AiMi: Q4 2025 Product Update

Q4 2025 focused on making change intelligence smarter, more explainable, and continuously aligned with evolving business requirements. Managing change is about understanding impact, validating decisions, and reducing risk. Too often that work is still manual, fragmented, and opaque.

To further streamline change management, we expanded our agents to also assess the impact of a change, learn from real outcomes, and provide a clear, auditable workflow for autonomous change creation and management.

Impact Agent

Once an actionable change has been identified, the next step is understanding the impact, a process that is typically manual and requires reconciling information across multiple sources.

Our latest foundational agent, the Impact Agent, evaluates each change using a set of weighted criteria and combines this with an understanding of which applications and services are in use to identify the actual impact.

To support this, we’ve introduced a new Subscriptions object. This captures trading and data subscriptions in a clear, organized way and links them directly to the applications that consume them. Subscriptions can be uploaded manually or automatically synchronised from source systems.

Learnings

As organisations evolve, it’s critical that agents evolve with them. Our new Learnings feature allows users to provide feedback on notices and changes, enabling agents to adjust future actions based on previous experience. The learnings are used by the agents to adapt their decisions to better reflect how your business operates.

Optimised Agent Instructions

Optimised Agent Instructions allow teams to describe their business and operations in plain language, whilst ensuring they are used consistently by the agents. This leads to more predictable behaviour and improves confidence, transparency, and accountability in how changes are handled.

Enhanced Agent Actions

To build trust in autonomous agents, teams require complete visibility into every decision. Enhanced Agent Actions provide a clear record showing how each notice was assessed, what was considered, and why specific actions were taken, or deliberately not. This gives teams confidence that automated decisions are transparent, explainable, and aligned with expectations.

Change Agent Preview

Safe experimentation is essential to successful agentic AI adoption. The new Change Agent Preview provides a sandbox environment where teams can validate agent behaviour and business context. This allows users to preview agent behaviour and outcomes without impacting live operations, making iteration faster and significantly lower risk.

Multi-Department Support 

Different departments within an organisation often have distinct workflows, tooling, and requirements. With multi-department support, teams can now operate independently within a single platform instance, each with their own agent configurations, application libraries, source context, and integrations (such as Jira). This allows departments to work independently while still benefiting from a shared, secure platform.

 Looking Ahead

Q4’s updates lay the groundwork for even deeper automation and learning in 2026. By combining richer impact analysis, autonomous, adaptive agents, and safer ways to iterate, we’re continuing to push toward a future where change is not just managed but understood.