The Qlik Cloud Discovery Agent – Part 3: Get the Most Out of Your Insights Feed

In part 1 of this series, we set up the Discovery Agent from the administrator's perspective. In part 2 we looked at how developers create insight triggers that deliver meaningful results. In this third and final part, we switch perspective to what matters most on a daily basis: you, as the end user consulting the feed.

The feed is where the Discovery Agent proves its value. Everything configured in parts 1 and 2 comes together here in a scrollable list of insight cards that tell you what changed in your data, where that happened, and why it's relevant. This article covers how to find the feed, how to read it, how to filter it for what matters to you, and how to dig deeper when an insight calls for it.

Where do you find your feed?

The Insights feed is accessible from the Insights or Analytics activity center in Qlik Cloud. Click Feed in the menu on the left. No further navigation is needed.

Your feed is personal. You only see insights from applications you have access to, and only for the measures for which insight triggers have been configured. If the feed appears empty, that's either because no triggers have been created yet in the applications you use, or because no refresh has taken place yet since the triggers were created. Both are worth checking with your application developer if you were expecting content.

Reading an insight card

Each insight appears as a card in the feed. Every card is designed to answer three questions at a glance: what changed, where it happened, and why it's relevant. The first two are facts straight from the data. The third is context generated by the Discovery Agent to help you understand what you're looking at.

Cards show the name of the measure, the time period in which the change occurred, and the nature of the detected change: a spike, a new record, a trend reversal, and so on. You can expand a card for more detail, including a small chart with the historical pattern and the specific data points that triggered the insight. Each card is also linked to the application it came from, so you can navigate straight from the feed to the full dashboard if you want to investigate further.

Note: insights in the feed expire after 90 days. If an insight catches your attention but you can't act on it right away, make a note of it or open the underlying application — you can't save it for later.

Filtering your feed

In a well-configured environment with multiple applications and many active insight triggers, the feed can quickly get busy. The filter options at the top of the feed are your primary tool for managing that.

You can filter by:

• Measures: show only insights for a specific metric, such as revenue or order count
• Time dimensions: focus on insights around a specific date field
• Time period: show only insights for a certain aggregation period, such as daily, weekly, or monthly
• Field values: limit to a specific value within a breakdown dimension, such as a certain region or product category
• Insight type: filter by the kind of detected change, such as spikes, records, or trend changes

These filters can be combined. If you're responsible for the European sales region and only want to see weekly revenue insights, set those two filters and the feed becomes a lot more manageable.

Digging deeper with Qlik Answers

When you want to better understand an insight card, you can open it directly in Qlik Answers without leaving the feed. The context of the insight card, including the application, the measure, and the relevant time period, is automatically carried over to Qlik Answers so you can ask follow-up questions right away.

For example, if you see a spike in customer returns from last week and want to understand whether it's concentrated in a particular product category or region, you can ask that question in plain language. Qlik Answers uses the underlying data from the connected application to respond, with references so you can trace the answer back to the source.

It's worth understanding exactly what Qlik Answers can consult here. When you open Qlik Answers from an insight card, it works with the structured data of the connected Qlik application. It doesn't automatically include any separately configured knowledge bases with unstructured content. If your organization uses Qlik Answers assistants with document-based knowledge for broader questions, those are available through the Qlik Answers interface, not from the feed.

Above model / Below model

These two types are the most advanced. Instead of comparing values to a simple historical average, the Discovery Agent builds a predictive model based on the historical data and evaluates whether the actual values come out significantly higher or lower than that model's prediction.

The model takes patterns in the data into account, including underlying trends and recurring cycles, when generating predictions. This makes above model and below model particularly valuable for measures with seasonal behavior. A spike detection algorithm might flag a peak in December retail revenue as anomalous without realizing that December is always high. A model-based approach factors that seasonal pattern into the prediction and only triggers an insight if December is exceptionally high, even by December standards.

