Skip to main content
pmpmaker
← Back to Posts

Why I Built This Site

By Wesley Erickson

• 3 min • 599 words
announcement project-management
Key Takeaways
  • The systems outlasted the projects. The tools and trackers I built on the side kept running after every project closed, which eventually became the point.
  • Healthcare IT requires translation, not just delivery. Aligning clinical needs with technical constraints is the job, and Microsoft 365 and Power Platform are my primary delivery stack.
  • AI works when it is grounded in defined workflow. It fails when it is layered on top of broken process and called transformation.
  • This site is a public build log. What I built, what failed, what changed, and what I would do differently.

I have spent over a decade leading enterprise transformation work. The artifacts that kept getting used after project closeout were never the status reports or steering committee decks.

They were the systems.

The pattern that kept running

On nearly every project, I set up an MS Project plan and a SharePoint workspace. Over time, I started improving the system around the plan. I built structured Microsoft Lists trackers, refined workspace design, and used Power Automate to remove repetitive follow-ups.

Nobody put that in the charter. It kept running anyway.

That pattern showed up often enough that I stopped treating it as a side effect of project work and started treating it as the point.

By day: healthcare transformation delivery

I am a Business Initiatives Consultant at UnityPoint Health in the Transformation Management Office. I have led healthcare IT efforts across clinical and digital domains, including Epic integrations with medical devices, telehealth and virtual nursing initiatives, and AI-enabled monitoring workflows.

I also led a multi-year patient experience digitization program through the MyUnityPoint portal. Patients can now schedule appointments, complete pre-visit forms, sign documents, make payments, and access test results and provider notes. A parallel effort integrated online scheduling into the provider directory so new patients can search and book directly.

Healthcare IT requires constant translation between clinical needs and technical constraints. My core delivery stack is Microsoft 365 and Power Platform: SharePoint, Teams, Microsoft Lists, Power Automate, Power BI, Copilot, and Copilot Studio.

Where AI fits and where it does not

I have been building AI capability into project delivery since 2023 through PMI coursework and direct implementation work.

In practice, AI works when it is grounded in a defined workflow with clear ownership. It fails when it is layered on top of broken process and called transformation.

At work, I focus on Microsoft-first patterns that align with enterprise governance. In my homelab, I test broader tooling, including Hermes Agent, Claude, Gemini, Open WebUI, and Ollama, to pressure-test ideas before recommending them.

Building things breaks assumptions faster than reading about them.

The homelab

Outside work, I run a self-hosted homelab on Ubuntu 24.04 behind Cloudflare Tunnel.

The AI layer is Open WebUI backed by Ollama. I connect it to selected MCP tools so workflows can interact with live systems, not just generate text. n8n handles orchestration. Vaultwarden, Vikunja, and Obsidian LiveSync support day-to-day operations. Home Assistant runs separately on a dedicated Dell Wyse thin client.

I use this environment to test architecture patterns before I suggest them in enterprise conversations.

Why I publish this work

Most AI transformation content stays at the vision level. I care more about implementation that survives real operating conditions.

This site is my public build log: what I built, what failed, what changed, and what I would do differently.

If you are trying to reduce handoffs, clarify ownership, and make workflows actually run, this is written for you.

Keep Reading

Stay in the Loop