Microsoft’s latest Build conference introduced a significant shift from basic AI assistants to autonomous AI agents capable of independent reasoning, decision-making, and task completion. The emerging “agentic web” will transform how businesses operate, with significant implications for organisational structure, governance, and competitive advantage. This week, we delve into these announcements and what they mean for the enterprise.
Focus On: How Microsoft’s Vision Transforms Enterprise AI Strategy
In our previous discussions on agentic AI and machine customers (see our November special issue), we examined how autonomous AI systems would transform business operations and customer engagement. Microsoft’s Build conference last week confirmed this direction is accelerating far faster than many anticipated.
Microsoft showcased the vision for what they are calling the “open agentic web”—a new ecosystem where AI agents make decisions and perform complex tasks independently across individual, organisational, and business contexts.
From Assistants to Agents
When interviewed on a podcast (available soon), I was asked what is new about agentic AI? Why do we need a new term to define what are, essentially, assistants. My reply has been that the difference is profound and has significant strategic implications because:
Assistants require explicit human direction and oversight for each task
Agents possess reasoning capabilities, maintain memory, and can autonomously pursue goals, often chained in sequence and across different AI technologies
Microsoft’s announcements—from GitHub Copilot’s evolution to an asynchronous coding agent to multi-agent orchestration in Microsoft 365—demonstrate how quickly we’re moving from human-directed AI to human-governed AI, where systems operate with significant autonomy, staying within defined boundaries.
Implications for Enterprise Leaders
For C-suite executives there are a few critical considerations. Firstly, as Microsoft revealed with its multi-agent orchestration capabilities, the future lies not in deploying individual AI solutions but in creating ecosystems of specialised agents that collaborate. This shifts the strategic focus from implementation to orchestration—how will your organisation design, govern and optimise interactions between multiple AI agents?
A new governance and risk management challenges emerges: “agent sprawl.” As autonomous agents proliferate across the enterprise, managing their identities, permissions, and actions becomes mission-critical. In the near future we should consider:
How will you assign responsibility for agent actions?
What governance framework will manage agent permissions and capabilities?
How will you monitor and evaluate agent performance?
Finally, Microsoft’s support for the Model Context Protocol (MCP) and introduction of NLWeb is an important signal: the future of AI agents will be built on open standards. This creates opportunities for early adopters to shape how these standards evolve within their industries.
Moving from Experimentation to Strategy
The acceleration of agent technology demands a shift from experimental pilots to comprehensive strategy. The base rules, though, remain the same:
Audit your current AI landscape: Identify opportunities for agent implementation with clear business objectives
Develop an agent governance framework: Establish clear policies for agent identity, permissions, and oversight
Invest in orchestration capabilities: Build the technical and organisational infrastructure to manage multi-agent workflows
Assess your data readiness: Ensure your data architecture supports the training and operation of domain-specific agents
Follow me
That’s all for this week. To keep up with the latest in generative AI and its relevance to your digital transformation programs, follow me on LinkedIn or subscribe to this newsletter.
Disclaimer: The views and opinions expressed in Chronicles of Change and on my social media accounts are my own and do not necessarily reflect the official policy or position of S&P Global.