Dear judges, please feel free to search through this blog post to find you key words for you categories. Enjoy reading, love PixelPoints <3
Solution
Overall technical sketch and diagram, The solution is a Power Platform–centric architecture that integrates external systems, SharePoint, Azure services, and a Minecraft client to enable secure data exchange, automation, and gameplay interaction through APIs and agents.
At its core, Power Platform acts as the orchestration layer, while Azure handles API exposure and backend services, and the Minecraft client consumes these capabilities through a custom API and agent.
Power Platform
Power Apps is used to build our user interface outside Minecraft. A player will be able to sign in with their Minecraft credentials to fetch relevant information about their player profile and inventory by triggering an API call to our Minecraft server. You have to be logged on in-game in order to fetch the information.
When logging in you call a custom API through a Custom Connector. It will retrieve information about profile and inventory that is used to populate the Power Apps with data such as credits and who the player is.
PixelStreet, our own Minecraft Wall street, for getting knowledge, ask question about player inventory, PointCoins (our own currency), and how to build and craft items using Copilot Studio. Our agent is connected to online resources to find information about the Minecraft and with connection to Microsoft Foundry and Azure Functions in order to get information about live data on our Minecraft server. You can get intel about prices that are fluctuating and updating depending on different world events and players buying and selling.
The Copilot Studio Agent we added a topic that has an action to perform a “Custom search”. It uses a web search and the input is coming from the conversation with the user. So the user is prompted to ask for what they want to craft, then how much and what they have in their inventory before doing a search online to fetch information. For the low code lovers, we have used Power Fx to formulas used in the topic.
The output wasn’t structured enough, so we added a custom prompt to transform the output to a useful format. It was given instructions on how to handle the input from the custom search and what the output should be.The custom prompt is built using instructions and natural language and help from Copilot to build the great instructions.
Pretty cool to build these custom prompts and get some very valuable output when combined with a clever little search to the big internet 😀
Minecraft
We have made some stellar interfaces in Minecraft. A custom Market place that function as a shop and to see your inventory.
When you hover an item, you’ll be presented a graph and dashboard showing trends in prices. You’re able to buy and sell items by right and left clicking them.
Autocrafting: show current options
Autocrafting: creating planks from available options
Demo
Azure
Read the post on client side salsa to understand a little bit more about how we use the different APIs, Azure Functions and custom APIs we have built: https://acdc.blog/wp-admin/post.php?post=10649&action=edit
Automatising deployment of Azure Resources: https://acdc.blog/wp-admin/post.php?post=10645&action=edit
SharePoint and their Microsoft 365 friends
SharePoint is where we store all our documents generated by our flows and external services. When a code of conduct is signed, it is sent to a SharePoint library and then updated with the correct metadata. It also proves easy access to the rest of the solution with the PixelStreet app, PowerBI reports and the Teachers’ Lounge. This is where we gather all the information from Dataverse about the Player, their Teacher and the Contact Person.
The Teacher’s lounge is the Team where the teachers hang out and gets notified by a Teams bot whenever new players arrive. In addition to the Team bot notification, the teacher responsible will get some Planner tasks appointed to them.
Governance
Please read our post on ACDC Craftmanship to get more intel on what we have done for governance: https://acdc.blog/wp-admin/post.php?post=11137&action=edit
Pipelines
Copilot Control System framework for agents
- Use principles to work by
Power Platform ALM – DEV – TEST – PROD
- Avoid using Default Environment
TEST and PROD being Managed environments configured with security groups.
DLP policies
Copilot Studio Overview
Power Platform Pipelines
Privileged Identity Management for activating Global Admin for our users only for a period of time instead of permanently. We needed to have the Global Admin role to enable features like Managed Environments and Power Platform Pipelines.
Purview
Source control
On the Low Code part, as showed before, we are using Power Platform Pipelines and the git control OOTB functionality. It was a great balance between easy implementation, tracking changes, having always a backup available and of course ease the deployment of our solutions through our environments.
On the Pro Code part, of course source control is an essential part of our solution as well! (Not using git for it would be criminal!). We optet for storing all the elements on our solution in a single repository, since we are just two developers who work well together and use branching for working on and deploying our changes.
This is what our repository looks like!
- CSharp: we are storing here our main C# solutions. That is:
- AzureFunctionsApp: different Azure Functions we use for integrations.
