Last month, Bloomberg Businessweek published an article entitled Microsoft’s CEO on How AI Will Remake Every Company, Including His. The piece offered a compelling inside look at Microsoft’s AI strategy through candid interviews with CEO Satya Nadella and several members of his executive leadership team. Together, they outlined a bold vision for the future of enterprise software—both within Microsoft and across the industry at large.
At the center of that vision is agentic AI: autonomous software agents that can reason, act, and collaborate with humans to get work done. Nadella didn’t mince words. In a move reminiscent of the historical “burn the ships” moment, he made it clear to his leadership team that Microsoft is going all in with agentic AI.
In this article, we’ll explore what that means not just for Microsoft, but for the entire enterprise software ecosystem. We’ll analyze how these software powerhouses are rethinking the role of AI, what this shift signals about the future of business applications, and what steps you should be considering right now as AI agents move from concept to cornerstone.
Let's dig in.
Moving Away From Forms Over Data
One of the most striking quotes in the article came from Charles Lamanna, Corporate Vice President of Business Apps and Copilot at Microsoft—echoing a directive from Satya Nadella himself: “The last five years (the solutions) we spent building, it doesn’t matter. It’s not worth anything anymore. Burn the ships.” It was a dramatic signal that the company isn’t looking to gradually evolve its products or just innovate around the edges. It won't happen overnight, but you're seeing a clear commitment to reimagine their entire business applications catalog from the ground up with AI.
To understand what this all means, let's look at the evolution we've seen with AI-infused apps in the past couple of years. The first wave of AI integration introduced copilots that were clipped to the corner of application screens like Clippy’s more capable cousin (see the example below in Figure 1). While these types of solutions have undoubtedly leveled up app experiences, it's been more about innovating around the edges than a full-scale transformation.
Figure 1: Copilot Pinned to the Side of Dynamics 365 Business Central
The next evolution of AI-infused apps is already taking shape, with app workspaces being reimagined from the ground up to deliver smarter, more contextual experiences. These new designs bake AI directly into just about every aspect of the app flow, offering automated insights, proactive decision support, and task suggestions that adapt in real time.
A good example to illustrate this is Microsoft's new Dynamics 365 Contact Center app. As you can see in Figure 2 below, Microsoft has reimagined the entire user experience with workspaces that dynamically adapt to business context, proactively surface insights, and take action on behalf of the user:
The omnichannel conversation interface is translating audio to text (and also potentially between languages).
Sentiment analysis is being used to help customer service agents monitor customer sentiment.
Generative AI technology is used to summarize the conversation (and in context of previous requests).
As the conversation progresses, the app surfaces relevant knowledge base articles to help agents respond to issues in the flow of work.
Figure 2: Application Workspace Completely Reimagined with AI
While the user experiences in apps like Dynamics 365 Contact Center are noticeably more intelligent and assistive, the underlying structure of these apps still leans on familiar form-based layouts. However, as agentic web technology continues to mature, another evolution is starting to unfold in which AI-powered agents are beginning to replace applications rather than simply enhance them.
Adopting an Agent-First Mindset
As the saying goes, sometimes the best app is no app at all. This is especially true for an increasingly mobile workforce that doesn't have time to toggle back-and-forth between apps all day long. These users crave solutions that deliver work into the hands of the right people at the right time within the normal flow of work.
When you think about it, this is what agents do best. To put this into perspective, consider the process flow illustrated in Figure 3 below.
In this scenario, a work order has been put on hold by a technician in an SAP ERP system. Before, operation supervisors were never directly alerted when these kinds of status changes occurred. As a result, part of their unwritten job description was to periodically log into SAP, run a list report, and scan for any work orders that had been placed on hold. From there, they had to drill into each individual order to investigate the cause—sometimes even reaching out to technicians manually to gather more context. If the root cause was something like a missing replacement part, the supervisor would then need to initiate a completely separate process to check inventory, locate the part, and potentially place an order to get things moving again.
