LandingAI, a pioneer in agentic vision AI technologies, has released an upgraded version of its Agentic Document Extraction (ADE), powered by the new Document Pre-trained Transformer-2 (DPT-2) model. This enhancement improves accuracy in extracting information from complex documents—such as tables without gridlines, angled scans, and embedded signatures—empowering organizations to make informed decisions in sectors like finance, healthcare, and compliance.
Traditional large language models (LLMs) often falter in capturing visual nuances in documents, leading to incomplete or inconsistent extractions. ADE with DPT-2 addresses this by breaking parsing into reliable, agentic steps, ensuring high-fidelity results from diverse elements like checkboxes, ID cards, and QR codes. Since its debut six months ago, ADE has enabled developers and Fortune 500 firms to handle billions of pages efficiently, reducing search times by up to 90%. "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."
DPT-2 exemplifies the shift from generic models to specialized processors, akin to hardware variations for different tasks. "We're past the era of one-size-fits-all models. Just as we use different processors today for different workloads... for AI workloads too, we need different types and amounts of intelligence to do different things," said Andrew Ng, founder of LandingAI. "ADE DPT-2, specialized in processing documents, works uniquely well in the space of document extraction." This focus on agentic efficiencies suits industries where precision is paramount, streamlining workflows in regulated environments.
The upgrade introduces targeted advancements:
These tools ensure comprehensive, structured outputs from unstructured documents.
DPT-2 is available in preview through the playground and API. The Parse API transforms documents into markdown and semantic chunks, while the Extract API performs LLM-reasoned, schema-driven field extraction with original grounding. For builders, the new Builder Program provides SDKs, cookbooks, priority support, early feature access, higher rate limits, and go-to-market assistance to accelerate enterprise applications. Learn more at landing.ai/partners.
LandingAI's ADE DPT-2 upgrade advances agentic vision AI, delivering trusted, efficient document processing to fuel precise decision-making across industries.
LandingAI is a pioneer in agentic vision AI, enabling developers and enterprises to extract actionable insights from visual data. Founded by Andrew Ng, the company focuses on specialized AI models for complex tasks like document intelligence, driving innovation in sectors requiring high accuracy and scalability.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.