Microsoft is heavily pushing artificial intelligence (AI) as a focal point for the future of how businesses will operate, and even how people interact within companies.
As it gears up for its Build 2025 developer conference, the tech titan is painting a compelling picture where AI agents can operate across different worlds, sharing knowledge and even performing tasks for one another, which holds the power to radically change how intercompany workflows operate.
Recently, the company’s CTO Kevin Scott emphasized the company’s goal of building an “agentic web,” comparing it to the early days of the internet when browsers made it possible for information to be shared regardless of operating system.
The decomposable vision and a baseline for widespread inquiry into AGI encourages the support for the Model Context Protocol (MCP), an open-source standard promoted by Anthropic.
The goal of MCP is to support an exchangeable language for AI agents created by heterogeneous systems to interact and cooperate effectively.
This could lead to AI systems that could work in areas such as collaboratively debugging software with outside organizations or jointly analyzing complicated reports, regardless of their source.
The new development reflects a change from the current situation in which AI tools mostly work in siloed setups across individual businesses.
Opening up cross-company AI collaboration is Microsoft’s vision of the future, when companies will be able to have access to specialized AI assistants that can flow across AI systems from their suppliers, partners, and customers.
This interoperability would simplify convoluted intercompany processes, remove friction in joint projects and unlock new efficiencies.
For example, the realm of supply chain management is ripe for this kind of disruption, with AI agents from the manufacturer, the logistics operator, and the retailer all communicating in real-time, predicting potential disrupts, and optimising delivery schedules autonomously.
And beyond that, Microsoft is battling a serious limitation of today’s AI: retention. The company is working on a technique to improve the memory and contextualization of past interactions in AI agents, which it’s calling “structured retrieval augmentation.”
That would enable AI tools to develop a richer understanding of the context over time, minimizing the kind of repetitive inputs that make for stilted, noncontinuous and potentially less productive engagements across company boundaries.
Consider an AI assistant that has memory of previous conversations with a vendor’s AI about some project, and can expedite follow-up and troubleshoot, without needing a human intermediary to re-tell the context.
Although it carries computational costs that come from intelligent personal agents, Microsoft is convinced that in the long run the potential benefit of having intelligent, cooperative AI can outweigh these challenges.
This strategic thrust mirrors a more generalized industry momentum toward treating AI as a core underpinning for the digital workplace of today, where the bedrock of AI reaches from internal operations to literally reshape the structure of intercompany collaboration.
With the Build 2025 conference and Microsoft’s vision for it, the industry is closely watching the vision for the future of work as well.