LandingAI, a pioneer in agentic vision AI technologies, announced its significantly upgraded version of Agentic Document Extraction (ADE) which will use a new model, Document Pre-trained transformer-2, to accurately extract information even from complex documents, to better inform organizational decision making.
Document Pre-trained transformer (DPT-2) builds on the strength of ADE’s initial launch six months ago, which enabled developers, startups, and Fortune 500 companies to process billions pages through ADE, with many reporting up to a 90% reduction in time spent searching for information.
ADE DPT-2 represents an even more refined foundation for document intelligence. While generic LLMs often struggle to accurately, fully and consistently extract all visual information from documents, DPT-2 combines structured deep learning models with agentic workflows to deliver higher accuracy and trusted outputs from documents filled not only with text, but also data in the form of tables without gridlines, invoices scanned at odd angles, signatures embedded inside forms, check marks, and so on.
“Documents contain the information that organizations need to make not only accurate but the best decisions possible. Key nuances can be lost if the visual representations are not adequately captured,” said Dan Maloney, CEO of LandingAI. “ADE addresses this gap and our enhanced release pushes capabilities even further.”
Also Read: OpenText and Fiserv Launch AI-Powered Content Next to Transform Information Management for Financial Institutions
LandingAI looks for rapid uptake in industries such as finance, healthcare, insurance, or compliance where accuracy is not optional and workflows are primed for greater agentic efficiencies.
“We’re past the era of one-size-fits-all models. Just as we use different processors today for different workloads (your smartwatch has a less powerful processor than your laptop than the datacenter), for AI workloads too, we need different types and amounts of intelligence to do different things.” Andrew Ng, founder of LandingAI. “ADE DPT-2, specialized in processing documents, works uniquely well in the space of document extraction.”
ADE DPT-2 breaks complex parsing tasks into smaller, reliable steps. New features in the release include:
- Agentic Table Captioning. DPT-2 can parse large, complex, no-gridline, and merged-cell tables with unprecedented fidelity. Every cell is preserved, aligned, and made accessible, enabling cell-level grounding so users know exactly where values came from.
- Figure Captioning, Refined. Logos, seals, and small figures are identified precisely and concisely, eliminating the noise of verbose descriptions.
- Smarter Layout Detection. Fewer chunks are missed, even in messy scans. DPT-2 can even detect stamps inside tables and process them separately critical for compliance workflows.
- Expanded Chunk Ontology. Beyond text, tables, and figures, DPT-2 recognizes signatures, checkboxes, ID cards, barcodes, and QR codes ensuring all document elements are classified consistently.
Source: LandingAI