Figure 3: Reimagining Process/App Flow with Agents
In this reimagined, agentic flow, however, that responsibility shifts. An autonomous agent now monitors work order statuses in real time and is capable of reasoning through many of the necessary remediation steps on its own. If a part is missing, the agent can search for available inventory, check lead times, and even prepare a purchase request if needed. And when human input is required—say, to approve an expedited order or choose between sourcing options—the agent surfaces a prompt directly within Microsoft Teams, reducing the cognitive load and context switching. Once the supervisor provides their input, the agent picks up the baton again and continues executing in the background, allowing supervisors to stay focused on managing operations instead of chasing down data across disconnected systems.
Reimagined User Experiences
As Microsoft has gone all-in with the agentic web, Satya Nadella is fond of saying that Copilot is the UI for AI. Marketing hype aside, in a recent article, we talked about how the UX for agents has evolved rapidly to support multimodal (or universal) interfaces where users can not only chat with agents, but also via voice commands, uploaded images/files, and so forth.
Figure 4: Working with Multimodal Agent Interfaces
As autonomous agents mature, we're also seeing significant advances in terms an of agent's ability to keep track of process context and "remember" where a request was left off. Moreover, as the reasoning capabilities of modern GenAI models mature, autonomous agents in particular are able to work more independently.
Bottom line: agents are getting a lot smarter. And, for the first time, developers now have access to the tools they need to move beyond the prototype phase and build enterprise-grade agentic web solutions that can take the place of traditional forms-over-data style applications.
Of course, this paradigm shift didn't just come out of nowhere. Prior to the explosion of agentic AI, there's been a broader, process-oriented design evolution that’s been unfolding for over a decade. UX design patterns have long advocated for moving away from monolithic applications—those sprawling interfaces that try to cram every data point and function related to a business object (like a purchase order or vendor) into a single, complex UI. Instead, we've seen a clear move toward role-based, task-focused applications that are simplified and streamlined to support how people actually work.
Figure 5: SAP's Fiori-Based UX Evolution
By centering apps around specific roles and workflows, developers have been able to deliver solutions that are more intuitive, more maintainable, and more responsive to business needs. The agentic web is, in many ways, a natural progression of this trend. Only now, we’re using AI to take things a step further. Instead of simply streamlining interfaces, we’re empowering autonomous agents to work in the background, handling repetitive tasks, guiding users through complex decisions, and only surfacing what’s necessary at the moment action is required.
Figure 6: Bringing Forms to Users Instead of the Other Way Around
Agentic AI & Development Economics
A recent IDC study made headlines with its bold prediction that over 1 billion new logical applications will be built by 2028. While the headline figures caught attention, many pundits and skeptics seemed to overlook a key detail: the emphasis on logical applications.
While there are many software vendors exploring full-scale AI-based product-suite renovations, the shift towards agentic AI is leading to an explosion of smaller, task-specific agentic or low-code solutions. These role- and task-focused tools, when designed and woven together thoughtfully, have the potential to completely reimagine how product suites and enterprise systems are constructed and consumed.
Figure 7: Composing Agentic AI Solutions Using Low-Code Tools Like Microsoft Copilot Studio
What makes this shift so impactful is the new composition model it enables. Instead of building sprawling monolithic apps, software vendors, implementation partners, and even customers can now rapidly compose lightweight, AI-powered solutions that sit on top of core ERP, CRM, and other foundational systems. This modular, agent-first approach enables you to move faster, iterate quicker, and deliver meaningful value to users in days or weeks—not months.
In doing so, agentic AI doesn’t just enhance the user experience—it also changes up the economics of enterprise software development. It lowers the barrier to innovation and dramatically reduces the time and cost required to build and deploy intelligent business solutions. In short, it’s opening the door to a new wave of digital transformation: one that’s faster, leaner, and more adaptive than anything we’ve seen before.
Closing Thoughts
While time will tell whether or not Microsoft’s “burn the ships” moment was prescient, the reality is that the age of agentic AI has arrived. What began as a few copilots tucked away in the margins is quickly evolving into a future where agents are not just enhancing traditional apps but functionally replacing them. As user experiences shift from form-driven to intent-driven, the software industry is entering one of its most significant transformations since the move to cloud computing we saw in the early 2010s.
For business leaders, developers, and IT teams, we think the message is clear. The tools, platforms, and design paradigms of the past are no longer enough to carry us into the next era. Agentic AI is changing how software is built, deployed, and used. Those who embrace this shift early will be best positioned to deliver faster innovation, better user experiences, and more flexible digital transformation strategies.
