At FutureStack16 in November, we shared a preview of our artificial intelligence (AI) technology, code-named Project Seymour. Since then we have worked closely with key customers in a pre-release program. They’ve provided great feedback which has helped us to refine Seymour’s user interface and algorithms. In addition, we have been working behind the scenes to harden Seymour’s capabilities as a core component of our Digital Intelligence Platform. We are already advancing our AI capabilities with Dynamic Baseline Alerts, and now it’s time to bolster that effort by starting a rollout of Seymour to the broader world.

Today, we are very happy to announce that we have begun to roll out Seymour to additional customers. By letting more customers work directly with Seymour, we can continue to iterate on the machine learning and user experience that make this technology unique—and, more important, useful.

Customers are already noticing. “The analysis in Seymour helped me find and fix potential issues I didn’t know existed—before they affected our customers. I’m looking forward to seeing Seymour continue to evolve to provide even deeper and broader insights,” said beta customer Eric Liang, CIO and founder of people search engine Spokeo.

Seymour AI: Powering the Digital Intelligence Platform

Seymour is a personalized view of analysis and recommendations about your digital business provided by AI algorithms that are constantly evaluating your data, looking for the most useful and relevant insights to surface. Seymour is designed to help you find, understand, and fix problems before they affect your business—or your customers.

seymour dashboard

Smart feed of personalized data provided by Seymour [click to enlarge]

What’s inside Seymour?

Seymour’s AI engine, which includes a patent-pending root-cause analysis system, surfaces results in a personalized data feed designed to constantly evolve and improve to meet customer needs. Seymour employs machine learning to understand the information relevant to each user based on how they use New Relic. It also uses a collaborative filter machine-learning algorithm to identify interesting content based on user engagement and how users respond to recommendations.

Personalized recommendations

Seymour offers individual recommendations and analysis about your data and your systems. We call each piece a “card.” There are currently four categories of cards (see below for examples of each type):

Advice. Advice cards take you right to the problem and suggest a solution. Some cards provide predictive analytics; for example, when Seymour projects that you’ll soon run out of disk space. Advice cards can also provide practical solutions and prescriptive analytics. For example, when Seymour identifies anti-patterns in your code that affect performance and suggests solutions.

We’re especially excited about our patent-pending root-cause analysis algorithm. If Seymour detects an anomaly, it will automatically perform a root-cause analysis to identify the most likely source of the problem.

Perspectives. Perspectives cards help you see the big picture of what’s going on in your system, and how it affects your business. We’ll show you what’s changed in your business, whether it’s where traffic is coming from or what applications are underperforming. 

Events. Managing a lot of systems? Working with dispersed teams that sometimes forget to tell each other what they’re doing? Events cards surface events, such as application deployments and new applications reporting to New Relic, to provide you with improved situational awareness.

Celebrations. Engineers have hard jobs. It’s easy to always look at what’s next while forgetting to appreciate how far you’ve already progressed. That’s why Seymour also includes Celebrations cards. Haven’t had an alert in a week? Website traffic up and performance improved? We’ll remind you! 

Which cards you see is based on your feedback, your preferences, and how you use New Relic. Our machine learning combines that information with a collaborative filter that takes recommendations based on affinities between cards. That is, it learns that people who like a certain type of card typically also like another card. The result? You get the information most useful for you.

Seymour keeps getting better

With this release, we’re looking forward to sharing Seymour with more customers. Seymour’s applied intelligence will only get better as more people use it. For more on what we’re doing, read about how we at New Relic think about separating AI reality from hype. If you’re a New Relic customer and interested in checking out Seymour’s limited release, contact your sales rep.

For more information, please check out the Seymour documentation. You can also see a demonstration and get updates on the launch here. And be sure to check out our new Baseline Alerts and join us at FutureStack New York to learn more about Seymour AI and Digital Intelligence at Cloud Scale.

Nadya Duke Boone is the Director of Product Management for Platform at New Relic, responsible for platform-wide features like Alerts, Applied Intelligence, Interconnectivity, and Insights. A licensed electrical engineer, she decamped for the world of software and has built everything from real-time operating systems to full-stack web applications. She is quite fond of airships. View posts by .

Interested in writing for New Relic Blog? Send us a pitch!