The most powerful 1040 automation technology just got better. We’re excited to announce the latest Artificial Intelligence (AI)-powered advancement in the SurePrep tax process: Do It Like Other People (DILOP). Let’s see how DILOP reimagines traditional Optical Character Recognition (OCR) technology and simplifies preparation for your staff.
What is DILOP?
DILOP is a new Machine Learning (ML) feature in SPbinder that automatically indexes non-standard documents based on how similar documents were indexed by your staff.
DILOP delivers the same benefit as Do It Like Last Year (DILLY), a 2022 SurePrep release that auto-indexes non-standard documents based on prior-year binders. But unlike DILLY, DILOP doesn’t require historical data to gather information. Instead, it leverages the collective intelligence of your firm in real time, meaning it can assist your staff within the first year of implementation.
Together, DILLY and DILOP complement each other to help alleviate the workload burden on preparers.
Why SurePrep developed DILOP
1040SCAN and SPbinder work together to automatically bookmark and index source documents into a standardized workpaper binder. But for years, this capability was limited to standard documents while non-standard documents required manual indexing.
Standard documents. Tax documents recognized by OCR technology. These forms contain consistent field layouts year over year, meaning OCR can interpret them through fixed templates.
Non-standard documents. Tax documents that are handwritten, altered, not printed, or complex forms (e.g., state- or city-based entities or municipalities containing format variations). Due to these disparities, OCR technology cannot recognize these forms.
In 2020, SurePrep began developing AI to expand our document recognition beyond the ceiling of standard OCR. We incorporated Natural Language Processing (NLP) to detect documents through text patterns rather than fixed templates. This flexibility has enabled our software to recognize complex items, such as property tax statements, payments and extensions, child tax credits, client notes, and more.
How DILOP works
- Attempt with DILLY. Our software first tries to auto-index non-standard documents using the DILLY method (i.e., matching historical data).
- DILOP Steps In. If no match is found from the prior year, DILOP is deployed to reference similar document matches within your firm.
- Document Handling. Non-standard documents confirmed by DILLY/DILOP are sorted into the appropriate folder within SPbinder’s index tree.
- Fallback Indexing. If a document cannot be matched (e.g., due to a deleted or renamed folder), it will be indexed to the next available folder, maintaining the structure of the index tree.
Benefits of DILOP
First-year efficiency
While DILLY and DILOP provide the same functionality, DILLY requires one year of non-standard manual indexing for each client return. This is done to supply the technology with a frame of reference for year two. DILOP, on the other hand, isn’t dependent on prior-year data and can begin auto-indexing non-standard documents immediately.
Collective sourcing
DILOP doesn’t just observe preparer actions on a return-by-return basis. It uses text patterns to enhance indexing accuracy across your entire firm.
Continued learning
As more documents are processed, DILOP’s AI/ML algorithms enhance their ability to match documents correctly. In turn, this increases the reliability of the auto-indexing feature for your firm.
What else is our Innovation Team working on?
Auto-indexing is just one component in our ongoing mission to simplify tax preparation. In the past few years alone, we’ve unveiled enhanced OCR, AI-powered OCR verification, and ML-based reclassification suggestion for Schedule K-1s.
Our AI-focused webinar demonstrates the transformative potential of AI in the 1040 tax space. Get an exclusive glimpse of SurePrep’s AI offerings and learn more about the implications of AI advancement for the tax and accounting industry.
View Webinar*This link directs to CPAacademy.com’s webinar registration.