Intelligent Form Reader
Intelligent Form Reader (IFR) is the Salesforce feature that uses OCR and machine learning to extract structured data from scanned forms, PDFs, and images uploaded to Salesforce, mapping the extracted fields to Salesforce records automatically.
Definition
Intelligent Form Reader (IFR) is the Salesforce feature that uses OCR and machine learning to extract structured data from scanned forms, PDFs, and images uploaded to Salesforce, mapping the extracted fields to Salesforce records automatically. Built on Amazon Textract under the covers, IFR is licensed inside the Industries Clouds (Financial Services Cloud, Health Cloud, Public Sector Solutions) where heavy form-based data intake is the norm. A common use case is a customer uploading a paper application or insurance claim; IFR extracts the fields and pre-populates the relevant Salesforce record before a human reviews.
The feature exists because forms (paper applications, signed claims, scanned identity documents) are still the predominant data-intake channel in regulated industries. Manual transcription of these forms is slow, error-prone, and a leading source of operational cost. IFR replaces transcription with automated extraction, leaving humans to review the AI-extracted output rather than type the whole form. Combined with OmniStudio for the front-end workflow, IFR can power a fully-automated paper-to-Salesforce pipeline.
How Intelligent Form Reader extracts and maps form data
Underlying technology: Amazon Textract
IFR uses AWS Textract as the OCR and form-recognition engine. Textract extracts text, key-value pairs, tables, and signatures from PDFs and images. Salesforce wraps the Textract output with industry-specific data models and the OmniStudio integration. The Salesforce-side admin never interacts with AWS directly; the integration is handled by the Salesforce-managed connection.
Form templates and field mapping
Admins define a Form Template for each form type the org handles. The template specifies the fields the form contains (Customer Name, SSN, Date of Birth, Income, Beneficiary) and the Salesforce object and field each maps to. Once defined, every upload of that form type runs through the template and produces a mapped record. Templates support both fixed-position forms (consistent layout) and flexible forms (Textract''s ML handles layout variation).
Confidence scores and human review
Textract returns each extracted field with a confidence score (0-1). IFR exposes the score in the Salesforce UI so a human reviewer knows which fields the AI is confident about and which need verification. Common pattern: auto-populate fields above 0.95 confidence, flag fields below 0.95 for review, present every extracted field with the source-image annotation for context.
Industries-specific use cases
Financial Services Cloud uses IFR for loan applications, insurance claims, and account opening. Health Cloud uses it for patient intake forms, referrals, and prior authorization. Public Sector Solutions uses it for permits, benefit applications, and identity documents. Each industry has form templates pre-built for common form types.
Integration with OmniStudio
IFR is commonly invoked from an OmniScript or Integration Procedure. The customer or agent uploads the form, the OmniScript triggers IFR, the extracted data populates the script''s data layer, and the user reviews or edits before the data is written to Salesforce. The integration is declarative; no Apex required for the common case.
Document storage and retention
The original uploaded form is stored as a ContentDocument in Salesforce (or in Files Connect if external storage is configured). IFR retains the link from the extracted record to the original document, so compliance audits can trace any field back to its source. Storage planning matters; high-volume orgs accumulate gigabytes of form images quickly.
Multi-language and handwriting support
Textract supports English, Spanish, German, French, Italian, and Portuguese for printed text. Handwriting recognition is supported for English with somewhat lower accuracy. Forms with mixed handwritten and printed fields work but may need more human review on the handwritten sections.
Set up Intelligent Form Reader for a form type
Configuration runs through the Industries Cloud admin app and the OmniStudio designer. Plan a few hours per form template.
- Enable Intelligent Form Reader
Setup, Industries Settings, enable IFR. Confirm the underlying AWS Textract connection is provisioned.
- Create a Form Template
Open the IFR admin app, click New Template. Upload a sample of the form. Define the fields and their Salesforce mappings.
- Configure confidence thresholds
Set the auto-accept threshold (e.g., 0.95) and the review threshold (e.g., 0.7). Below 0.7 the form requires manual entry; above 0.95 the field auto-populates.
- Build the OmniScript or Integration Procedure
Create the UX that uploads the form, calls IFR, and presents the extracted data for review. Most admins start with a Salesforce-supplied template.
- Test with sample forms
Upload several variations of the form (printed, handwritten, slightly off-template). Verify the extraction quality. Tune the confidence thresholds based on the results.
- Roll out and monitor
Launch to the user team. Track the auto-acceptance rate, the review rate, and the manual-correction rate. Adjust thresholds and templates based on real-world usage.
- Handwriting recognition is weaker than printed text. Forms with significant handwritten content require more human review and lower confidence thresholds.
- Form templates need to be re-tuned when the underlying form changes. Even small layout changes can reduce extraction accuracy.
- Document storage adds up quickly. High-volume orgs should plan ContentDocument storage growth or external Files Connect storage.
- IFR is Industries Cloud-licensed. Confirm the org has the right feature license before designing extensively on the capability.
Trust & references
Cross-checked against the following references.
- Intelligent Form ReaderSalesforce Help
- Set Up Intelligent Form ReaderSalesforce Help
Straight from the source - Salesforce's reference material on Intelligent Form Reader.
About the Author
Dipojjal Chakrabarti is a B2C Solution Architect with 29 Salesforce certifications and over 13 years in the Salesforce ecosystem. He runs salesforcedictionary.com to help admins, developers, architects, and cert/interview candidates sharpen their fundamentals. More about Dipojjal.
Test your knowledge
Q1. What does Intelligent Form Reader do?
Q2. What problem does it solve?
Q3. What pairs well with Intelligent Form Reader?
Discussion
Loading discussion…