4 human skills that are worth more than a perfect AI prompt
Execution is getting cheaper by the second. Here's a formula that maximizes your market value.
Last week, we talked about the seismic shift rocking our work lives: how jobs are being sliced into tasks, how AI is absorbing the “busy work,” and the erosion of entry-level roles.
We can’t stop companies from reorganizing workflows around AI. But we can control how we show up in this new landscape. As we watch AI churn out reports and slide decks in under 30 seconds, the question we all need to ask ourselves is:
When a machine can do the “doing,” what makes me irreplaceable?
The end of just doing the work
For decades, companies bundled tasks into roles because coordinating humans required a lot of heavy lifting. We needed judgement, memory, and for people to be in the same room. AI is in the process of collapsing those “coordination costs.”
Now, anything that can be measured or standardized is at risk of being automated or outsourced. What’s left? The messy, high-stakes, “human-in-the-loop” parts of work that require accountability.
Doing the work is cheap. Overseeing the work is where the value (and the money) is.
If AI can generate a marketing plan or a financial forecast in the time it takes you to grab a latte, execution isn’t the bottleneck. The bottleneck is discernment, or knowing what to trust, what to do with it, and who takes the fall when things go wrong.
I’m a trained journalist, and my writing process looks a lot like a collab with ChatGPT, Gemini, Perplexity, and Claude these days. But because I know the steps: how to fact-check, how to spot a “hallucination,” and where AI hits a wall, I’m the one in the driver’s seat. Just asking AI to write an article doesn’t make it a good article.
The new career currency
To stay relevant in 2026, we need to trade in a new kind of currency. Think of this as your career “cheat code”:
Market Value = Critical thinking x Judgement × Communication × Learning velocity
It’s not about competing with AI, it’s about becoming the person who directs, evaluates, and adds context to what AI produces. In real life, that’s:
Audit, don’t just deliver: Don’t just hand in a project, show how you evaluated it.
Design, don’t just use: Don’t just learn a new tool, show how you redesigned a workflow.
Advise, don’t just assist: Document the trade-offs. Explain your “why.”
1. Critical thinking: Your professional BS detector
Think of critical thinking as your first line of defense against AI-generated nonsense. It’s not about being unreasonably skeptical, it’s about spotting when something looks right, but is built on shaky ground.
AI consistently produces polished reports, but the risk is that we’re drowning in work that looks legitimate but isn’t. We’ve seen this in the legal world, where lawyers have been fined for submitting AI-generated filings with fake citations. As one judge put it: “The duty to check sources remains unchanged.”
The takeaway: Your value isn’t in how much you produce. It’s your ability to look beyond the polish and know when something is wrong.
How to build it: Treat every AI output like a rough draft, never the final answer.
The Proof Point: Keep an audit log of mistakes you’ve caught. Showing a hiring manager how you think is the ultimate flex.
2. Judgement: Where the buck stops
Judgement is what turns raw info into action. You need it when there’s no clear “right” answer, whether you’re weighing ethics, office politics, or how to balance efficiency with trust.
AI can give you options, but it’s not going to take responsibility for the outcome. When those lawyers got caught using fake citations, their firm had to admit: “There is no excuse.” The tech was plausible, but the humans owned the failure.
The Proof Point: Create a one-page “decision memo” for your portfolio. Show the options, the trade-offs, and why you made the call.
Get started with this article: The Secrets of Better Judgement
3. Communication: Meaning-making
Communication is about making meaning out of noise. You need to be able to:
Translate AI insights for people who aren’t “techy.”
Be honest about uncertainty and limitations.
Navigate the emotional fallout when a process gets automated.
The Proof Point: Show a “Before and After.” Present the raw data next to the version you tailored for humans, with notes on what you changed and why.
I highly recommend Charles Duhigg’s book Supercommunicators - I borrowed the audiobook from my local library. It’s excellent.
4. Learning velocity: Future-proofing your skillset
Tools now have a shorter shelf life than ever. Your edge is no longer what you know, but how fast you can unlearn the old way and update your systems.
The Proof Point: Document a workflow you rebuilt from scratch. What did you intentionally stop doing?
I’ve hired a lot of people in my time, and being able to walk me through your thought process is the fastest way to get a “yes.”
Watch this video from Sandeep Swadia (TheMITMonk) for a masterclass in learning how to learn:
What this means for your career stage
If you’re early-career: Let’s be honest: this makes things harder for you. The “training wheels” tasks like basic research, drafting, and routine analysis are exactly what AI does best. The ladder is thinning, and the “entry-level” bar is higher than ever.
As Kevin Roose put it when he wrote about the trend last May: employers are increasingly saying, “These tools are so good that I no longer need marketing analysts, finance analysts and research assistants.”
Along with your resume, build a Portfolio of Visible Reasoning. Show how you evaluated info, how you made decisions, and how you translated complexity. These are the receipts that prove you can handle higher-level judgement.
Show how you evaluated information
Document decisions and what you learned
Demonstrate how you translated complexity for real people
These can live on your personal site, GitHub, Notion, or as portfolio attachments. They show decision-makers what your resume can’t: visible reasoning.
If you’re mid-career: Execution skills eventually plateau. Now, influence goes to the people who can navigate the messy stuff like ambiguity, risk, and organizational reality. Your value is no longer doing the work, it’s shaping how the work gets done.
Ethan Evans’ newsletter, Level Up is a great resource to learn how to influence and navigate the corporate world.
It’s not about adapting faster
It’s really important to play with AI tools and get some real experience with them. But all those tools will be out of date in 6 months. The need for human accountability isn’t going anywhere.
In this new world of work, your only real leverage is how well you think, decide, explain, and adapt.
Let’s talk: I’m curious. especially for those of you in hiring positions, what’s the one “human” skill that makes a candidate stand out to you right now?
Drop your thoughts in the comments.
AI in the news
AMD’s Lisa Su says AI isn’t replacing people, but is changing who gets hired (CNBC) Lisa Su says AI hasn’t slowed hiring at Advanced Micro Devices, but it has changed who gets hired, with priority going to candidates who are “AI forward” and actively integrating the technology into their work. While some policymakers warn AI may dampen hiring overall, Su argues AI is augmenting productivity at AMD rather than replacing workers, allowing the company to grow and bring more products to market.
Google and AI start to settle lawsuits alleging chatbots led to teen suicide (Guardian) Google and Character.AI have reached a settlement in principle to resolve multiple lawsuits alleging their chatbots harmed minors, including a Florida teenager who died by suicide. The agreement covers cases across four states, though the specific terms remain undisclosed and subject to court approval.
OpenAI unveils ChatGPT Health, says 230 million users ask about health each week (TechCrunch) OpenAI announced ChatGPT Health, a dedicated space for health-related conversations that keeps medical context separate from other chats, can integrate data from apps like Apple Health and MyFitnessPal, and will not be used to train models. While pitched as a way to improve access and continuity of care, OpenAI stresses it’s not for diagnosis or treatment, and the rollout raises concerns about AI hallucinations and accuracy in medical advice.