- DataverseJobs: one time jobs and scrapping and testing to Dataverse and endpoints.
- EarlyBound: generation of Early Bounds to work more effectively.
- Docs: we store all our docs here, which are sincronised used the Wiki as code functionality in Azure DevOps.
- infra: we store here all our infrastructure (CI) code, and some of the CD part of it.
- biceps: our biceps code is stored here, as we use IAC for Azure.
- scripts: we have a lot of useful scripts for easing the configuration of our VM, setting up our Minecraft Server, updating our plugin…
- plugins/PointPixelAPI: this is our Minecraft Plugin. Here resides all the functionality we developed for it, mainly our Minecraft shop, autoctrafting system and API endpoints.
Continous Implementation (CI)
The first part we consider in this process is how we deploy all our resources in Azure. Since we are continously deploying new things, we use biceps code to ensure every customization made happens without manual intervention and, in case the unimaginable happens, we can always roll back.
That includes AIFoundry, AzFunctionApps, Keyvault, Purview, a VM…
We are also using Managed Identity, so we also set the access between resources through biceps code.
Also, we have some scripts for automatically installing PaperMC and configuring the Virtual Machine so that we are able to connect to it and play. Also, we assign the VM a static IP.
Security and authentication
For security and authentication in Azure we opted to go for RBAC access management + automatically created Managed Identitities.
And of course, we are using Keyvaults for managing our secrets throughtout our whole solution:
Continous Deployment
We have been mainly using scripts for deploying our plugin to our Virtual Machine. That way a process that might see quite tedious (building in Maven -> connecting to VM -> transfering plugin build -> restarting server -> checking server restarts successfully) happens seamlessly!
Small comment: I am the developer and I’m very biased of course, but it’s tiny little part absolute fauvorite of the project! Hehe.
After that, we also set a pipeline to run that command at the speed of a commit push merge to main branch.
Testing process
Since we are two developers developing our Java Plugin at the same time, we agreed on setting some unit testing. That way, we make sure core functionality in the app never gets broken as we agidly add new functionality!
Documentation
Part of the good communication and ease to work together on a team comes from having good documents we all read, contribute to and comment!
Best practices
Wrapping custom APIs in a Custom Connector to retrieve information about a player and it’s inventory with two different endpoints instead of using HTTP requests in Power Automate for instance, making it more secure and safe, but also enable others to use it in the organisation PointPixels so it’s reusable and centralised when it comes development and maintenance.
The Custom connector has two actions, Inventory and Player Information.
Using solutions… Yes, we know it’s given, but we mention it because it’s important. All of our components are packaged in solutions, and these solutions are then used to update TEST and PROD environments with changes and updates
The solutions are connected to Git so that we have source control through Git Connection.
To keep our secrets secure in Power Platform, we’ve added environment variables of the Secret data type, which are linked to our Azure Key Vault.
Code
Autocrafting API
Retrieve all the recipes in for use in AI Foundry
You can also find information about all the existing recipes in the game. Here there is an example:
We created an API with that information that serves as an MCP layer we use for showing available crafting recipes to Minecraft players.
Here are the explanation an example of all the different endpoints:
[GET] /recipe/{id}/{count} returns the recipe for a given item based on the id. It is also provide an amount to recalculate the ingredients needed.
[POST] /showCurrentCraftingOptions display all the available crafting opportunities based on a inventory and tools input:
— If a crafting table is available more items will be unlocked.
[POST] /showRemainingItems/{id}/{count} returns a list of missing ingredients to craft an item based on the inventory and tools available
[POST] /simulateCraftingResult/{id}/{count} simulates the inventory state after crafting is done for a specified item and amount based on a inventory and tools input. math: inventory – recipies expended + newly crafted items.
Microsoft Foundry: Economic Agent to predict future prices based on events
To understand the power our AI Foundry Agent has over our Minecraft world, it is convenient to have a small reminder about what our solution consists on, and also how the Data Model looks like (very briefly!).
Microsoft API -> Canvas App
The Pro Code layer: Minecraft API
We have a really cool Minecraft plugin built in Java that, among other thing, exposes an API that shares real time information about the players.
As an example, here we have an endpoint for retrieving the inventory of a player:
About our solution
We created our own Minecraft server in which we simulate a living economy system. That sounds really cool, right?! That’s how we designed our data model to achieve such a goal!
