Helix Interactives
Build the Infrastructure, and Understand Why
In Helix Labs, learners build real infrastructure in a live Azure environment. Helix Interactives add another layer: they help learners understand the decisions behind what they build.
Learners do more than follow instructions. They compare options, apply rules, make choices, and explain their reasoning. This helps them use what they have learned when they face a new situation.
This page includes three of the eleven Interactives created for Microsoft Azure Administrator (AZ-104). They run directly in the browser, with nothing to install.
Every Interactive Produces Something
A learner can choose the correct answer without fully understanding why it is correct. Helix Interactives go beyond multiple-choice questions by asking learners to work through the idea themselves.
In each activity, learners create something that shows their thinking. They may build a set of rules, choose and defend a position, work through a series of decisions, or complete a comparison table before seeing the answer.
The result shows more than whether the learner was right or wrong. It shows how they reached the answer, and helps them understand the thinking they can apply in a real Azure environment.
The three examples below use different activities because different topics require different ways of thinking.
Assemble the rule set, then prove it against traffic nobody showed you
The lab builds myNSGSecure in live Azure and associates it with a real subnet. This Interactive has the learner design one from nothing first.
Security signs off on four requirements. The learner assembles a rule set from a palette of six and proves it against three tests, all visible before the build begins. Two further tests appear once those three are green. One checks the Azure load balancer health probe, the requirement that gets forgotten, where a denied probe takes a healthy service down without a single rule mentioning it. The other checks RDP between the web servers, which one appealing shortcut rule opens silently while passing everything on screen.
Both shortcuts satisfy every requirement the learner was given. Being told that changes nothing. Assembling a set that meets all four requirements and then discovering what it left open is how a learner reaches least privilege as a conclusion of their own.
Running herem04-nsg.html
Commit to a position, then build the case for it
Geo-redundant storage is a sound choice for an archive workload, and the lab has the learner create a real account with real redundancy behind it.
Three answers are defensible here, and picking one is only the beginning. The learner commits to locally redundant, zone-redundant, or geo-redundant storage, sees what each option costs and what each one actually covers, and then answers a colleague who disagrees. The rebuttal adapts to the option chosen. Defending geo-redundant storage means meeting a distinction the exam tests: the copy sitting in the paired region cannot be read until a failover happens.
A room that splits between two options is the intended outcome. Constructing the argument for a position is what turns a preference into knowledge, and both sides have to name the failure they are paying to protect against.
Running herem07-storage.html
Build the distinction, then walk it until it holds
Lab 10 deploys a virtual machine, protects it with a Recovery Services vault in its own region, and replicates it into a second vault in a second region. Building both for real is what makes a comparison possible at all.
Constructing that comparison takes two passes. First the learner fills a matrix by failure mode, cell by cell, before any answer appears: a deleted file, a corrupted deployment, an encrypted disk, a lost region. Backup goes backwards in time. Site Recovery goes sideways in geography. Then the learner walks one machine from unprotected through to failed over, with a running readout of what could be recovered at each state.
Two of those states recover nothing, and one of them is a policy assigned three weeks ago with no completed job behind it. A learner who walks into that state themselves has built a permanent distinction between intent and protection.
Running herem10-protection.html
Nothing here is scored, because a score cannot build anything
An Interactive is not an assessment and it does not report on anyone. Nothing a learner does leaves the page, which is what makes a wrong answer useful rather than costly.
The feedback on a wrong option describes the belief that produced it and why that belief fails. A learner who reads that repairs the model, where a learner who reads a correct answer only memorises a replacement. Options shuffle on display, so nothing can be remembered by position. Reset loads the next round rather than replaying the last one.
Learners see which part of their own model is load bearing and wrong
Four of the eleven pages close with a readout that contains no answers. It sorts every check the learner completed into four boxes, using the confidence they reported alongside each one: confident and right, confident and wrong, unsure and right, unsure and wrong.
Confident and wrong is the study list, and building a model requires knowing which part of it to rebuild. A learner who is certain and mistaken and a learner who guessed correctly look identical on a percentage, and only one of them is about to make an expensive decision.
One part of what a course arrives with
Interactives are not sold as a product on their own. They are one component of what XtremeLabs delivers around a title, and they are built from the same lab set as everything else in it, so the resource names, the rule priorities, and the cast are the ones the learner meets in Helix Labs.
For AZ-104 that component is eleven Interactives, one per module, about 39 minutes of learner time across the course. An instructor guide arrives with them, carrying the full answer key, the model behind every activity, and facilitation notes for each page, including which case splits a room and where to spend the debrief. A map shows what each module builds and where it sits on the Interactivity Maturity Model.
Each Interactive is one self-contained HTML page. It runs beside the labs in Helix Labs, inside an LMS over LTI, or from a folder on a laptop with no connection at all.
- 11
- Interactives, one per module
- 39 min
- Learner time across the course
- 33
- Activities in total
- 8
- Interactive models
Eleven for AZ-104, and the method fits any course with a lab behind it
Interactives are built per course from that course's own lab set, which is why the AZ-104 set argues about NSG priority numbers rather than about networking in general. The more a lab has a learner build, the more there is to reason about first.
Tell us which course matters most and we will show what its Interactives would cover, and what else would come with them.