Welcome to this special report from Cloud Wars’ Agent & Copilot, where we analyze the opportunities, impact, and outcomes that are possible with AI.
In this episode, Brent Wodicka, CTO of AIS, discusses the software design requirements to make software accessible to AI agents as they create blended workforces with humans.
Highlights
Software for Blended Workforces (01:30)
Wodicka explains that most enterprise software was designed with the assumption of a human user, which becomes a bottleneck when agents enter the workforce. That software still functions but may be bypassed if it doesn’t adapt to the new workflow involving humans and agents.
Agent-Focused Design Requirements (03:58)
Design requirements in the AI era apply to both ISVs and in-house developers: they need to build structured, predictable interfaces. For ISVs, this integration may appear on a product roadmap, while for in-house developers, it involves rethinking the architecture of existing systems. Wodicka notes that agents’ effectiveness depends on the tools and data they have access to.
Community Summit North America is the largest independent innovation, education, and training event for Microsoft business applications delivered by Expert Users, Microsoft Leaders, MVPs, and Partners. Register now to attend Community Summit in Nashville, TN from October 11-15.
Creating Pathways (06:54)
APIs are the foundation for connecting agents to external systems. He explains that existing APIs were designed for system-to-system integration, not for autonomous agent use, leading to gaps in discoverability, error handling, and completeness. Command-line interfaces (CLIs) are another pathway — they are structured, text-based, and composable but they have limitations: they can’t reach mobile or web platforms. Structured tool interfaces provide a formal description of a tool’s capabilities in a machine-readable format, and machine-readable schemas, which are tailored towards software or machines.
Usable Work Surfaces and Guardrails (11:37)
To give agents the access they need, software must expose usable work surfaces, which are structured interfaces that allow agents to understand and take actions safely. Here’s an analogy: giving an employee access to a building versus an office with a whiteboard and computer. The latter is a usable work surface
Wodicka states that governance for agents needs to be more explicit and structural, including first-class identity, a permissions model scoped to tasks, time-granted access, and explicit human reviews at critical decision points.
Current State of Agent Readiness (16:11)
Most vendors are still in the early stages of meeting design requirements for agents to access their software; many are adding AI agents or co-pilots directly to the UI as a point solution. A few important platform-level plays, like Microsoft’s investments in Model Context Protocol (MCP) and Agent frameworks, are moving faster, but overall, the agent readiness question hasn’t been fully answered.
When I asked if building agent readiness will elongate software development cycles, Wodicka noted that it involves expanding the “design aperture” rather than starting from scratch.
More AI Insights:
- Brent Wodicka of AIS Details AI Organizational Model That Increases Human Scalability
- MCP Enablement Brings AI Automation to Dynamics 365 at Vast Scale
- Microsoft Agent Framework Enables Complex, Multi-Agent Actions
- Microsoft Aligns AI Agents With Traditional Software Development Workflows
Ask Cloud Wars AI Agent about this analysis
