The new work language: asking AI the right way
Two teachers sit down to build the same exam. One finishes in 15 minutes. The other takes an hour. The gap isn't dedication or talent. It's language - the ability to give AI clear, concise, contextual instructions.
AI prompting is no longer just a command. It's the new lingo required to operate in tomorrow's workplace. The advantage goes to those who can frame problems, set constraints, and ask better questions.
Why questions now signal leadership
"The biggest change is cultural, not technological," says Ignasi Llorente, a CEO advising companies on AI-driven shifts. Organizations are learning that AI isn't simply another tool. It's a new way of working.
The value has moved from tools to thinking. In a world where everyone has access to the same systems, the differentiator is the quality of the request. High-leverage prompting demands four things:
- Clarity: Define the outcome, not just the task.
- Synthesis: Cut noise. Provide only relevant inputs.
- Criteria: Spell out success standards and trade-offs.
- Context: Share audience, constraints, tone, and risks.
What executives should do this quarter
- Set prompt patterns: Publish 5-10 company "prompt blueprints" for core workflows (planning, briefs, customer emails, market scans). Make them easy to copy, adapt, and improve.
- Add review loops: Require a human verification step for decisions, numbers, legal text, and anything that affects customers.
- Create a no-AI zone for judgment calls: Strategy debates should use AI as input, not as a decider. Capture reasoning before any model output is considered.
- Measure outcomes: Track time saved, quality lift, error rate, and rework. Reward teams that improve prompts and documentation, not just speed.
- Invest in literacy: Treat AI skills like writing or spreadsheets. Budget time and training, not just licenses.
What writers should do now
- Push for originality: AI can draft. You supply angle, voice, and taste. Your edge is thinking, not typing.
- Front-load constraints: Audience, purpose, POV, must-avoid phrases, and examples. The first 10% of your prompt sets 90% of the outcome.
- Iterate in public: Keep a prompt library. Add before/after samples. Share what works with your team.
Prompt blueprints you can use today
- Executive brief (strategy)
"You are my strategy analyst. Goal: produce a 1-page brief to decide X. Audience: exec team. Context: [market, constraints, risks]. Criteria: concise, evidence-backed, options A/B/C with trade-offs. Output: bullet summary, assumptions, open questions, decision triggers." - Editorial outline (writer)
"You are my editor. Topic: [topic]. Audience: [who]. Angle: [original take]. Constraints: avoid clichés, no fluff, 1 stat per section. Output: headline options, outline with thesis, section goals, sources to fact-check."
The skills gap is linguistic, not technical
In schools, younger teachers often tap AI with precision while veteran peers use it sparingly or without full context. The advice that works across generations is simple: offer context, state the goal, and be direct about constraints.
In offices, forced rollouts backfire. One administrative team was told to write all emails with a model, which many read as a critique of their writing. Adoption rose only after leaders framed AI as a draft assistant, kept usage optional, and clarified where human voice still mattered.
Don't outsource thinking
Engineers like José Torró use strong prompts to automate hours of work, but technical depth still decides quality. Without domain knowledge - networks, architecture, development - the output drifts. Tools don't replace judgment. They amplify it.
The bigger risk is dependency. Some teams now defer big calls to a chat window and end meetings without conclusions. That's not efficiency. That's atrophy.
Guardrails that keep quality high
- Decision hygiene: For material decisions, write your reasoning first. Then consult AI. Compare, revise, decide.
- Verification rules: Fact-check numbers, sources, and names. Ban unverified citations.
- Source transparency: Flag any AI-assisted text in drafts so reviewers can apply the right level of scrutiny.
- Red-team critical outputs: Ask the model to critique its own answer. Then have a human do the same.
Beyond "faster and cheaper"
Leaders who treat AI as a typing assistant will stall. The real upside is amplification - surfacing options you wouldn't think of, testing assumptions quickly, and compounding team intelligence through shared prompts and patterns.
The organizations that win won't just automate tasks. They'll improve thinking at scale.
Education and policy are catching up
Researchers urge schools to teach AI prompting alongside reading and research skills, especially in secondary education where critical thinking takes shape. Expect policy to push basic literacy across sectors as regulation matures. For context on Europe's direction, see the EU's work on the AI Act here.
Five practice drills for your team this week
- Outcome-first rewrite: Take a recent prompt. Rewrite the first two lines so they state only the goal and constraints.
- Context pack: Build a 5-bullet context block you can paste into any prompt: audience, objective, constraints, tone, examples.
- Counter-prompt: After getting an answer, ask: "Argue against this. What did we miss?"
- Assumption audit: Extract all assumptions from an AI output. Mark each as validated or speculative.
- Style anchor: Provide two contrasting examples of your brand or voice. Instruct the model to justify style choices back to you.
Hiring and evaluation signals
- Defines clear outcomes before touching a model.
- Packages context succinctly and omits noise.
- Sets explicit criteria and constraints for quality.
- Verifies facts, cites sources, and documents changes.
- Improves prompts over time and shares them with the team.
Where to upskill
If you're formalizing training, start with role-based prompt practices and governance. You can browse practical courses by role and skill here: Complete AI Training - Courses by Job and focused prompt resources here: Prompt Engineering.
Bottom line
Access to AI is now a commodity. The moat is thought quality. In this economy, those who ask better questions will set the pace - and those who don't learn this new language will feel it in their results.
Your membership also unlocks:
