AI Assist took three years of development with the aim of bringing quality insights across multiple Adobe clouds.
Adobe has launched a tool called AI Assist that will weave machine learning throughout its analytics platform with the aim of giving customers the ability to uncover currently unknown insights and understand business context.
The effort, which took about three years of development time, is part of an ongoing artificial intelligence theme for Adobe. The company has been working to bring more AI and machine learning into all of its clouds.
John Bates, director of product management for Adobe Analytics noted that enterprises are only accessing 1 percent to 3 percent of the data produced and stored. Adobe’s aim is to take the grunt work out of data science using AI and machine learning to create intelligent alerts and ultimately prescriptive analytics for businesses.
Want to publish your own articles on DistilINFO Publications?
Send us an email, we will get in touch with you.
“AI Assist is introducing machine learning and AI into the interface. It will analyze preferences by user and highlight opportunities,” said Bates. “The problem isn’t the amount of data produced over time. The problem is the amount of data that’s untapped.”
For instance, AI Assist in Adobe Target may suggest an A/B test. It will also provide insights based on learned user habits and preferences. Bates added that AI Assist is designed to automate data mining because it’s impossible to manually comb through the data.
Adobe spent years working with billions of data points on its platform to train users and ensure quality.
Broadly speaking, AI Assist aims to uncover unknown unknowns, understand business context and adaptively learn. The lottery ticket of data science is finding data that hasn’t been tapped before and adding real business value.
AI Assist is part of Adobe’s broader roadmap.
Bates added that users can override AI Assist, but part of the machine learning is scoring data that is interesting to the end user. Adobe has validated its recommendations for multiple users and adapted its insights. In a nutshell, insights from one week are likely to change weeks later as new data and insights are uncovered.
Date: September 26, 2018
Source: ZDnet