In 2024 we saw a massive swing in Microsoft messaging to bring Copilot and Agentic AI to the forefront of its business productivity offerings. Behind the messaging, we have also witnessed the evolution and merging of many core Microsoft technologies to bring this vision closer to reality. Over the next few blogs, I will be exploring these advancements and providing my view on if, and how, the Microsoft vision can deliver on the current hype cycle promises.
Through 2024 the Microsoft machine made several key announcements through online posts and global events:
- March 2024: Microsoft announces the general availability of Azure OpenAI Service updates, introducing GPT-4 capabilities for enterprise use.
- May 2024: Microsoft Build Conference showcases advancements in Copilot integrations, highlighting real-world use cases.
- July 2024: Microsoft hosts the “Empower AI” webinar series, focusing on AI applications in the Power Platform.
- September 2024: Launch of new AI-driven features in Microsoft Viva at the Microsoft Envision event.
- October 2024: Microsoft’s AI Tour comes to London, celebrating AI innovation in the UK and showcasing practical applications of Agentic AI. Read more about the event here.
- November 2024: The Microsoft Ignite Conference explored the future of Agentic AI and its impact on business automation. Details about Ignite 2024 here.
In 2025 developments continue, with the announcement of $80 billion of likely investment to build out AI-enabled data centres to train AI models and deploy AI and cloud-based applications around the world. They, like many of their competitors, are going all in.
Agentic AI for Business Process Automation
What does this mean for building and delivering AI-backed business automation? Well, let’s break down some of the key Microsoft investments in, and building up to, Agentic AI:
1. Microsoft Azure OpenAI Service
What Microsoft says:
The Azure OpenAI Service is a cornerstone of Microsoft’s AI strategy, providing businesses with seamless access to advanced large language models (LLMs). This service allows developers to create AI-driven applications, enabling real-time decision-making, customer interaction and process automation more rapidly than ever before.
My opinion:
We are in the early stages of reviewing the setup, as well as costs, and will come back to you with educated commentary soon!
2. Copilot Integration
What Microsoft says:
Microsoft has introduced AI-powered Copilots across its suite of tools, including Word, Excel, PowerPoint, Teams and the recently launched Copilot Pages. Copilot Pages enhances content creation by allowing users to rapidly generate, organise and refine web pages or documentation using AI. These Copilots act as intelligent assistants with the goal of simplifying workflows, enhancing productivity and providing actionable insights to users in real-time.
My opinion:
The productivity co-pilots have come a long way in a short period of time; however, they are far from the finished article. I constantly find myself opening Chat-GPT, Claude and you.com, which do a much better job – even with me having to stitch the answers back into the Office suite for use.
3. Azure AI Foundry
What Microsoft says:
Azure AI Foundry is a Microsoft initiative designed to accelerate the adoption of advanced AI technologies across industries from one single location in Azure. By offering tailored AI models, tools and consulting services, the Foundry helps businesses solve complex challenges and innovate faster. This program fits into Microsoft’s broader Agentic AI investments by providing industry-specific solutions and trying to enable more collaboration with partners and customers to drive impactful AI implementations.
My opinion:
Azure AI foundry looks great, it is a little too soon to pass any judgement, with it being so new and with so many capabilities being planned or in preview.
4. AI and Machine Learning Capabilities in Power Platform
What Microsoft says:
The Power Platform—comprising Power BI, Power Apps, Power Automate and Power Virtual Agents—is embedded with AI features that aim to allow businesses to build custom applications and automate processes with minimal coding, while democratising AI.
My opinion:
The Co-Pilots that help you to build automations are getting better, although they still do a better job for prototyping than for production builds and asking co-pilot to adjust an existing flow can a be heart-stopping moment as it wipes out half your flow. For debugging and suggesting issue resolution it is helpful. To consume AI services in the power platform, again this is getting better with more and more connectors becoming available. However, the dream that you have instant access to all your corporate knowledge and that a power app can have AI to make useful decisions and act upon them in a simple way, is not here today. You need to be extremely technical and have a lot of background in the tech stack to build used and safe business process apps.
5. Copilot Studio
What Microsoft says:
Microsoft’s Copilot Studio empowers organisations to create and customise AI-driven Copilots tailored to their unique needs. This platform enables integration with Microsoft 365 and the Power Platform, providing businesses with the flexibility to build AI solutions that aim to streamline workflows, enhance productivity and address specific operational challenges.
My opinion:
Co-pilot studio is a great start, however it is buggy and requires a lot of experience in understanding chatbot (virtual agent) frameworks to go past the templated co-pilots. In addition, the knowledge sources which can be used to good effect are still very limited.
6. Investments in Autonomous Agents
What Microsoft says:
Microsoft is also heavily investing in autonomous agents, a critical component of Agentic AI. Through advancements in reinforcement learning and simulation environments, Microsoft is enabling the development of AI agents that can make complex decisions independently. These agents are designed to optimise workflows, manage dynamic tasks and provide proactive solutions in various industries, from logistics to customer service. These innovations further solidify Microsoft’s leadership in scalable and practical AI applications.
My opinion:
With the merging of the MS stack, Autonomous agents are now possible. The trick here is can you make them ‘Agentic’ in the true sense and how can you run them at scale for business processes. Now, this is EXCITING and a core part of what we are working on at jaam.
Where are we with Agentic AI?
So, do these investments and capabilities mean that we can all build and have our work run by AI agents today? Well, the answer to that question is a firm NO. It is, however, moving in that direction. How far and how quickly we reach it is anyone’s guess.
What can be achieved today can be used to give productivity gains now, and pave the way for taking advantage of incredible advancements in AI capabilities and use cases in the near future. The key to success is going to be staying knowledgeable on what can actually be done at what cost today and designing for what is yet to come. And it’s not going to be a “one platform solves it all” answer.
The Future of Agentic AI
The fusion of Agentic AI, process automation technology and jaam automation’s expertise marks a new era in business process automation. As technology continues to push the boundaries of AI, we remain committed to delivering innovative solutions that empower our customers to thrive in an increasingly automated world.
Think you might have some processes that could be boosted with the use of agentic AI powered business automation? Talk to us.