Instabase and NatWest Unlock Unstructured Data

In an innovative partnership, NatWest and the University of Edinburgh are using Instabase to explore how your health affects your wallet. The Healthy Habits study uses Instabase’s AI platform and advanced statistical modeling to reveal key insights that could help NatWest improve customers’ financial well-being.

PRODUCT UPDATES

• Split Classification: Commercial users can split multi-page documents into appropriate classes by enabling document splitting in project settings.

• Data Types: In extraction prompts, commercial users can select text, list, or table data types , which helps Build return more accurate results. The list data type supports extracting multiple occurrences of an object along with attributes you specify.

• Validation prompt: For commercial users, validation prompts let you tell Build how you want to validate a field in your own words. Your prompt is used to generate a custom validation rule written in Python code.

• Expanded Source information: Reasoning fields now highlight the general region within the document that information was gleaned from to better understand the source.

• Stop managing file paths: The Batches endpoint introduces a simpler way to upload and manage input files. This eliminates the need to manage file paths.

• Improvements to Review: For commercial apps that include downstream integrations , if any document in a run fails validation, the run is now paused for review with the status “Review required.” After completing review, results are sent to downstream integrations.

• Native sheets viewer: Excel and CSV files are now uploaded and viewable in their native format.

• Process more file types: Upload a much broader range of supported file types across AI Hub, including most text and code files.

LATEST NEWS AND INSIGHTS

What’s Really Needed for Document Understanding?

The popularity and accessibility of large language models (LLMs) like ChatGPT make it possible for anyone to copy and paste document contents into the context window of a LLM, but this is only sufficient for the simplest documents with “one-dimensional” text. The vast majority of business use cases deal with “two-dimensional” documents, where layout and location are crucial to understanding its content.

LLMs only have a one-dimensional understanding of text, meaning companies need more than just LLMs to truly understand, analyze, and process documents. Instead, it requires several components that are assembled into one holistic system.

Get a deeper understanding of the challenges with applying LLMs to document understanding and all the critical components Instabase provides to get you the best results in our newest whitepaper “LLMs Are Not All You Need: Full Stack Document Understanding with Instabase AI Hub.”

Read Whitepaper Here ?→

UPCOMING EVENTS

InsTech Webinar: Right Here, Right Now - Finding the Right Data for You

APR 29 ??| ? 8:00 AM PDT/4:00 PM BST | On-Demand

The BIBA Conference

MAY 15 - 16 ??| ??Manchester, UK

Webinar | The Next Generations of Insurance: Building a Tech-Driven Ecosystem

MAY 30 ??| ??1:00PM - 1:45PM EDT | Virtual

Insurtech Insights USA

JUN 5-6 ??| ??New York, USA

Insurance Tech Innovation Conference

JUN 5-6 ??| ??London, UK

WHAT’S HAPPENING IN AI

Mistral’s Latest Model Focuses on Efficient Performance

Mistral just released Mixtral 8x22B , an open-source model designed with cost efficiency and optimal performance in mind.

CIOs Are Investing in AI, But Will They See Results?

Although 42% of CIOs plan to deploy generative AI by Q3 2024, Gartner warns that their timelines and investments are highly aspirational. To ensure positive payoff, CIOs need to figure out how to calculate ROI and look for proofs of concept.

Meta Releases Llama 3 AI Assistant

Meta’s latest iteration offers search integration with popular apps such as Facebook, Instagram, WhatsApp, and Messenger and includes faster image generation than previous versions.

Interested in learning more about AI Hub?