Changelog

Filtering and saved views

Now you can filter all of the data on your Inspect dashboard by granular parameters that will allow you get down to the data that's most important to you and your team.

You can filter by: 

Keyword: this will allow you filter by specific keywords that co-occur or are related to the topics in your Inspect dashboard.

Sentiment: filter down by what your customers hate or love

Data source: target specific channels in the customer journey.

Cohort start date: understand how different topics evolve

Other enhancements

Dramatically improved performance of loading topics and source data feed.

Increased accuracy of highlighting in source feed.

Highlighting in Source Feed

You'll now see highlights of the phrases that are relevant to the topical item you clicked on in the source feed. This should help you find exactly which portion of the conversation is relevant to your analysis a lot faster.

Improved Inspect Layout

We responded to feedback regarding the original feed-based layout on the Inspect feature. We've now condensed that information into a more scannable and scalable table-based layout. The new view in Inspect has the following columns available to be analyzed: 

Topic name: a label for the automatically discovered group of topics and subtopics.

Topic quotient: How often the topic occurs throughout the entire corpus.

Topic evolution

Sentiment evolution

Net sentiment

Enhanced Source Feed

Rather than having to click into a feed of source data and then click in again to see the entire source document— we've condensed the source feed into a convenient drawer that opens up over Inspect table view. This allows you to stay in context what you're analyzing, and quickly switch between topics without multiple page loads.

V1 NLP Pipeline released

We've dramatically improved the quality of our natural language processing pipeline. You'll notice richer and more verbose phrases that are highly characteristic of what your customers actually say. Most topics and subtopics will be at least 2-3 words— if not 4 words long.

Role-based access control and field-level permissioning

Artifact admins can now create custom roles or permission groups inside of Artifact now, and then control which fields from each imported data source are visible for each role. This will give you granular control over sensitive customer data. It also gives our publicly traded customers peace of mind regarding information retrieved from data sources related to revenue operations.