If you feel behind on AI, you are not alone.

Even people working at the leading edge can feel behind because the field advances faster than any one person can fully absorb.

That does not mean we should stop trying to keep up. It means we need to continually experiment and focus on the capabilities that will matter even as the tools change.

Let’s establish some practical goalposts to help you assess your AI fluency and identify what you should build next.

AI can serve two complementary roles.

As a thought partner, it can help you challenge assumptions, examine alternatives, synthesize information, and make better decisions.

As a doing partner, it can help you research, organize, analyze, draft, build, and execute.

Using AI as a replacement for a traditional search engine can be useful, but retrieving an answer is not the same as strengthening your thinking or advancing meaningful work.

If you use AI as both a thought partner and a doing partner, you have established the right foundation. The next step is making that usage intentional, repeatable, and capable of improving over time.

The emerging divide is no longer simply between people who use AI and people who do not. It is between people who occasionally prompt AI and those who intentionally design systems around it.

A prompt can improve a task. A system can improve how you work.

The In Pursuit of Elite writing workflow uses a Project, Instructions, Skills, and improvement Loops. Agents are available when a workflow benefits from an additional execution layer.

The foundational elements of an effective AI operating system can be understood through a simple progression:

Place → Direction → Capability → Action → Improvement

The example below shows how these building blocks can create an efficiency flywheel.

The flywheel leverages the best of AI for speed, scale, and iteration while keeping human judgment responsible for direction, verification, and final approval.

By combining a Project with clear Instructions, reusable Skills, and an improvement Loop, I am building a system that becomes incrementally more efficient each time I use it and retain what works. It saves me time while helping me maximize the value delivered to the reader.

A Project gives an ongoing body of work somewhere to live.

It can hold conversations, source material, instructions, decisions, drafts, and completed outputs. Instead of starting every interaction from zero, you create an environment where related work can accumulate.

This newsletter is being developed as part of my In Pursuit of Elite Project. The Project contains the publication context, while separate threads make individual posts easier to locate and continue reading.

In this screenshot, you will see the In Pursuit of Elite writing project, along with supporting threads listed below in chronological order.

Starting a new thread within the same Project can serve as a safeguard. Very long ChatGPT or Claude conversations can make it harder for the system to consistently identify and apply relevant information, increasing the risk of mistakes, inconsistent answers, and occasional hallucinations.

When the task, phase, or deliverable changes, start a new, focused thread to provide clearer context and more reliable work.

A Project gives the work a place.

Instructions tell the AI how the work should be approached and can reduce the need to repeatedly explain your standards.

They can establish objectives, standards, preferences, constraints, responsibilities, and ways of working. My In Pursuit of Elite Project instructions define the development gates, supporting Skills, evaluation threshold, revision process, and human responsibilities.

For reference, here is a screenshot of part of the project instructions:

Without instructions, you may repeatedly explain what you want. With them, the standards travel with the work.

The human determines the direction. AI follows the instructions.

A Skill packages a methodology, expertise, workflow, or supporting resources for repeated use.

A Project organizes a body of work. A Skill packages a repeatable capability.

Because Skills are reusable, I can apply them inside other Projects or invoke them independently for ad hoc work.

In a corporate workspace, Projects, Skills, and their underlying standards can also be managed and shared across teams, subject to the platform’s capabilities and workspace controls. This can turn individual efficiencies into greater consistency and leverage across the organization.

My In Pursuit of Elite Project taps into three Skills:

This is not about relinquishing my voice or judgment. The writing Skills codify and protect how I write, while the focus group Skill pressure-tests the work from multiple perspectives. Together, they enable ChatGPT to function like an editing team while I retain responsibility for the ideas, direction, and final approval.

My synthetic focus-group article describes how I originally built that panel inside a ChatGPT Project. I have since converted it into a reusable Skill while preserving the fixed panel and evaluation protocol.

The panel provides structured synthetic feedback. It is not customer research, market validation, or statistical evidence.

Agents help move an outcome through multiple steps. An agent might research a subject, organize its findings, apply defined criteria, make bounded decisions, and deliver the result.

In Staying AI Fluent, I shared how I use Manus agents each week to monitor selected thought leaders, identify relevant AI content, summarize it, and send the results to specific email addresses.

Agents are an optional execution layer in some workflows and a core requirement in others, especially when work must run with limited prompting, make decisions across multiple steps, and deliver an outcome on a set schedule.

Agentic workflows require clear outcomes, appropriate access, defined boundaries, verification, and human oversight.

Loops help drive measurable, verified improvement.

The first output is rarely the best. A disciplined loop moves the work through:

Create → Critique → Compare → Revise → Test

This newsletter uses gates for the concept, outline, draft, and final publication review. At each gate, the Leadership Recall Focus Group independently evaluates the work. Meaningful weaknesses trigger revision and another test.

The target is for at least 17 of the 20 panel members to rate the work excellent. But the score does not replace judgment. A factual problem, credibility issue, or important reader objection can still require revision.

Looping does not mean polishing forever. Stop when the standards are satisfied, and another cycle is no longer worth the marginal improvement.

Block time every week to learn and experiment with AI workflows. The goal is simple: reclaim minutes, and sometimes hours, from recurring work and reinvest that time in the high-value work you are uniquely wired to deliver.

That is where AI creates real freedom.

Choose one recurring body of work:

The leaders who create the most leverage with AI will not necessarily know the most tools. They will deliberately combine context, direction, capability, action, and improvement.

Look at the AI work you repeat every week. Choose one recurring workflow and turn it into a system.

There are two ways I can help.

If you want help putting these foundational AI concepts into practice, Special Ops is an accessible monthly membership inside my Elite Ops Academy built around focused learning, practical application, and accountability.

As leaders and operators, we must also continually tune our Learning Operating Systems to keep pace with developments in AI and AI tooling. If you do not have a deliberate way to discover, capture, and recall important developments from curated sources such as Feedly and carefully selected YouTube channels, you are operating at a disadvantage.

You can get a no-cost copy of Leadership Recall to help strengthen that system.

If you’re interested in Special Ops, reply to this email, and I’ll send you the details.

Let’s leverage our time better,

Steve Kahle |  ELITE OPERATORS
In Pursuit of Elite