Laurent - stock.adobe.com
It's a twist on conventional content-centric AI tools in that Nuxeo, known for its low-code, open source framework, hopes to enable nondata scientists to create custom data models its AI can use, in the mode of Salesforce's recently released CRM-oriented data modeling tools.
This new generation of AI tools helps stitch back together fragmented enterprise content, which used to reside in one orderly enterprise content management (ECM) application in an on-premises service, according to David Jones, product marketing vice president at Nuxeo and AIIM content association director.
As enterprise content spreads across separate cloud platforms, such as Box, Salesforce and marketing automation repositories, AI content services can help find and tag content in place. That way, it doesn't have to be migrated to a centralized location and updated, and it also makes that data accessible and usable for automated tasks.
"That Nuxeo is opening up the cloud to allow people to access the repositories and interrogate them is a big differentiator," said Jim Lundy, founder and CEO of Aragon Research in Morgan Hill, Calif. "Most cloud ecosystems are closed. While they say they have connectors, a lot of times federated content management is not an easy thing to do."
Platform offers multiple AI connectors
Jim LundyFounder and CEO of Aragon Research
Nuxeo Insight Cloud, released June 4, is part of Nuxeo's newest platform update, called LTS 2019, and it also can interface with other AI tools, such as Amazon Rekognition, Amazon Comprehend and Google Vision, for tasks such as automated image recognition and tagging.
Those tools have inherent limitations in helping organizations organize, tag and make unstructured content accessible, Jones said. It's because the tags can be generic; for example, one of the publicly usable Amazon or Google services might tag a Daffy Duck cartoon image simply as a bird, when the organization needs more specificity to complete a task.
Nuxeo Insight Cloud, Jones said, can help employees mold data models from their company's own data stores and switch on the AI.
Early use cases emerge
So far, early users who have found uses for the tools include media companies organizing and managing rights for their content, as well as insurance companies using, for example, policyholder pictures or dashcam video of auto damage to manage accident claims. Another emerging class of users includes document-intensive business departments, such as HR.
"It allows you to define what data you want to use to train the model and what fields you want to predict," Jones said, adding that the cross-cloud reach of Nuxeo Insight Cloud is part of the company's competitive strategy for taking on bigger, more recognizable vendors, such as IBM. "That's the promise of content services platforms, to me: to deliver where ECM failed," he said.
A company's internal data, broken out into separate models for AI to perform specific repetitive tasks -- such as retrieving customer data or connecting images to marketing materials -- is where AI and machine learning can prove their value to saving human labor costs.
"[Nuxeo] can actually allow enterprises to do specific model work, an advanced AI capability, and most of the other firms aren't doing that yet," Lundy said.