Radiant is a product analytics platform for teams building with Generative AI.

Radiant allows product teams to: 

• Quickly see and characterize what users are sending to models and what the models give back using evaluators, filters and clustering. 

• Be able to see the interactions how the user seems them, with detailed visibility into complex interactions

• Understand how well AI is actually reaching business objectives by integrating with external business metrics

Product Updates

Improved Field Management In Data Explorer

• We’ve now consolidated Field creation and management into a single panel. You can now add fields from Requests, Interactions, and Sessions together in the data explorer for filtering and exploration. 

• It’s now much easier to manage the many fields – including static functions, LLM judges, and metadata fields –  that you create for data exploration and evaluation.

Consolidated fields selection panel in Data Explorer and Dataset Manager

As the number of analyses in Radiant increase, we’ve seen users create more and more derived fields to do exploration and filtering. As more fields get added, the dataset explorer view gets wider and wider, making it hard to see the fields that you care about at that moment.  We’ve fixed this so that you can easily show/hide relevant fields in your current screen, both in the Data Explorer and Dataset Manager.

Adjudicating between multiple (human and machine) evaluators

We’ve seen cases where multiple human and machine evaluators take many passes over a given request / interaction / session, and sometimes disagree. This is common in cases such as:

• Moving from a v1 of an evaluator to a v2

• Having a human reviewer as a final pass on top of machine evaluators

• Having multiple human reviewers disagree

When this happens many users want a final human adjudication. We’ve made this easy to do in-product. If a user is an owner of an evaluator, they can select the preferred response of the many available, and provide a rationale.

Easily copy and paste Radiant data into Excel or Google Sheets

At Radiant, we often use spreadsheets to prototype new functionality in Data Explorer and our datasets product. We’ve also seen our customers do the same – nothing beats a spreadsheet for ad-hoc analysis and sharing data externally. We’ve now made it really easy to copy data from a Dataset Manager or Data Explorer into a spreadsheet. This works really well if you want to

• Copy the results of an investigation into Excel

• Share a dataset with an external party using Google Sheets

We’re curious to learn where you do use a spreadsheet alongside Radiant, and if there are ways we can build more functionality into the product. Please reach out!

Bug Fixes and Quality of Life Improvements

Previewing filled-in judge templates before running them

Last week, we shipped a way to easily test LLM-based judges in the UI on a few rows. We realized it was also helpful to see what a populated judge template would look like before executing it. So, we built it. We’ve already caught some typos in judges we’ve been writing internally with this fix

About Radiant

Radiant is the Enterprise AI platform to take you from idea to production deployment. From security and governance to scaling and anomaly detection, we make it simple to make AI a critical part of your business. 

Try out a demo here, sign up here to get your own instance, or reach out to our founders directly at [email protected].

We’re also hiring. If you know of someone great that is interested in helping every company build AI into their products and operations, we'd love an introduction.