How we use AI to support planning processes smarter

By Jerry Suiker, Power Platform Consultant, Dynamics4Business, Jan. 28, 2026

At Dynamics4Business, we are continuously investigating where AI and Copilot Agents can actually add value. Not as a loose experiment, but as practical support for existing processes. A recent example is the development of a Copilot Planning Agent, built to support a recurring planning process.

Why a planning agent?

Planning processes are often repetitive and require coordination between multiple people, availabilities and ground rules. In practice, this takes time and is error-prone, especially when done manually. The question we asked was simple: can a Copilot Agent help support this process more consistently and efficiently?

An agent who supports, not takes over

I developed a Copilot Planning Agent that:

  • Works with predefined ground rules
  • takes into account availability and preferences
  • Supports planning
  • Helps structure a recurring task

The agent is set up as a tool, with humans always in control. Exactly how we prefer to deploy AI.

What I learned while building

During development, I encountered typical AI challenges. Instructions that are interpreted differently than intended. Answers that keep coming back. Or suggestions that seem logical but don't make sense process-wise yet.

These experiences provided valuable insights:

  • the importance of clear and unambiguous instructions
  • incorporating checks and validations
  • testing different scenarios
  • Thinking from process logic, not just technology

These are lessons that I bring back to every subsequent Copilot Agent.

Knowledge sharing and further development

I shared the learnings from this use case during an internal team day. This way, we not only build solutions, but also collectively build knowledge about what works and what doesn't when developing AI agents.

AI only works when it fits your process

This case shows that AI only becomes valuable when it connects to a concrete process. No loose experiments, but targeted applications that support where it makes sense. Wondering where AI can really make a difference within your organization? We would like to think along with you.

Content

Read the next article

February 18, 2026
Blogs

Automated testing: not a differentiator, but a requirement