Topics in a nutshell

In the Crystal world, a topic is a subject related to your business metrics that you can talk to Crystal about. For example, you may want to ask Crystal about your company’s Sales over time: that would be a topic. When you ask questions and interact with your advisor to gain business insights, the conversations will always be focused on the topics that the Admins created.

For Admins, in fact, topics are one of the most crucial parts of the crystal experience: this is the reason why they should spend the majority of their time in the Console, setting up topics.

But how does this process work? Let’s see.

The four topic creation steps

  1. Connect

In order to retrieve insights on your business, crystal needs to get data from somewhere! That’s why the first step in the topic creation process is called connect. During this phase the Admins will connect a data source (i.e., database or data warehouse systems) and build topics around that data.

They will also choose the tables to use for the configuration.

2. Configure

During this passage, Admins will choose the topic’s objective based on their needs and on what they want to achieve with that data: for example, a comparison among different data, a ranking of data or an over time progression. Each objective could be visually represented by one or more visualizations, for example a single line chart, a table or a donut chart.

In this phase Admins will have to select the specific columns from the data source’s tables to involve in the topic creation and they will decide whether to group or aggregate data in a certain way. This phase is crucial because it determines how the topic will be.

Note that every combination of objective and visualization has a different configuration with different fields to be completed.

3. Filter

This is the only non-mandatory step of the topic creation process. Admins can decide to add filters (for example, for country) to their topic in order to better detail the questions and to enrich the conversational experience for the members: such filters will, in fact, be available for the queries. It is therefore highly recommended to use filters.

For example, if the topic is Sales over time, a filter might be “in Italy”, if referred to the column “country”.

4. Train

The train step is the fourth and final step of creating a topic. In this stage, admins will help train crystal, so that she will be able to understand the users' requests and their language and to learn the business specific words and jargon.

To do so, they will have to define entities and aliases, which are name variations or trigger words that refer to the metrics used when talking to crystal to make sure she provides the right insights. For example, if a topic is named Sales, the alias could be Incomes: during the conversation, the advisor will understand that both words refer to the same topic.

During the train step, the system will suggest some sample questions to ask crystal based on the chosen configuration.

Completing the topic

After the topic creation process is completed, there are still two passages to go through before the topic is ready for the conversation: the first one is to assign permissions, or, in other words, to decide which member users will be able to see this topic in the advisor and in the Dashboard and ask questions about it.

The second one is publishing: a topic can be kept in draft or ready status, but it’s only when it is published that it will be available for members and their questions during the conversation.

We hope this article helped you understand what lies in the backstage of crystal.

Keep exploring the Tutorials section to find out more on crystal!

If you have any questions about crystal’s features, you have encountered a problem or you would like to share your feedback, contact us at support@igenius.ai.

Last updated