Skip to main content

Decision Tree

Updated over a week ago

Use decision trees in Elium to help your users solve complex problems more efficiently through a clear, step-by-step guided process. Unlike traditional content formats, decision trees are specifically designed for diagnostics and structured navigation.

1. Principe

A decision tree works the same way as an article in Elium. It must be published in a space, can be saved as a draft, and follows the same access rights rules.

All standard actions are available: move, delete, archive, etc. A decision tree can also be part of an approval workflow or have an expiration date set.

In short, a decision tree is an article structured in an interactive, step-by-step format.

2. Use Cases

📘 Guides
Guides become interactive with decision trees: instead of reading through a long document, users answer a few questions or select options to be directed straight to the information that applies to them. This helps tailor the guidance to their specific context.

🎓 Tutorials
A tutorial can be broken down into conditional steps: if the user has already completed certain actions or has specific knowledge, they can skip ahead. Otherwise, they can be guided step by step. The tree adapts to their level and choices.

🩺 Diagnostics
This is the most obvious use case: users are asked a series of questions based on their previous answers (e.g., “Does the product turn on?”, “Do you see a red light?”) to reach a conclusion or action. This avoids having to dig through lengthy technical documents.

🛠️ Internal or External Support
In support scenarios, decision trees reduce resolution time by guiding the user or agent to the right answer. Internally, they help formalize procedures (HR, IT, etc.); externally, they automate part of the customer response while keeping the journey clear and understandable.

3. Define the Scope

Before creating your decision tree, you need to clearly define its scope.

🎯 Target a specific need or frequently asked question
Start with a recurring issue or situation users need help with, like “I can't log in” or “Which tool should I use for this task?” A good decision tree addresses a concrete need—not a broad theme.

🔍 Check whether the topic requires conditional reasoning
Ask yourself whether the answer depends on multiple factors (context, symptoms, user profile, etc.).


If the answer is the same for everyone, a regular article will do. Decision trees are best when the answer depends on user input or choices.

🎯 Keep the scope to a single end goal
One tree = one clear objective. Examples:

  • Find the right password recovery procedure

  • Identify a technical issue

  • Choose a suitable training option
    If you have multiple end goals, consider creating multiple trees.

🧩 Use the tree as an orchestrator of atomic content
A decision tree doesn’t need to include everything. It can link out to existing articles, each focused on a specific action, concept, or procedure.

This approach allows you to:

  • Reuse existing content without duplication

  • Improve searchability: linked articles remain searchable on their own

  • Keep the tree simple and easy to follow

Example: at the end of a decision path, the tree could point to:

  • “Follow the password reset procedure”

  • “Read the full refund policy”

4. Internal Structure

A decision tree is made up of two main components: steps and branches.

  • Steps represent the different nodes of the journey. They can include the same content blocks as a regular article. This is the content users will read as they move through the tree.

  • Branches connect the steps together. They define the possible paths and allow users to navigate from one node to another. When viewing the tree, branches appear as links with titles between steps.

5. Creation

Decision trees are created from the standard content creation menu in the top right corner.

You’ll then enter a dedicated editor with a tree that starts with a single step.

  1. The title of your decision tree – this will appear as the title of an article.

  2. Visualization of your decision tree – it always starts with a root step.

  3. The space where your decision tree will be saved

  4. Draft status

  5. Options – full width, protected, and thumbnail.

  6. Go back in the editing process

  7. Preview as an end user

  8. Zoom option

6. The Editor

Each card represents a step in the editor.

Root Step

  1. Step title

  2. Indicates that this is the starting step of your decision tree

  3. Number of directly connected steps

  4. Quick access to the step’s content

  5. Options to delete or duplicate the step

  6. Quick preview of the step

Add Steps

When you move your mouse to the bottom of a card, two new elements will appear: a black circle and a "+" button.

Click the "+" button to add a new step.

You’ll then see two steps and a line connecting them—this is called a branch.

The rounded rectangle lets you give this branch a title, which will appear as a button for the end user.

The second step follows. Steps without a title will appear greyed out.

Link Steps

When you click the (+) button, you automatically create a new step with a parent-child link.

However, it's also possible to connect two existing steps.

By clicking on the black circle, you can link existing steps together. You can link steps at any level of the decision tree.

If you “don’t go downward”—meaning you link a step to another one that is higher up or on the same level in the tree—a shortcut step will be created (see example below).

Shortcut steps are ghost representations of the target step. They help maintain a clear visual tree structure.

Add Content to a Step

You can think of a step as a mini article. A step can include any element found in a standard article.

Click the "Open Editor" button on any step to start editing its content.

The (+) button in the step editor gives you access to familiar elements such as lists, links to articles, files, and more.

Preview

You can preview a step at any time by hovering your mouse over the eye icon on the step card.

You can also get an interactive preview of the entire tree by clicking the "Preview" button in the top right corner of the editor.

This allows you to step into the user's shoes and better understand how they will interact with the decision tree.

7. Viewing Mode

Once the decision tree is published in a space, it becomes accessible to all users who have access to that space—just like articles.

Here’s an example of a decision tree titled “Refund Request,” made up of three steps:

And here is how it appears to users when viewed on the platform:

The two branches “Yes” and “No” appear as buttons, leading to the steps “Expenses over €1,000?” and “Create a new supplier,” respectively.

Just like articles, users can comment on, share, or favorite decision trees.

An icon in the top right corner also provides a global overview of the decision tree, allowing users to click on a specific step to jump directly to it—handy when you already know where you want to start navigating!

8. Good to Know

Decision trees are searchable just like standard articles.

Did this answer your question?