Siganova OCR · Available on AppExchange
Siganova OCR reads the ID cards, bank details and invoices attached to your records and fills in the fields, each with a confidence score.
The problem
How it works
In the visual builder, pick the fields to read, their type, the target Salesforce field and a confidence threshold.
When a file is added to a record, extraction starts automatically based on your condition, or on demand.
Each extracted value is written to the record with a confidence score from 0 to 100%.
Every extraction is logged. If one fails, you can rerun it from the console.
Benefits
Start with the built-in ID card and bank details templates, or build your own extractor visually.
Every extraction gets a score your Flows can act on.
Siganova OCR verifies that the uploaded file is the document you expect.
Trigger extraction from a Flow action or the Apex API, and react to the extraction-complete event.
History, retries, settings, a global pause and raw data purge, all in one place.
A built-in guide and tutorial help your admins get started, in English or French.
Security and trust
Your Mistral API key stays stored in your own Salesforce org.
Salesforce permissions apply: users only reach data they already have access to.
No document content is ever written to logs.
From the console, you can purge raw extraction data and pause the service at any time.
Use cases
Context
An onboarding team receives an ID card and bank details for every new customer, attached to the record in Salesforce.
Before
An advisor opens each document, retypes the name, date of birth, ID number and IBAN, then double-checks everything.
After
Fields are filled in as soon as the files are uploaded. A Flow routes only the extractions below your confidence threshold to manual review.
Context
Claims handlers process reimbursement requests that come with an invoice.
Before
The amount, date and supplier are copied by hand, one request after another.
After
The admin has built an “Invoice” extractor in the visual builder. Values land on the record automatically, and handlers focus on the low-score ones first.
FAQ
Siganova OCR is a managed package on AppExchange. Install it in your org, then your admin adds the Mistral API key and turns on the templates you need. You can start in a sandbox.
The interface is available in English and French. For supported document languages, get in touch with us.
Your files stay stored in Salesforce. For extraction, the document is sent to Mistral Document AI using your own API key, stored in your org. Siganova secures this flow between Salesforce and Mistral: no document content appears in logs, and you can purge raw data from the console. Processing by Mistral is governed by Mistral’s own terms and data protection rules, which we encourage you to review; Siganova is not responsible for it. Siganova also does not guarantee the Zero Data Retention policy offered by Mistral: if you need it, request it directly from Mistral.
Extraction runs on your own Mistral API key, so usage is billed by Mistral to your account at their rates. For Siganova OCR license pricing, get in touch with us.
Like any automated reading, an extraction can be incomplete or inaccurate. That’s what the confidence score and your threshold are for. For supported file formats, maximum size and volumes, get in touch with us.
A guide and tutorial are built into the app. For anything else, write to us from the Contact page.
Install Siganova OCR from AppExchange, or book a 30-minute demo to see it work on your own documents.
