Your Signature AI-assisted Project – Target: 15 minutes 

You have spent twelve weeks building the foundation. You know how the tools work. You can prompt, iterate, check output, and get results that are genuinely useful. That is real. Most people around you cannot do that yet.

But here is the honest problem. Nobody else knows you can do it.

A skill that lives only in your own head — that has not produced anything a real person has seen and valued — is not a career asset yet. It is potential. And potential without evidence is not enough when you are trying to advance, change roles, win a client, or simply be taken seriously as someone who has genuinely embraced AI in their work.

That is what this block is about. Block 2 is called From Tool User to Career Changer because the shift we are making is from capability to proof. And we start with the most direct route to proof: completing one signature project that demonstrates real AI-driven value to a real stakeholder.

That is what this episode covers. By the end of the lesson, you will have chosen your project and a clear plan for completing it.

Core Concept

A signature project is not a practice exercise. It is not a prompt you typed into ChatGPT to see what happened. It is a complete piece of work — something a real stakeholder needed, that you delivered using AI in a way that made the outcome meaningfully better, faster, or more thorough than it would have been otherwise.

Let me be specific about what makes a good project. There are four criteria.

First: it involves a real stakeholder. This might be your manager, a client, a colleague, your board, or someone in your community. The point is that the work has to matter to someone other than you. Internal projects count — a report for your GM, a briefing paper for your team, a process audit for a colleague who needs it. External projects count too — a proposal for a client, research for a board paper, a strategy document for a community organisation. What does not count is something you built for yourself that nobody else will ever see or use.

Second: it normally takes two to four weeks. We are not talking about a quick win here. We are talking about something substantial enough that completing it with AI assistance represents a meaningful difference — in time, quality, depth, or scope. If the task would normally take an afternoon, it is probably too small. If it would take three months, it is probably too big to complete well this week. The sweet spot is something that would ordinarily be a serious undertaking — the kind of work that has been sitting on the back burner because you have not had the time or capacity to do it well.

Third: it demonstrates improvement that is explainable. When you hand over the finished work, you need to be able to say — honestly and specifically — how AI contributed to the result. Not just that you used AI, but what AI made possible that would not have been possible otherwise. Maybe you produced a more thorough literature review than you would have had time for. Maybe the draft came back faster and left more time for refinement. Maybe the analysis covered more scenarios than you could have worked through manually. The AI contribution needs to be real and articulable.

Fourth: it is documented. This is where most people miss the opportunity. They do the work. They hand it over. They get a good response. And then they move on and the evidence evaporates. A signature project is only useful to your career if you document it — specifically, as a case study. Before the project, write two or three sentences about what the work was, who needed it, and what the normal approach would have been. After the project, write two or three sentences about what you actually produced, how AI was involved, and what the outcome was. That before-and-after is the beginning of your portfolio.

Now let me talk about the workflow. This is the four-tool sequence that works well for projects like this.

You start with Perplexity for research. If your project requires any kind of background knowledge — market context, sector data, precedent cases, relevant frameworks — Perplexity is your starting point. It is faster and more reliable for grounded research than most AI tools, and it cites its sources, which matters when the output is going to a stakeholder.

From there, you move to Claude for analysis. Once you have your research inputs, Claude is where you work through the thinking — structuring an argument, identifying the key issues, stress-testing a position, comparing options. Claude handles complexity and nuance well, and it is good at holding a large amount of information in view and drawing out what matters.

Then you go to ChatGPT for drafting. Once the thinking is solid, ChatGPT is a strong drafting tool — particularly for documents that need to be readable and well-structured. Briefing papers, proposals, reports, executive summaries. You give it the structure and the substance; it produces professional prose that you then refine.

The final step — whichever AI tool you prefer — is refinement. You read the output carefully. You check for accuracy. You adjust the tone where needed. You make sure the AI contribution is invisible in the sense that the final document reads as if a sharp professional produced it — because one did. You.

Let me also be honest about where this breaks down. Three failure modes are worth knowing before you start.

The first is choosing a project that is too small. If the project does not feel like genuine work — if it would not be impressive to a manager or client — it will not be impressive in a portfolio either. Set the bar honestly.

The second is not documenting as you go. By the time you finish, the details of what you did and how AI contributed will have faded. Write the before description before you start, and the after description within twenty-four hours of finishing.

The third is hiding the AI involvement. Some people are nervous about this. They feel like using AI makes the work less theirs. That is worth examining. The work is yours — the judgment, the scope, the quality standard, the stakeholder relationship. AI is the tool you used to produce it. A journalist who uses a camera is still the journalist. A consultant who uses AI is still the consultant. Own it.

The three levels

Let me walk you through this week’s exercise. As always, choose your level based on what is realistic this week. All three are legitimate.

Minimum — twenty minutes. At the Minimum level, you are working with something that already exists. Take a deliverable you have produced recently — a report, a briefing, a proposal, a presentation — and use the four-tool workflow to improve it in one specific way. Maybe you run it through Claude to strengthen the argument. Maybe you use ChatGPT to sharpen the executive summary. Maybe you use Perplexity to fill a gap in the background evidence. Pick one concrete improvement and make it. Then write your before-and-after summary: three sentences about the original and three sentences about what changed. That is your Minimum output — a real before-and-after documented for your portfolio.

Standard — thirty-five minutes. At the Standard level, you are completing a new project from scratch. Use the project selection criteria to identify something real — a genuine stakeholder, a genuine need, something substantial enough to matter. Complete the full project using the four-tool workflow: research with Perplexity, analysis with Claude, drafting with ChatGPT, refinement at the end. Then produce a one-page case study covering what the project was, who it was for, how AI was used across the four steps, what the outcome was, and what you could not have produced without AI. If you can, share the work with the stakeholder and get a response — even an informal email saying it was useful is worth capturing.

Ambitious — fifty to sixty minutes. At the Ambitious level, you are targeting something substantial — the kind of project that would normally consume a full working week. Use the full four-tool workflow, document every step as you go, and produce a comprehensive case study with specific metrics: time saved, quality difference, depth of analysis. Where possible, present the final work to the stakeholder — even a short conversation about what you produced counts as a stakeholder presentation. This level produces portfolio-quality output: the kind of case study you could reference in a performance review, a job interview, or a client conversation.

Action and close

Here is exactly what I want you to do after this episode.

Go to the lesson. Read the key concepts section — it is a short read that reinforces what we have just covered. Then download the exercise guide PDF. It has the full instructions and templates for whichever level you have chosen.

Before you close the lesson, decide which project you are going to do. Write the project down — in one sentence: who the stakeholder is, what the deliverable is, and what the AI contribution will be. That sentence is your commitment. If you cannot write it in one sentence, the scope is not clear enough yet.

Complete the exercise by end of this week. The project does not have to be perfect. It has to be real, delivered, and documented.

The PDF cheat sheet covering all key frameworks from this block is available at the end of the course — it will include this episode’s project selection criteria and case study template.

You have the capability. This week, you start proving it.