Operationalizing Large Language Models (LLMs) is the next big opportunity in AI. Any organization…
Six steps to Automated Data Integration from Paper Documents
Even as the world and its data goes digital, paper documents continue to be an important medium for information. From paper receipts for expenses to insurance documents, inspection reports, agreements, and certificates – our reliance on paper continues, and frankly in many cases it is practically easier to be using paper. However, the transfer of data from paper to digital often creates huge efficiency issues for many industries. The COVID-19 crisis has exposed this issue in our healthcare system where fax machines being a key part of the patient report workflow have become a big bottleneck at the intersection of our digital world of health records and the analog world or medical reports. Nexla is on a mission to enable frictionless data exchange across organizations, teams, and data systems. Seeing that the paper-to-digital problem will remain for years, the Nexla team wanted to make it easier for enterprises to include paper documents in their workflow. Scans and pictures taken from mobile phones solve the capture part of the problem.
Read along to learn how the Nexla team leveraged a modular architecture to quickly build an enterprise grade solution for this paper-to-digital bottleneck by automating the flow of data from paper documents directly into APIs, databases, data warehouses, and spreadsheets.
- Email data source automates an otherwise tedious and repetitive task with a user-friendly input.
- Documents automatically flow through a text-recognition OCR process to capture data
- No-code UI allows any user to setup data validation and enrichment as part of the flow
- Validated data can flow into a report, spreadsheet, database, or API while error data can be sequenced into a review workflow
The backstory: A logistics company checks up with Nexla
It started when a modern, digital freight broker saw Nexla’s ability to take email content and attachments as a source of data, detect relevant data from them, and automatically create a data flow. They shared that they get various documents such as insurance certificates by email. Their customers and partners would simply take a picture of a document on their mobile device and send it over.
We quickly outlined the use case. In this particular one, the customer receives ACORD-25 formatted insurance documents. From these documents it is important to verify policy dates, limits, and insurance holders before approving a job. Needless to say it is time-consuming and tedious to do so as a manual process. Not only that, but the process creates a bottleneck that holds up the entire subsequent business workflow. Having an elegant, automated solution could speed up their operations and create an automated workflow with their other digital systems such as CRM, order tracking, loading dock reservations, etc.
The Nexla Approach
Paper documents are certainly not limited to the transportation world. They pose a data challenge everywhere, whether it is receipts to be expensed, HR documents, medical reports or any other situation that requires someone to to check a physical piece of document.
This is where Nexla comes in making document processing is as simple as sending emails with documents into Nexla. Let’s take a look at how this system works.
Step 1: Setting up the email data source
With just a few clicks, Nexla generates a unique and secure email address. The user can simply email any document as an attachment to this email source connector. Documents coming here are automatically detected and classified.
Step 2 : Data flows through text recognition engine
In Nexla, depending on the type of document it gets automatically routed to the right processing engine. For scanned documents and images, it is sent to the OCR system. Nexla is able to connect to multipleOCR systems to digitize the data. Depending on quality of results and confidence scores, Nexla automatically applies the right set of results, giving the user a higher quality synthesized view of data in the form of a dataset.
Step 3. Design your Data
Users can now manipulate this data in Nexla’s easy to use, no-code interface. They can select attributes, enrich them, and apply filters by confidence level.
Step 4. Apply Data Validations
With digital documents it is very valuable to be able to apply validations such as fields that must be available or correct formatting of data such as phone numbers or email addresses. Nexla makes it easy for any business user to apply validations without writing code.
Step 5. Data checked automatically
After the flow is finalized, any new documents get automatically processed. At each step of the process, changes to schema, characteristics, and errors are recorded. Any error records are automatically quarantined, flagging potential discrepancies to users through notifications.
Often records that fail validation are then routed into a manual review queue because no matter how good the OCR it is possible to detect data incorrectly or simply have a document with missing information. Some of our users automate this part of the flow by connecting Nexla to their Sendgrid or Twilio accounts resulting in another email or phone based flow.
Step 6. Results loaded instantly to destination
Once validated, the data flow can be routed to any destination system in any format. The choice is all yours whether you need the data in a spreadsheet in Google Drive, as an entry into your Database, or sent via API call to any of the hundreds of systems we connect with, including your own internal application.
The final word: From our friends in logistics
With our freight brokerage user, automating document processing and visually setting data flows on their own created huge time-savings and value for their business. No more routine manual work and full control over data flows for business users, not just engineers or coders.
It took the Nexla just team just 2 weeks to bring OCR capability to production which is a testament to our modern micro services based plugin design. We are excited to share this new capability that has further extended our ability to detect data in email content and spreadsheet attachments.
Summary of Customer Workflow
TL;DR? Here is a step-by-step to getting your paper documents into an automated digital workflow.
- Setup email as a source of data
- Customer sends email with Certificate of Insurance (COI)
- Email gets automatically picked up by Nexla as a data source
- Nexla automatically calls one or more OCR systems, gets digital data
- Nexla validates confidence scores of fields
- User can set up rules for the data set, such as validations – e.g. for insurance doc it can be, must have policy, must have contact info, name matches, etc.
- As more documents come over time they get automatically processed through step 3-6
- Any documents failing validation or not meeting confidence score are sent into an error queue and processed separately
- Resulting clean data can now be used in multiple ways
- Pushed out to the customer’s internal system via their custom API.
- Nexla integration with Sendgrid is used to send email notifications
- Nexla integration with Salesforce is used to update the CRM with insurance information.
Want to learn more? Request a demo with our team and get a free trial account.
Request a demo to learn how Nexla can work for you and then get started with a 14-day trial
Unify your data operations today!
Discover how Nexla’s powerful data operations can put an end to your data challenges with our free demo.