The Qlik Cloud Discovery Agent – Part 1: What is it and how do you set it up?

Most of us know this scenario: you open a Qlik dashboard at the start of the week, scan the charts and try to work out whether anything significant has changed. Maybe you spot something, maybe you don’t. If something unusual happened three days ago, chances are you will only discover it if you happen to look at the right chart at the right moment.

This reactive approach to trend monitoring has always been one of the quiet limitations of traditional BI. Dashboards are powerful, but passive. They show you data when you go looking for it. If something important changes between your visits, that insight simply waits until the next time you open the dashboard.

The Qlik Cloud Discovery Agent is Qlik’s answer to this problem. Instead of waiting for users to discover trends themselves, the Discovery Agent proactively monitors your data after every reload and delivers relevant changes straight to the user, in a scrollable feed, as they happen. In this first article, we explain what that means in practice, how the functionality is structured and what you need to do to get it working in your environment.

What is the Discovery Agent?

The Discovery Agent is an AI capability in Qlik Cloud Analytics that automatically monitors metrics in your applications for meaningful changes over time. When data is reloaded, the Discovery Agent evaluates preconfigured rules, known as insight triggers, and determines whether something noteworthy has happened. If so, the result appears as an insight card in the user’s personal feed : a scrollable overview of everything that has changed and why it matters.

The core idea is a shift from pull to push. In a traditional dashboard environment, users retrieve information by opening an app and looking around. With the Discovery Agent, insights are brought to them automatically. Users no longer have to hunt for trends themselves; relevant changes come to them.

The Discovery Agent consists of three components:

• Insight triggers are the rules that determine what is monitored. A developer creates them in a Qlik application and specifies a measure (for example revenue or the number of open support tickets), a time dimension (the date field over which changes are tracked) and optionally a breakdown dimension (for example by product or by region). The trigger also defines which types of anomalies or trend changes to look for. We go into this in more detail in part 2 of this series.
• Feeds are the personalized streams of insights that users see. Every user with access to an application sees the insights generated for that application in their own feed. The feed is accessible from the Insights or Analytics activity center in Qlik Cloud and is designed to be scanned quickly, much like a news feed. In part 3 of this series, we cover how to get the most out of your feed.
• Qlik Answers integration lets users go beyond the insight card itself. When an insight catches your attention, you can explore it further right away by asking follow-up questions in natural language through Qlik Answers, without leaving the feed or opening the full dashboard. This makes the Discovery Agent part of a broader AI-driven analytics workflow in Qlik Cloud.

Who is it for and what do you need?

Before we go any further, it is worth being clear about who can use the Discovery Agent and what is required.

License requirements

The Discovery Agent is available on Qlik Cloud Analytics Premium and Enterprise licenses, as well as on Qlik Talend Cloud Premium and Enterprise. It is not available on lower tiers and is not supported in Qlik Cloud Government environments. If you are not sure which subscription your organization has, check with your Qlik administrator or your Qlik account manager.

Cross-region inference

This is the most important prerequisite, and the one that most often catches teams off guard. The Discovery Agent uses Qlik’s generative AI capabilities, which means certain data is temporarily processed in Qlik Cloud regions outside your own. This is called cross-region inference, and it must be explicitly enabled by a tenant administrator before the Discovery Agent becomes available.

If you have already enabled cross-region inference for other Qlik AI features, such as AI-assisted data quality rules or GenAI documentation in data products, you are already set up. If not, this is the very first step.

The Qlik Answers connection

Exploring insights through Qlik Answers requires additional configuration: your tenant needs Qlik Answers capacity and the applications you want to explore must be indexed for Qlik Answers. This is separate from the Discovery Agent configuration itself, but it is worth planning for if you want to give your users the full experience.

What does the workflow look like?

The Discovery Agent has three distinct roles, each with its own responsibilities. It is important to understand this distinction before you start configuring.

• The administrator enables the functionality and decides who can use it. This means enabling cross-region inference and granting the right permissions to the right users.
• The application developer creates and manages insight triggers within Qlik applications. This is a configuration task that requires knowledge of the application’s data model, in particular the available measures and time dimensions.
• The end user consults the feed. You don’t configure anything; you simply open your feed, read the insights and act on them. The experience is deliberately kept simple: scroll, filter, explore.

In a small organization, these roles may overlap. The developer may also be the administrator and one of the users at the same time. In larger organizations, they are often separate. Either way, configuration follows this order: administrator first, developer second, users last.

Setting up the Discovery Agent: a step-by-step guide

Step 1 – Enable cross-region inference

Navigate to the Administration activity center in Qlik Cloud. Under the AI settings, you will find the option to enable cross-region inference. Enabling it gives the tenant permission to temporarily send certain data to other Qlik Cloud regions for generative AI processing (see also the Qlik documentation). This setting applies to the entire tenant, not to individual spaces or users.

Once this is enabled, the Discovery Agent is available in principle, but users still cannot use it until they have the right permissions.

Step 2 – Assign the Insight triggers permission

There are two separate permissions that control access to the Discovery Agent, and it is important to understand the difference between them.

The first is the Insight triggers permission. This determines who can create and manage insight triggers in applications. You will find it under Content types > Applications in the default user settings or in a custom role. Assign it to developers and power users who create and maintain the triggers. End users who only consult the feed do not need this role.

Step 3 – Assign the Insights feed permission

The second permission is the Insights feed permission. This determines who can see and use the feed. You will find it under Features and actions > Agentic AI in the default user settings or in a custom role.

An important detail: both permissions are disabled by default in the default user settings. This means that even in a tenant where cross-region inference is enabled, nobody will see anything in their feed until you explicitly activate these permissions. This is a deliberate choice. Qlik wants organizations to make a considered decision about who gets access, rather than switching it on for everyone straight away.

You can enable the permissions broadly through the default user settings, which apply to all users, or more selectively by assigning them to specific user groups via a custom role. The selective approach makes sense if you want to start with a pilot group before rolling out more widely.

Step 4 – Check space permissions

In addition to the permissions above, users also need the right space permissions in the spaces where the relevant applications are located.

• For shared spaces: users need at least the Can view role to see insights in their feed. To create and manage insight triggers, they need Can edit or higher.
• For managed spaces: users need at least Can view to see insights in their feed. Creating and managing insight triggers requires the Can operate role or higher (space owner or manager).

As a tenant administrator, you can view insight triggers and feed insights by default, but you cannot create or manage insight triggers in spaces you are not a member of.

About personal spaces

If you want to create insight triggers in your personal space, for example to try things out before publishing to a shared or managed space, you need the Private Analytics Content Creator role in addition to the standard Discovery Agent permissions. This role must be assigned by a tenant administrator and is easily overlooked, so check it before you start.

You’re all set. What’s next?

Once cross-region inference is enabled and the permissions are configured, the Discovery Agent is ready to use. The next step is creating insight triggers in a Qlik application. That is exactly what we cover in part 2 of this series: which insight types are available, how to configure triggers properly, which data requirements apply and which pitfalls you should be aware of.