Introduction
Knowi’s slack integration combines our powerful search-based analytics capabilities with Slack’s intuitive messaging user interface. This enables users to ask questions of their data, receive answers, and visualize those answers within Slack.
There are two steps in the process of setting up Knowi’s slack integration. The first step is installing the Knowi bot, which must be done by a Slack admin. If your admin has not installed the Knowi bot, you will see an error message, in which case you must refer them to this page.
Sections
Manage Queries and Notifications
Once Slack integration is set up, enter /knowi in any channel to bring up the Manage Queries and Notifications menu.
Dashboards and Widgets
From Slack, you can interact directly with dashboards and widgets in your connected Knowi account. Select the dashboard or widget from the dropdown to Get, Ask, or Explore it.
Get
Retrieves a PDF report of the dashboard/widget sent directly to the Slack channel.
Ask
Ask allows you to filter the results of the dashboard/widget report based on a value(s) you enter. For example, if you have a dashboard/widget with data for several employees and you want to filter for just one, simply enter the employee name when prompted:
When the PDF report is returned, it will only show data for "Justina Burdi":
Explore
Explore will automatically redirect you to the dashboard/widget in the Knowi app. Note: It will prompt you to enter login credentials if you are not already logged into the account the dashboard/widget resides.
Dashboard & Widgets Results
Once results are returned from a Get or Ask operation on a dashboard/widget, you have the option to Share, Rerun, Ask, or Explore the results.
Share
Share the dashboard/widget report to another channel or conversation in your Slack workspace with an optional message.
Rerun
To re-run the Get or Ask operation on a dashboard/widget.
Ask
Similar to the Ask operation here, you can also use Ask to further filter the dashboard/widget report returned.
Explore
Similar to the Explore operation here, you can also use Explore to be redirected to the dashboard/widget from the report returned.
Create Your Own Question
To create your own question from scratch:
- Enter /knowi followed by your question:
/knowi <your_question>
- In a few moments, your answer will be returned in the form of a visualization
Select a Dataset
This allows you to select the dataset you want to query by adding “-d” at the end:
- Enter the command:
/knowi <your_question> -d
- The Knowi Search Analytics window will pop-up, with a drop down to select a dataset and a text box with the query string of your question
- Use the dropdown to select the dataset you want to query (by default, it chooses the best match)
- Click Submit
- In a few moments, your answer will be returned in the form of a visualization
You can also specify the dataset name upfront:
- Enter the command:
<your_question> -d <dataset name>
- The Knowi Search Analytics window will pop-up and the dataset name you entered will be pre-selected in the dropdown
- Click Submit
- In a few moments, your answer will be returned in the form of a visualization
Example:
/knowi total email sent -dQuestion Results
Once the results of your question is returned, you have the option to Download, Explore/Validate, save as Favorite, Share, or Schedule the results.
Download Data
Select to Download Data to export your results to CSV.
Explore/Validate
Selecting Explore/Validate will automatically redirect you to the Visualization settings of your results where you can edit the visualization type and settings.
In the Data tab, you can validate and edit the results, ask a new questions in the search-based analytics bar, or download and reset the results.
Share
Allows you to share the question you asked to another channel or conversation in your Slack workspace with an optional message.
Schedule
Allows you to schedule your question/query to run at a chosen interval. Once you click on “Schedule”, you’ll see the following settings:
- Query String - The query syntax of your question (this can edited)
- Alert Frequency - Enter how frequently you want the alert to run
- Alert Frequency Type - Select the alert interval (minutes, hours, days, month)
- Schedule report to a Conversation - Select the Slack channel or conversation the query results will be sent to
- Starting At (optional) - Select the start date for the alert
Favorite Queries
Adding a question/query to Favorite allows you to save and reuse your questions and organize them in categories. Very useful for commonly asked questions.
To set up Favorite Queries:
- Enter your question preceded by /knowi (i.e. /knowi <your_nlp_question>)
- Once the results are returned, click Favorite
- Click Create new category and give it a name
- Click Save
- The next time you enter the command /knowi, the Category you created and the query you saved will appear as a drop down option under Favorite Queries
- Under Favorite Queries, you can run the pre-configured query, modify, or delete it
Ask Questions of Your Documents
To ask questions about the documents uploaded to Document AI, use the /knowi-docschat command followed by your question:
/knowi-docschat what is the total contract value for each contract?
The bot searches only your documents and returns an answer along with the source references. This requires AI Agents Access on your account. For more details, see Agentic BI in Slack.
Example Questions
Below are some commonly used question/query formats to get you started. Note that your actual question will vary based on your dataset and use case.
“All Details” Command
A handy query for getting detailed results for a string value (i.e. customer, product, sales associate, etc.) in your dataset. This will return all records in the dataset that match the value.
Format:
/knowi <"all details for"> <string value> -dExample:
/knowi all details for Wells Fargo -dRolling Date Range
Format:
/knowi <”all details for”> <string value> since <time frame> Example:
/knowi all details for Wells Fargo since yesterday “Where” Statement
Format:
/knowi <”all details for”> <string value 1> where <column name> is <string value 2>Example:
/knowi all details for Wells Fargo where message type is Marketing