Don't Know Where to Start with AI? Begin with Your Worst Task

"I don’t even know where to start with AI."

We hear this from clients all the time.

In healthcare especially, the idea of AI transformation can feel overwhelming. Systems are highly interdependent — your Epic or EHR instance connects to insurance billing platforms, patient-facing messaging tools, clinical decision support tools, and patient portals. Everything has been carefully constructed to work together, usually after years of painstaking integration.

It can feel like if you touch one piece of the system, you risk unraveling the whole ball of yarn.

And then there’s the pace of AI itself. Many clients tell us: “Anything I want to try already feels out of date” “If it’s not cutting-edge, is it even worth doing?” "I was just learning how to write a good prompt, and now everybody is saying things like 'agents' and 'swarms'... I'm so far behind!"

If you lead a small to medium-sized healthcare provider, or you work inside a health tech company, these feelings are perfectly normal — and paralyzing.

So where should you start?

Here’s our advice: Start with the job you hate.

Why Start with the Worst Jobs?

When we say "the job you hate," we mean the task you dread doing — the one that drains your time and your spirit. It could be a task in your own personal workflow, or a responsibility your team collectively wishes would disappear.

It’s often a job that is:

✍️ Repetitive

😓 Tedious

🧠 Mentally draining

😔 Unrewarding

These are perfect conditions for automation and augmentation with AI. Starting here has several benefits:

• It delivers immediate personal value — you’ll get time and sanity back.

• It creates a safe space to experiment with AI without overhauling core clinical or operational workflows.

• It builds confidence and momentum inside your team.

  
  

Some "Jobs You Hate" — Real Examples

We see these types of jobs across healthcare organizations:

Clinicians

• Manually entering encounter data or billing details.

• Rewriting patient notes to fit coding requirements.

• Confirming test results have been delivered and logged correctly.

Administrative & Operations Teams

• Ordering and tracking supplies.

• Manually scheduling follow-up appointments and sending reminders.

• Processing inbound faxes (yes, they still exist, and no one likes them).

• Updating software across all laptops and tablets.

Health Tech & Product Teams

• Generating standardized operational reports.

• Triage of user support tickets.

• Drafting templated customer messaging.

Pro tip: Ask your team “What is a task you would be thrilled to never do again?” — that’s your starting point.

How to Start Automating the Pain Away

Don’t launch a massive AI initiative right out of the gate. Start small:

1. Try an Off-the-Shelf Tool

   1. Use something like ChatGPT, Gemini, Claude, or Microsoft Copilot.

   2. Give it one instance of the hated task:

      1. Draft a message.

      2. Generate a form entry.

      3. Summarize and tag a fax.

      4. Clean up a messy spreadsheet.

      
      

      Important: Don’t expect perfection on the first try — iterate and experiment.

2. Explore No-Code / Low-Code Tools

   1. If a task needs more automation, explore tools like:

      • Lovable

      • Zapier

      • Make

      • Retool

      
      

      Example: build a web form to auto-read inbound faxes and file them with patient records — no more manual entry.

3. Rule of Thumb

   1. If AI can save you even one painful hour per week, that’s a win worth pursuing.

Congratulations — You’ve Built a Proof of Concept!

You now have a live example of AI automation working inside your team.

🎉 Celebrate it. Show it off. Build curiosity. 🎉

Too many teams hesitate after small wins — they think “This isn’t a big enough project to matter.” But that’s exactly how real AI transformation starts: one valuable, practical step at a time.

Ready for More? We Can Help.

At Third Axis Consulting, we help healthcare organizations and health tech companies move from proof of concept to pilot and scale.

We can help you:

🔎 Identify the best candidates for AI-driven automation

🏗️ Build scalable prototypes

🗺️ Navigate compliance and regulatory risks

🧍 Manage organizational change and staff adoption

💥 Avoid common pitfalls of “AI hype” and focus on real value

AI doesn’t have to start with a moonshot. It can start with the job you hate.

And once that job is automated, you’re ready to tackle the next.

Stay tuned — in future posts, we’ll cover how to pilot and scale your AI initiatives for sustainable success.