and handwritten data consistently across all future documents of the same layout. Once configured, the same template can be reused indefinitely for identical forms.
Setup: Define all required header fields and line-item columns, assign correct data types (text, number, date, amount), and enable handwriting recognition for fields containing manual input. Use fixed template zones to capture handwritten text accurately.
Setup complexity: 5/10
Recognition accuracy: ~90–94% (depending on handwriting clarity and image quality)
Processing recurring whiteboards with handwritten metrics—such as daily goals or shift dashboards—works similarly to typed templates with handwriting. DigiScan reads structured layouts and extracts handwritten values.
Setup: Define fields, line item fields and create validation rules if needed.
Setup complexity: 4/10
Recognition accuracy: ~90–93% (depending on handwriting clarity and image quality)
DigiScan can read numeric values and text from device screens or lab instruments, automating field data collection. Templates can also include calculated or conditional fields that return specific results (e.g., PASS / FAIL) based on captured values.
Setup: Define numeric and result fields for the device layout.
Setup complexity: 4/10
Recognition accuracy: >95%
DigiScan supports fully customizable scenarios where users define any fields, rules, or validation logic they can imagine. You can design templates that interpret visual scenes, describe objects, compare items, or return context-based conclusions—like in this example, where the system identifies and classifies multiple everyday objects from a photo.
Setup: Create custom fields, specify output format or logic (text, numeric, descriptive), and define validation or reasoning rules as needed.
Setup complexity: variable (1–10/10)
Recognition accuracy: depends on task complexity and image clarity
© 2025. DigiScan. All rights reserved
Please enter your email address.
