I was sipping my cafecito, letting my morning skim agent sweep the news, analyst notes, and client pings, when two need-to-know bullets popped up back-to-back.
First: Citi is training 175,000 employees on how to prompt. The article's key takeaway frames prompting as “the new Excel.” Translation: this isn’t a nice-to-have...it’s basic job literacy. I’ve been in those executive rooms; when a company that large says “everyone,” they mean it.
Second: a quieter but punchier newsletter about agentic AI. Essentially, sharing that it's not just about prompting the tools for an answer, but about designing little digital teammates that run steps for you, check their work, and request your approval. That one made me nod, because it aligns with what I’m seeing with my clients.
Prompting helps you participate in the AI era. Agents help you differentiate.
Here’s the picture I keep envisioning: two managers, same-size team, same tools. Manager A writes a solid prompt and gets a decent draft. Manager B turns their team’s SOP into a tiny agent that fetches sources every Monday, summarizes changes, flags risks, and drafts a 200-word update for review. After three weeks, Manager B isn’t just faster; they’ve built a repeatable system that the whole team can reuse. That’s the edge.
This makes me think of the times my mom used to say,
“No te quedes en la orilla si puedes cruzar el río.”
(Don’t stay on the edge of the shore if you can cross the river.)
Prompts get you to the water. Agents get you to the other side.
Myth: “Prompting is the finish line.”
Reframe: Prompting is the starting line. The edge now is turning your steps into small agents that do the work the same way every time, with receipts.
Myth: “Agents are only for engineers.”
Reframe: Many agent tools are no-code/low-code. The real skill is workflow thinking: clear steps, guardrails, and a human approval point.
This week, try the following framework and Prompts to Steal to help you take action. These will help stay ahead of the curve.
Agent.ai
No-code agent builder + professional network.
Insider tip: Browse and clone public agents, then tweak to your team’s sources and tone; publish your version to show measurable impact.
Instruct
A “Truly no-code” agent platform where you build and run agents in natural language.
Insider tip: Start with a 3–5 step workflow and add guardrails last; shorter chains ship faster.
CopyOwl
One-click research agent for deep background briefs.
Insider tip: Require source snapshots or links in every output so a human can verify before sharing.
Steal these prompts and use them in ChatGPT, Copilot, Gemini, Claude, or your company-approved platform. Doctor them up with your context.
Your move this week:
Pick one tiny task (weekly update, meeting recap, market scan). Use a prompt today, write the process tomorrow, and ship a mini-agent by Friday. Make it verifiable. Add citations, a “what changed since last week” section, and a required human sign-off before publishing.
It’s tempting to stop at good prompts and call it progress.
But the leaders who’ll stand out now are the ones turning know-how into repeatable systems. Small agents that run the play the same way, every time, with human judgment in the loop.
Prompting is table stakes. Agents are how you lap the field.
In this new world of AI, prompting is table stakes. Agents are how you lap the field.
¡Hasta la próxima, un abrazo fuerte!
(Until next week, a big hug!)