As a full-time Power Platform consultant and part-time nerd, I've been closely watching how Microsoft has rapidly integrated artificial intelligence models into its cloud products. I've watched the Power Platform grow since its launch and the changes over the past year feel much more significant than in years past.
Since Microsoft's partnership with OpenAI and the introduction of Copilot, it has become abundantly clear that AI will be an integral part of the platform going forward. To predict possible future Power Platform offerings, we can take a look at the current trajectory of AI and other emerging technologies to imagine how Microsoft may incorporate them into the core platform tools.
Over the next three to five years, I believe we'll see huge advancements in AI (hot take I know) - specifically in the fields of autonomous agents and customized fine-tuned models, which are designed to analyze vast amounts of data to identify patterns and insights. We'll also likely see mainstream adoption of augmented reality devices such as the Apple Vision Pro and the competitors it will bring. These will be big driving forces that will guide some new features that I believe we could see one day in the platform. The following are my predictions for what may be next for the Power Platform.
Power Platform Automated Connectors
OpenAI has already done work on making APIs accessible through natural language with custom GPT actions. I believe we'll see that mixed with browse capabilities and function calling to a point where you don't even have to define actions or create a custom connector to interact with APIs - instead, an OpenAI assistant will define a custom connector based on documentation it finds online, prompt you for authorization credentials, and then you're good to go. An assistant's ability to browse the internet also means it will be able to recommend APIs and integrations based on how your company operates and what services you use. Connection references will be stored the same way they are today, but the APIs that they authenticate will be abstracted into one central interface that allows you to manage what systems you are talking to and when those actions are relevant.
Power Automate AI Approvals
This feature represents a monumental shift toward automating complex decision-making processes, utilizing fine-tuned AI models to analyze data and execute business decisions based on predefined criteria such as Service Level Agreements (SLAs), contractual agreements, and internal policies. This development is not so much a technological change as it is about humans trusting an AI assistant to make these decisions. While it may not seem so on the surface, this would be a leap toward a future where AI's role in business operations becomes central, reshaping how organizations approach decision-making at its core. Making sub-second data-driven objective decisions based on strict guidelines would expedite and ensure consistent outcomes from key business areas such as legal and HR departments, which in turn increases operational efficiency across every department that relies on them.
Power Automate Process Optimizer
As deep learning and fine-tuned AI models become increasingly accessible, I envision systems capable of comprehensively analyzing a company's data across files, emails, and cloud platforms. These systems, designed to understand the nuances of a company's operations, would work alongside autonomous AI assistants to identify inefficiencies, automate routine tasks like data entry, and suggest operational improvements. By integrating with tools like Process Mining, they could pinpoint redundant processes, streamline workflows, and leverage predictive analytics to anticipate and mitigate potential bottlenecks. This not only promises to free up employee time for strategic endeavors but also paves the way for a more centralized and Copilot-ready infrastructure. Ultimately, this proactive approach to leveraging AI would transform internal processes, accelerate informed decision-making, and drive long-term innovation, revolutionizing how companies operate.
AI Builder Studio
AI Builder Studio would revolutionize how we interact with and utilize AI within the Power Platform, by allowing developers to go beyond the pre-built models currently available. This would include a no-code interface to fine-tune models provided by OpenAI, offering a level of customization and efficiency currently unattainable with AI Builder models. Developers could tailor AI models to specific business needs, optimizing processes such as content generation for customer service and marketing or legal and compliance documents. This advancement would not only streamline workflow automation but also open up new possibilities for predictive analytics, natural language processing, and image recognition, making the platform even more powerful and versatile.
Suggested Solutions
Suggested solutions would be Copilot-driven suggestions for new apps, flows, reports, or other resources that solve business problems based on interactions you have through email, chat, or meetings. As Copilot gets more context into how your business operates, it will attempt to solve common problems that deliver the most impact to the most users. For example, if someone mentions in a meeting the need for a helpdesk system, Copilot can begin to build the entire solution as the meeting progresses.
Microsoft 365 Copilot Expansion
Microsoft's investment in expanding Copilot has led to some of the most significant recent changes to the Power Platform and beyond. Copilot will undoubtedly become a tool that we use daily in communication, development, and general tasks. Copilot will likely incorporate OpenAI's assistant framework to improve interaction context and perform operations within your organization. I also see Copilot leveraging a generative AI-focused virtual machine that could replace the current RPA tools such as Power Automate Desktop by simply feeding instructions and having Copilot interface with legacy interfaces like desktop executables or websites that do not provide an API. This shift towards AI-driven automation would mark a considerable advancement in how businesses automate their processes, enabling even more complex and nuanced tasks to be automated with ease. By providing a more intuitive and flexible approach to automation, Copilot would significantly reduce the technical barriers to entry for automation even more so than Power Automate did, democratizing access to powerful digital transformation tools across a wider range of business users and scenarios.
Augmented Reality Power Apps
The Apple Vision Pro opened my eyes to the value of having a device built around the idea of not virtual reality, but augmented reality. These devices will allow you to truly work from anywhere, no desk required. The implications for this in the construction, warehousing, and medical industries are also apparent. Being able to overlay data and objects into the real world will make manual tasks and interactions much easier. Power Apps will add support for these devices that will allow you to leverage object detection as well as object and data overlays for an augmented reality experience that will rapidly increase productivity.
Immersive Teams Meetings
Although this is more Power Platform-adjacent, I'll still include Teams. Augmented reality will also usher in immersive meetings. Being able to conduct immersive meetings is not a futuristic concept, with some companies like Meta, who acquired the Oculus VR product line in 2014, leading the charge into this space. However, the adoption of this as the default way to meet with an increasingly hybrid/remote workforce still feels like a fairly long way out. With the potential to transform remote collaboration by offering participants a more engaging and interactive experience, it feels inevitable that Teams will soon push for their immersive spaces to be the primary way of collaborating. I believe all it will take is more Copilot integration into the immersive experience (as well as the widespread adoption of VR/AR devices) for this to begin to emerge as the primary way we meet.
Power Platform Test Environment
As the adoption of the Power Platform continues to grow, the need for a true test environment will grow along with it. I envision a new environment type that allows you to do automated integration tests with your Power Apps and Power Automate flows through simulated events such as emails, Dataverse records created and updated, or any other trigger. Certain changes or new content can be AI-generated for testing edge cases and tests can be automated when new solutions are imported. AI assistants will have the ability to interact with your Power Apps perform tests based on natural language prompts for test cases and write detailed failure reports along with suggested (or automated) fixes when an error occurs. This will be built into Pipelines so that all unit tests must pass before a solution can go into a production environment.
Beyond the futuristic AI-infused tools that may come along one day, there are still some enhancements to the current generation that I would love to see. Here are some features that could be integrated into the Power Platform with existing technologies as it continues to evolve.
Real-time Power Apps Data
This has been on my wish list since I first started building Power Apps years ago. Given that Power Apps uses standard APIs for data ingress, there would need to be some limitations for it to be available for all data sources. Azure Cosmos DB offers real-time data processing, which suggests that this feature could eventually be introduced to specific data sources, possibly starting with Dataverse, which is currently SQL-based. Real-time data in Power Apps would allow users to see changes on a screen without having to manually refresh the data or use a timer control. Currently, there are options using custom PCF controls and SignalR for this functionality, which gives me hope that it could become an out-of-the-box feature soon.
Power Platform Integration Studio
A Power Automate integration studio would enable the creation of a secure no-code API for your organization. Imagine a flow with an HTTP trigger expanded to include multiple endpoints along with methods (GET, POST, etc.) that your customers will use to interact with your business. Use cases include creating helpdesk tickets, checking project status, account balances, and automated document processing, all through your API. This would give automation-heavy organizations a significant incentive to do business with you. Authentication would be managed through a Power Page interface (similar to Azure API Management) or directly through an Azure B2B connection.
A Power Apps integration studio would allow for external-facing canvas apps to replace Microsoft Forms and incorporate current Power Page features. This would also be an avenue for Microsoft to offer native apps with the wrap feature, providing your company with a standalone app.
More Integrated Azure Products
Many products in the Power Platform are abstractions of Azure products, such as custom connectors (Azure API Management), Hosted Machines (Azure Virtual Machines), and even Power Automate itself (Azure Logic Apps). I'd like to see more Azure resources available directly within the Power Platform, especially Azure Functions. Azure Functions are micro web services that can perform tasks in various languages such as C#, Python, and PowerShell. Introducing a Power Automate Functions product that allows passing inputs and receiving results from pro-code directly within the Power Platform would significantly enhance data processing capabilities. Adding the ability to write functions in PowerFX would make them accessible to existing low-code developers while allowing data to be processed in a separate and reusable method.
Synchronous Power Automate Flows
This feature is likely to be introduced sooner rather than later, given the preview product for low-code plugins that may evolve into this functionality. If you've created a Dataverse plugin, you know that when you register a plugin step, you can choose for it to run before a record is created or modified on the server. This allows for the validation of certain inputs and the interruption of the save function if there's an issue with the data. I have often wished for the ability to create similar functionality in a Power Automate flow, as they are much quicker to build and easier to maintain. Even if this feature were limited to Dataverse operations, it would open up new ways to validate data and decide how the server responds to inputs.
I believe all of the above are within the realm of possibility in the next 3-5 years. AI will continue to change the way we work and build tools for our organizations, and the barrier to entry will only get lower. Understanding these tools, the systems they operate within, and the unique ecosystem of apps and automations that exist in your organization will make you valuable in this new era of instant optimization. I encourage everyone reading this to learn how these systems interact and how you can best build solutions with solid foundations. This way, when Copilot inevitably starts making suggestions, you understand (and can modify) the architectural decisions behind those suggestions. AI is not going to replace us yet, but it will certainly facilitate our growth. Let's embrace these advancements and prepare ourselves to innovate and thrive in the future.
Comments