TransformX 2022: Operationalizing AI

Today we kicked off our third annual conference, TransformX, where the world's brightest AI leaders, visionaries, practitioners, and researchers engaged in conversations around the present and future of AI/ML.

At this year’s event, we felt it was important to shine a light on how AI is going to define the next era of technology and power. Just like at the dawn of the internet, or cloud computing before it, the businesses that leverage AI will have an incredible competitive advantage to dominate their industries for years to come. 

In fact, for many businesses, we’re already at a nexus where it’s AI or die. And that’s because not all companies are equipped with the same resources to leverage the power of AI. Some organizations, often the biggest tech companies, are AI builders—they have the talent, resources, and experience to build their own, differentiated AI. But they often lack the high-quality data needed to feed their algorithms. On the other hand, there are organizations that aren’t rooted in technology, but are prepared to leverage AI as a key asset and differentiator. They are looking to leverage this technology into their businesses without having to build their own ML team.

Whether an organization is already building AI or is at the starting line wanting to apply AI, it’s clear that all organizations can reap tangible benefits from AI if they have the right tools and solutions tailored to their business. That’s why today at TransformX, Scale introduced a suite of new products to unlock and operationalize AI for everyone – from startups to researchers and Fortune 500 companies to the U.S. government. These tools have already been battle tested in early access beta by companies such as OpenSea, Voxel and Pickle Robot, and now they are fully available for everyone to adapt, develop and scale their use of AI. 

Studio

Some labeling projects require in-depth subject-matter expertise. For example, medical initiatives have unique terminology and require deep domain knowledge. Studio provides organizations with the capability to run high quality, efficient labeling projects with their own team of experts. Features include the ability to pre-label datasets to reduce labeling time and auto-annotation to make it easy for anyone to label complex objects by simply drawing a box around the specific object. Studio supports all data types, including images, videos, audio, and text. Any team can quickly get started with Studio by signing up here. 

Nucleus Models 

As many ML teams know, shipping a model isn’t the end of the AI journey. To ensure a model performs well in all scenarios and avoids model failure in production, it is critical to test key scenarios beforehand. In 2020, we launched Nucleus, a new way to develop and test ML models. We are now improving Nucleus with new ML model testing functionalities that help catch regression over time and help teams to choose the best model through efficient model comparisons and scenario-level insights. Learn more about Nucleus and its model features here.

Content Understanding AI

One of the biggest challenges for companies that invest in user-generated content is to recommend relevant content that keeps users engaged. With Content Understanding AI, product teams can easily enrich user-generated content to customize and iterate their content’s taxonomy - ultimately increasing engagement. Data enrichment also gives product teams insights into their user’s content, such as sentiment and mood, to help identify growth areas. Additionally, trend detection leverages ML techniques like clustering, topic aggregation and extraction, as well as Scale human-in-the-loop features to detect and understand new content trends surfacing on platforms. Learn more about Scale’s Content Understanding here. 

Document AI Go 

In 2020, we launched Document AI to process complex documents up to near perfect accuracy in seconds. This year, we launched Document AI Go. Document AI Go is a self-serve machine learning platform that enables companies to quickly classify and extract data from their documents in just a matter of seconds. Unlike competing products, Document AI and Document AI Go are powered by foundation models and don’t rely on templates, so they can handle the most complex and unstructured documents. Any organization can quickly sign up and start uploading documents here.

We’re thrilled to introduce these tools and put them in the hands of every company. We can only imagine what’s possible when every company can access the best quality AI tools and tailored models. I believe that by making it easier for more companies to unlock and operationalize AI, we will experience a world where AI meaningfully supports humans as they solve the world’s biggest problems.