🎮 The Next Input — Issue #147

In partnership with

⚡ The Briefing — 60 sec

• Brutal AI prediction that will shock Australia and the world Not to scare you, but that comfy job title is not exactly bulletproof anymore. The “AI will reshape work” line is now well past prediction stage and drifting into operating reality.

• AI is quietly redrawing career paths – and the divide is already visible This is literally the point of this newsletter. The people learning to work with AI are already pulling ahead, while everyone else is slowly finding out the ladder has moved.

• The Pentagon is developing alternatives to Anthropic, report says When institutions this large start building fallback stacks, the message is obvious: nobody wants to be fully dependent on a single model vendor. Even the AI race is becoming a hedging game.

🛠️ The Playbook — The AI Career Moat Engine

Mission
Build a repeatable internal system that helps teams become AI-leveraged operators instead of AI-exposed bystanders.

Difficulty
Intermediate

Build time
3–4 hours

ROI
Faster output, stronger internal capability, and a workforce that compounds with AI instead of getting outpaced by it.

0) Why This Matters

The career divide is no longer theoretical.

Recent HR reporting says frequent AI users are earning more, improving performance faster, and feeling more confident about career progression than non-users. In that piece, Human Resources Director cites research from Northern Kentucky University showing frequent AI users reported earning about 28% more annually than non-users, while also being more likely to say their performance improved.

At the same time, the bigger institutions are signalling the next phase of the market. The Pentagon is reportedly building alternatives to Anthropic after a breakdown in its relationship with the company, which is a clean reminder that serious operators do not want their core workflows tied to one provider.

So the play here is not “use AI a bit more.”

It is:

• make AI fluency part of daily work

• turn prompting into workflow design

• build portable capability across tools and models

• create a practical moat around your people and processes

1) Architecture

2) Workflow

1. Identify 10 recurring tasks where team members spend time drafting, researching, summarising, or structuring information.

2. Classify each task into automate, assist, or keep-human based on judgment and risk.

3. Build one strong prompt and one fallback prompt for each task across at least two model providers.

4. Run the tasks live for two weeks and compare output speed, quality, and edit load against the old method.

5. Store the best prompts, examples, and failure cases in a shared prompt library.

6. Train the team on the workflow, not just the tool, so capability survives vendor shifts.

3) Example Prompts

Task Rewriter

You are an AI workflow designer.

Take the task below and convert it into:
1. a repeatable AI-assisted workflow
2. the parts that should stay human
3. the best prompt structure for the AI portion
4. the likely failure modes

Task:
[insert task]

Career Moat Prompt

You are reviewing a team's workflow for AI exposure risk.

For the role below, identify:
- which tasks are most vulnerable to AI replacement
- which tasks become more valuable with AI leverage
- what skills the person should build immediately
- one weekly practice routine to stay ahead

Role:
[insert role]

Cross-Model Fallback Prompt

You are rewriting this prompt so it works across multiple AI models.

Requirements:
- remove provider-specific assumptions
- keep instructions clear and structured
- include output format
- include quality checks

Original prompt:
[insert prompt]

Manager Review Prompt

Review this AI-generated output as a team lead.

Check for:
- factual weakness
- missing context
- poor judgment
- tone problems
- anything that still needs human intervention

Return:
1. approve
2. edit
3. reject
With a short reason.

4) Guardrails

• Do not confuse AI access with AI capability.

• Train on live workflows, not toy examples.

• Keep prompts portable across more than one model.

• Require review on anything client-facing, strategic, or sensitive.

• Track where AI genuinely improves performance versus where it only feels faster.

• Build judgment alongside tooling, or you just scale sloppiness.

5) Pilot Rollout — 3 hours

1. Pick one team and list its top 10 repeatable knowledge tasks.

2. Select 3 tasks with obvious upside and low downside.

3. Write a primary prompt and a fallback prompt for each task.

4. Test the same workflow in at least two models.

5. Measure time saved, quality of first draft, and amount of human correction needed.

6. Publish the winning workflows into a shared library and train the team on them.

6) Metrics

• Weekly hours saved per operator

• First-draft acceptance rate

• Human edit load per output

• Adoption rate by team

• Number of reusable prompts in production

• Percentage of workflows that work across multiple models

• Change in employee confidence on AI-assisted tasks

Pro Tip: The safest career moat right now is not “being good at ChatGPT.” It is knowing how to turn messy work into portable AI-assisted systems.

🎯 The Arsenal — Tools & Platforms

• ChatGPT · flexible daily AI workspace for drafting, synthesis, and workflow execution · ChatGPT

• Claude · strong reasoning layer for structured writing and critique · Anthropic

• Gemini · useful third model for portability and workflow redundancy · Gemini

• Airtable · simple task inventory and adoption tracking layer · Airtable

• Notion · shared prompt library and workflow documentation hub · Notion

Copy-paste prompt block:

You are helping me build an AI Career Moat Engine for my team.

For the workflow below:
1. break it into discrete tasks
2. classify each task as automate, assist, or keep-human
3. identify the skills that become more valuable with AI
4. write one strong primary prompt
5. write one fallback prompt for a different model
6. list the top 5 failure modes
7. propose a 2-week pilot

Workflow:
[insert workflow here]

Return the answer in markdown with sections for:
- Task map
- AI leverage opportunities
- Human-only tasks
- Primary prompt
- Fallback prompt
- Risks
- Pilot rollout
- Metrics

💡 Free Office Hours

If you are trying to make your team more AI-capable without locking everything to one tool or one vendor, I run free office hours to help map the workflow, the prompts, and the fastest pilot path.

Book here: https://calendly.com

Here’s how I use Attio to run my day.

Attio is the AI CRM with conversational AI built directly into your workspace. Every morning, Ask Attio handles my prep:

• Surfaces insights from calls and conversations across my entire CRM

• Update records and create tasks without manual entry

• Answers questions about deals, accounts, and customer signals that used to take hours to find

All in seconds. No searching, no switching tabs, no manual updates.

Ready to scale faster?

Start now →

🕹️ Game Over

AI is redrawing the ladder in real time. Better to build a moat than stand there arguing with the excavator.

— Aaron Automating the boring. Amplifying the brilliant.

Subscribe: link