Release Notes - 17 August 2022

We could not be more excited to release a complete redesign of our product, making it easier than ever to take your unstructured data and transform it into business knowledge. Our platform is easier to use, more performant and packed with new features.

Joining us to chat this month is founding team member, frontend and product wizard Dan Palmer!

New data Explorer makes no-code analysis a breeze

The Explorer is how you create dashboards in Relevance, making it easy to view your data, examine topics/trends and analyse relevant metrics. We’ve rebuilt our Explorer from the ground up to create a significantly better experience for users!

Why does it matter?

Over the last 6 months we’ve paid close attention to how our users explore their dashboards. We’ve shipped so many features that usability was becoming a problem. Too many buttons and too many options! We decided it was a good time to re-think our user experience from first principals, and rebuild a significantly more friendly Explorer.

How does it work?

Pages and Views

Start by creating your first page! In the new Explorer you can create any amount of pages - that come in three types:

Category View
Chart View
Document View

You’ll notice that all of our views have received face lifts! They also should load significantly faster too.

Filter Sidebar

Click the filter button in the top left to open our new Filter Sidebar. We’ve moved filters here to give you more real estate for exploring your data. You can pop open and close the sidebar at anytime.

Global and Page Specific Filters

You’ll notice that there are Global Filters and Page specific filters. Global filters apply to every page in your dashboard. Page specific filters apply to only the page you’ve set them up on.

TIP: Want to know which filters you’ve applied to the page?

If you set up page specific filters, they will also appear in a quick use bar on top of your pages (even if your filter sidebar is closed). This is great for helping non-technical users quickly filter or search through your views.

New onboarding wizard helps you preview in real time

We also want to make it super easy for you to onboard new users onto the platform. We’ve created a streamlined onboarding experience for getting started!

Why does it matter?

Getting the most out of Relevance is easy, but often involves running workflows that take time to transform your unstructured data. This new onboarding flow helps you see our AI in action, on a smaller portion of your data (or a sample dataset), in real time.

How does it work?

When you create a new account in Relevance AI, you’ll automatically be sent into our onboarding wizard. Alternatively, you can get back to it any time by creating a new dataset and selecting “Onboard again”:

Onboarding will take you through one of two flows: our no-code text analysis, or advanced Python SDK quick start. You’ll probably want our no-code flow!

Step 1 - Upload your dataset

Upload your data or get started with a sample dataset. We provide a range of sample datasets if you don’t have a CSV on hand.

Step 2 - Select the most relevant text field to analyse

You’ll be asked to select the “most relevant text field” in your dataset. We’ll use this to run some real time analysis and extra the most relevant topics. Try to select a text field that contains meaningful content, such as sentences or short paragraphs.

Step 3 - Describe two of the AI-detected categories in your data

After we’ve finished our real time analysis, you’ll be prompted to help label some of the AI-detected categories in your data. We’ll learn from your labelling style and use that to automatically label the rest of the categories!

Step 4 - Boom! Welcome to Category View

When that’s done, you’ll land in a Category View where you can analyse the AI detected categories and find trends in your data. Voila! ✨

TIP - the onboarding wizard analyses the first 1000 documents in your data

Now remember, the AI will only have run on the first 1000 documents in your dataset when going through this onboarding wizard. To perform the same analysis on the entire dataset, head over to Workflows to vectorise, cluster (AI categories) and more.

New workflows help you transform your unstructured data

Speaking of workflows, we’ve been launching them like they’re going out of fashion! You now wield significantly more power over your unstructured data as a Relevance AI user.

Why does it matter?

Workflows are how you transform your unstructured data into business knowledge, through cutting-edge AI technology. For you - all it takes is a few clicks!

How does it work?

Click View All workflows in the navigation bar to see all workflows we’ve made available! We now have more than 20 no-code workflows, and 30+ Python workflows.

Our no-code AI workflows mean you only have to follow a few simple steps to get started. We’ve made things as simple as possible, so anybody in your team can transform data!

We’re always releasing new workflows, so make sure to come back to the Workflows page often. Check your email as we’ll start keeping users posted as new no-code options become available.

That’s a wrap!

It’s been a huge month for the Relevance AI platform. We’re always innovating and working extremely hard to build a better experience for our users. If you have any questions, or feedback, you can contact me directly. Find me at: [email protected]