Our data model
The most important part to understand here is the Minecraft Item Value. We represent the ever changing value of a Minecraft Item by creating a record for every variation in the price, recording both the price and the time range in which it is valid.
Cool! Where comes AI into play?
Right, right! We have the perfect scenario for generating simulations. And of course, we are not going to develop a custom algorithm that calculates it. We have data driven Generative AI !
So we create a beautiful AI Agent in Foundry that we are using from Azure Functions to generate the data and create the records in Dataverse.
With that in place, we decided to make it fun by adding the possibility to define economy crashing events, such as: wars, economic recessions, inflation, economical crisis or even Trump invading Greenland!
So, to make that work, we added this prompt as instructions:
You are a price simulation engine for a Minecraft economy.
Return ONLY a JSON array of objects with exactly these fields:
– id (string)
– price (number with 2 decimals)
– validFrom (ISO 8601 string)
– validUntil (ISO 8601 string)
Rules:
– Generate prices for every item in “items”.
– Use “history” as the trend signal.
– validFrom starts at “from”.
– validUntil = validFrom + stepMinutes.
– Generate exactly “points” time steps per item.
– Prices must be > 0.
– No extra keys. No explanations. No markdown. Output JSON only.
And, having an input as follows, we simulate Trump is taking over Greenland, for example:
event: the economic event that’s taking place.
items: list of items we want to simulate.
history: last 2 prices of the items given.
from: initial timedate to generate data.
stepMinutes: interval between every fluctuation.
points: how many iterations we want to generate for each item.
That is the result!
At the heart of the code section of our delivery is a virtual machine responsible for hosting the Minecraft server
Low code
Low-Code Excellence — Minimal Code, Maximum Impact 🧱✨
In the Arctic Cloud Developer Challenge submissions on ACDC.blog, low-code isn’t a fallback — it’s a super-power. Across multiple team builds, the focus is on using Power Platform’s drag-and-drop, builder-friendly capabilities to deliver real world value quickly, while keeping implementations maintainable, practical, and delightful.
Here’s how low-code mastery shines through:
1. Full functionality with zero PCF components
Several solutions rely entirely on Power Platform building blocks — Canvas Apps, Model-Driven Apps, Power Pages, Power Automate, Dataverse, and Copilot Studio — without writing custom PCF or heavy code. These show that low-code can still be complete and compelling.
2. Responsive interfaces built with Power Pages
A highlight is the Minecraft Builder Interface built as a Power Pages website — fully responsive, intuitive, and template-driven. It lets users interact with game elements, build structures, and trigger events without ever seeing a line of code.
3. Orchestration flows that feel like magic
Power Automate flows weave business logic, condition paths, approvals, and external requests into automated pipelines that literally build structures in Minecraft. This shows how low-code logic can span from data triggers to API actions, replacing manual toil with orchestration brilliance.
4. Copilot and Power Fx unlock embedded intelligence
Teams embed Copilot Studio agents and use Power Fx formulas directly in topics and data-model prompts — essentially weaving AI-assisted logic into low-code constructs, not just UI/screens. This elevates low-code solutions with intelligence without scripting.
5. Reusable templates and connectors
Low-code is taken further with template-ready setups and custom connectors (built with paconn) so that solutions can be deployed in minutes and reused across scenarios. This is practical low-code architecture, not just quick prototypes.
Power Fx in Agents
Data, AI & Analytics — From raw blocks to data diamonds 💎
In the PointTaken26-1 PixelPoint category on ACDC.blog, the team showcases creative and technically rich solutions that weave data engineering, AI-driven insights, and dynamic analytics into one cohesive ecosystem — even inside a Minecraft world.
1. Building a data-centric architecture
The PixelPoint solution is anchored by a Power Platform and Azure architecture that collects, orchestrates, and stores data from external systems, APIs, and gameplay interactions. At its core, Dataverse acts as the central data store for players, inventories, prices, and economy trends.
2. Real-time and trend analytics
Rather than static reports, data flows into real-time dashboards — sometimes literally inside the Minecraft interface itself. For example, hovering over an item in their custom shop presents trend data such as current price, hourly and 24-hour price movement, and calculated trends, all derived from Dataverse records. This is a vivid demonstration of surfacing analytics where it matters most.