The minimum data requirement is 4 data points for non-daily aggregations and 7 for daily, the same as spike detection. The model can be built on a relatively small amount of data, but more history naturally produces better predictions. If your data has a strong seasonal component and you notice spike detection is producing too much noise, switching to above/below model is the logical next step.

Data volume requirements at a glance

Before you create your first trigger, it's worth understanding the minimum data requirements. Creating a trigger on a measure with insufficient data doesn't produce an error message, but no insights will ever be generated either. That can be tricky to diagnose if you don't know what to look for.

When are insights generated?

The Discovery Agent evaluates insight triggers at most once a day, each time the application data changes as a result of a refresh. If an application is refreshed multiple times a day, only the first refresh of that day triggers the insight evaluation.

There's an important distinction between the first evaluation and all subsequent ones:

• First evaluation: when you create a trigger and the data is updated for the first time, insights are calculated over the past seven data points. For a daily trigger that's seven days; for a monthly trigger that's seven months. Historical data beyond those seven points can be used as analytical context, but insights are only generated for the most recent seven.
• Subsequent evaluations: after that first update, the Discovery Agent only looks at data points that are new since the last evaluation. If you refresh daily, each daily refresh only generates insights for that day's new data.

This has a practical implication: if you create a trigger today on an application that has been running for years, your users won't immediately see a backlog of historical insights. They'll see insights for the past seven data points, and after that new insights will appear as fresh data comes in.

Another timing detail worth knowing: current and future time periods are always excluded from insight calculations. A monthly trigger won't generate insights for the current month until that month has passed. This makes sense, but it's easy to forget when testing a newly created trigger and wondering why nothing is showing up.

Tips and tricks

Start with Spike detection

If you're setting up triggers for the first time, start with Spikes up and Spikes down. They have the lowest data volume requirements and start generating insights quickly. Once you've confirmed the pipeline works end to end, you can gradually add more insight types.

Check your date fields before you start

Before creating triggers, check whether the date fields you want to use as a time dimension are actually loaded as date values in the data model. Open the data model viewer, find the field, and check whether the type is shown as date or timestamp rather than as a text string. If it's a text string, update your load script with date() or timestamp() functions. This is one of the most common reasons a newly created trigger produces no insights, and it can be tricky to diagnose if you don't know what to look for.

Be deliberate in your choice of insight types

Enabling all eight insight types for every trigger feels thorough, but it produces a busy feed. Think about what each measure actually means. For a stable operational measure like average order value, spike detection and above/below model make sense. For a strategic KPI like net promoter score, change in baseline and change in trend are probably what you're looking for. Choose the types that match the question you're actually trying to answer.

Use breakdown dimensions strategically

Breakdown dimensions multiply the number of calculations and the number of insights in the feed. Adding a breakdown dimension with 50 product values means every evaluation now performs 51 calculations: 50 individual values plus the total aggregate. That's powerful, but only if the breakdown at the product level is actually useful to your users. Start without breakdown dimensions, let users work with the aggregated insights first, and add breakdowns once you know which dimensions they actually want to filter on.

What publishing and duplicating does to your triggers

Insight triggers live in an application, but they aren't carried over in every scenario. When you publish an application to a managed space, the triggers need to be recreated in the published version. When you duplicate an application, or export and reimport it, the same applies: triggers don't survive duplication.

Moving an application between spaces is the only scenario where triggers are preserved. Keep this in mind when planning your development and deployment workflow, especially if you maintain a development version of an app that you periodically publish to production.

Section access is a hard blocker

If your application uses section access for row-level security, insight triggers aren't available for that application at all. This is a fundamental limitation of the current version of the Discovery Agent, not a configuration issue you can work around. Keep this in mind if section access is a requirement in your environment.

In the next part

Now that your insight triggers are configured and your refresh schedule is set, the Discovery Agent will start filling your users' feeds after the next data update. In part 3 of this series, we switch perspective to the end user and look at how to read and navigate the feed effectively, how to use filters to cut through the noise, and how to use Qlik Answers to dig deeper into an insight that catches your attention.