3. AI powered data generation and simulation
The team uses Microsoft Foundry AI Agents to simulate economic data for their Minecraft economy. Instead of hand-crafting models, they rely on data-driven generative AI to produce price variations, allowing them to model economic events like recessions or inflation within the game world. This showcases AI not just as insight but as a data generator that feeds back into analytics.
4. Diagnostics and operational insights
For deeper operational visibility, PixelPoint leverages Azure Application Insights to monitor both Azure Functions and AI Foundry models. This not only helps with debugging but also adds another layer of analytics around backend performance and usage patterns.
5. Agents and contextual intelligence
Their Copilot Studio Agent taps into Foundry, external search, and structured AI prompts to bring contextual insights to end users — enabling players to query for crafting information or item trends. By transforming unstructured search output into usable data, they demonstrate AI as a bridge between raw data and decision-ready information.
Digital Transformation
We have developed an interactive learning platform that helps students understand trading, economics, and market dynamics in a practical and engaging way. The solution is built inside Minecraft, where students can buy and sell items in a simulated market that reacts to real-world concepts such as supply and demand, pricing, scarcity, and external events like wars or global changes.
The platform is teacher-controlled, allowing educators to manage student activity. This ensures a safe learning environment while giving teachers full oversight of progress and outcomes.
To create a scalable and intelligent solution, the platform is integrated with Microsoft technologies including Dataverse, APIs, Canvas Apps, Microsoft Teams, and SharePoint. Game data (transactions, prices, inventory, and student actions) is sent through APIs into Dataverse, where it is stored and processed. Canvas Apps are used for dashboards and administration, giving teachers real-time insights and control, while Teams and SharePoint are used for collaboration, communication and information storing.
This solution helps educational institutions do more with less by automating data collection, reporting, and monitoring, reducing manual administration for teachers. It significantly improves the student experience by turning complex economic theory into hands-on learning, while giving educators clear, actionable insights through automation and centralized data.
The platform demonstrates a real-world, intelligent digital transformation by combining automation, data integration, and immersive learning to improve both educational outcomes and operational efficiency.
Power BI
To help users gain insight of how the economy fluctuates, we have set up some simple and easy to understand dashboards in PowerBI, accessible from the PixelPoint SharePoint site. Are the price of gold going up or down? This is where you find out!
Redstone Realm — Business Logic, Built to Adventure ⚙️🟥
Rather than focusing on a single app or interface, the teams design connected ecosystems that work across devices, modalities, and user contexts — desktop and web, keyboard and touch, dashboards and chat interfaces.
1. Business-first solutions across the Microsoft stack
The solutions integrate familiar tools such as SharePoint, Teams, Viva experiences, Dataverse, and Azure data services to solve concrete business problems. This approach grounds the creativity in reality — proving that the same tools used for collaboration, document management, and operations can also power imaginative, interactive experiences.
2. Multi-modal, inclusive experiences
Whether users interact via web apps, Teams chat, dashboards, or conversational agents, the solutions are designed to meet people where they are. Accessibility and usability are treated as first-class concerns, ensuring smooth experiences regardless of device, input method, or working style.
3. AI as an infused capability, not a bolt-on
AI is woven directly into the fabric of the solutions. Using Copilot Studio, agents, LLM-powered reasoning, and intelligent prompts, teams create systems that help users mine insights faster, make better decisions, and automate complex reasoning — the equivalent of forging an AI-infused pickaxe rather than placing individual blocks.
4. Redstone-style orchestration and automation
Behind the scenes, logic flows, agents, and integrations act like redstone circuits and command blocks — triggering actions, reacting to events, and coordinating services across the platform. These intelligent automations turn static data into responsive systems that adapt to user input and context.
5. Trust by design: privacy, governance, and experience
Despite the creativity and speed, the solutions remain rooted in enterprise principles. Data privacy, security, and governance are respected throughout, ensuring that innovation doesn’t come at the cost of trust. The result is solutions that are adventure-worthy and production-ready.
Redstone Realm Takeaway
The Redstone Realm category highlights solutions that balance imagination with impact. By combining Microsoft 365, Dynamics 365, Azure, and AI-powered agents, the teams show how modern business platforms can be assembled like redstone machines — modular, responsive, and endlessly extensible.