To use Qlik Answers from the feed, you need two separate permissions that are set by your administrator. The Insights feed permission (under Features and actions > Agentic AI) gives you access to the feed itself. The Data analysis permission (also under Features and actions > Agentic AI) is additionally required to be able to use Qlik Answers. If the option to explore further in Qlik Answers from an insight card isn't available, ask your administrator whether both permissions are enabled for your account.

This combination of proactive detection and conversational exploration is the core of Qlik's agentic analytics experience. The Discovery Agent tells you what's worth looking at; Qlik Answers helps you figure out what to do with it next.

Tips and tricks

A quiet feed isn't a broken feed

The Discovery Agent doesn't show every change in your data. It evaluates statistical significance and only produces insights for changes that genuinely deviate from the normal pattern. If the feed is quiet on a given day, the system is working exactly as intended: it means your metrics are moving within expected bounds and nothing currently requires your attention.

If the feed stays empty for an extended period in an application you know has active data, that's the moment to check with your developer whether the triggers are configured correctly and whether the data has enough history for the insight types being used.

Manage a busy feed with filters

When the feed is first activated for an application, or after a new trigger is created, you may see more insight cards than you'd normally expect. The first refresh after creating a trigger performs a one-time historical scan over the most recent data points, which is why more cards typically appear in that first run than usual. After that first update, each refresh only produces insights based on newly arrived data.

Don't try to read every card — instead use the filters to focus on the measures and dimensions you actively manage. A filtered feed that's relevant to your role is far more actionable than an unfiltered stream of everything happening across all applications.

The refresh frequency sets the pace of your feed

Insights only appear after a data refresh. If an application is refreshed once a day, your feed for that application is updated once a day. If it's refreshed weekly, you'll get insights at most weekly.

This is useful to know when setting expectations with your team. If you want to monitor a measure daily, the underlying application needs to be refreshed daily too. If your current refresh schedule doesn't match the frequency at which you want to receive insights, discuss that with your application developer or data team.

Use Qlik Answers for the follow-up question

An insight card tells you that something changed. It doesn't automatically tell you why that happened or what to do about it. That's where Qlik Answers comes in. Don't just use it to confirm the numbers — use it to ask the follow-up questions: which customer segment caused the spike? Which product caused the margin to drop? In which region is the trend change occurring?

The more specific your question, the more useful the answer. The context automatically carried over from the insight card gives Qlik Answers a head start, but asking precise, targeted follow-up questions gets you to a usable conclusion faster than broad, open-ended questions.

Don't use the feed as an archive system

Insights expire after 90 days. If an insight is important enough to record, export the relevant data from the application or document the finding in your regular reporting. The feed is meant for timely awareness and action, not for long-term archives.

What the future holds

The Discovery Agent is still a relatively new capability within Qlik Cloud, and there are a few current limitations worth knowing about. The most significant for many organizations is section access: applications that use section access for row-level security currently can't generate insights in the feed. Support for section access is on the roadmap for a future release, so if this applies to your environment, it's wise to keep an eye on the Qlik release notes.

In addition, several further capabilities are planned. A Following tab that lets you follow specific applications or insight categories is coming, as well as delivering insights by email and integration with external tools. If the current feed-based experience is already adding value for your team, the roadmap suggests using it will become even easier in the future.

Wrapping up the series

This third article wraps up our series on the Qlik Cloud Discovery Agent. Across three parts, we've covered the full picture: how an administrator enables the functionality, how a developer builds effective insight triggers, and how an end user gets the most out of the feed that results from it.

The Discovery Agent is a meaningful shift in how Qlik Cloud can be used on a daily basis. Instead of waiting for someone to open the right dashboard at the right moment, it proactively brings the data to the user. Combined with Qlik Answers for deeper investigation, it starts to feel less like a BI tool and more like an always-on analyst keeping an eye on your KPIs for you.