▸ USE AI TODAY · a workflow to steal

Ask AI to grade its own answer and redo it as a 10

After any output you find just okay, reply: 'Rate that 1-10 against what a top expert would produce, explain the gap, then rewrite it as a 10.'

Why it works: Models settle for 'good enough' on the first pass. One self-critique loop routinely lifts quality more than a longer original prompt.

⌘ TODAY'S PROMPT · copy & paste  ·  Meetings

Turn a call transcript into a follow-up email + task list

You have a recording or transcript and owe people a follow-up.

From this transcript, produce two things: (1) a short follow-up email that thanks people, confirms what was agreed, and lists next steps; (2) a task list with owners and dates in table form. Keep the email under 120 words. Mark anything ambiguous as [confirm]. [paste transcript]

Why: The follow-up is where deals and projects keep momentum — this makes sending it a 2-minute job.

✦ GET BETTER AT AI

AI can state wrong facts confidently. For names, numbers, dates, and quotes, verify before you rely on them — ask it to cite the source, or check it yourself.

◦ Worth knowing in AI · 5 moves in full

◆ YOUR ROLE TODAY · pick your seat

One AI move + one copy-paste prompt built for your role. Read the one that's yours — 30 seconds each.

▣ FOR THE CMO  ·  marketing & growth

Ranking for one keyword is a dead end; owning a topic is durable. AI can turn a single subject into a content cluster — one authoritative pillar plus the supporting pieces that link to it — so you build topical authority instead of one-off posts.

⌘ Build a topic-cluster content plan · copy & paste

You are an SEO content strategist. For the core topic below, design a content cluster: one pillar piece (the definitive overview) and 6-8 supporting articles that each answer a specific question a buyer actually searches. For each, give the working title, the search intent it serves, and how it links back to the pillar. Flag the 2 pieces to write first for the fastest authority gain, and any subtopic too competitive to bother with. [state the core topic + your audience]

Why: Search and AI answer engines reward the site that owns a topic, not one lucky post.

▤ FOR THE CFO  ·  finance & operations economics

An audit or due-diligence request is calm if your house is in order and a fire drill if it isn't. AI can turn the likely question list into a readiness checklist — the documents, reconciliations, and explanations to have ready before anyone asks.

⌘ Get audit / due-diligence ready · copy & paste

You are preparing us for an audit or investor due diligence. For the scope below, produce a readiness checklist: the documents and schedules a reviewer will request, the reconciliations that must tie out, the accounting judgments they'll question (and the memo to have ready for each), and the 3 areas most likely to raise a follow-up. Rank by what to fix first and flag anything that currently lives only in one person's head. [describe the scope / what's being reviewed]

Why: Diligence goes smoothly for the CFO who prepared the answers before the questions arrived.

▧ FOR THE CSO  ·  strategy & planning

AI collapses the time from question to framework. Describe a decision and it will structure it — options, criteria, tradeoffs — so your team debates the substance instead of arguing about how to argue.

⌘ Structure a messy decision fast · copy & paste

You are a decision architect. For the decision below, produce: a clear statement of what we're actually deciding, 3-4 real options (including 'do nothing'), the 4 criteria that matter most, a simple scored comparison, and your recommendation with the biggest reason it could be wrong. [describe the decision]

Why: Gives the room a shared map so the meeting reaches a decision, not a loop.

▨ FOR THE CTO  ·  product & technology

Legacy and unfamiliar code is now readable on demand. AI can explain a system you inherited, so onboarding and 'why does this exist?' archaeology stop eating your week.

⌘ Understand code you didn't write · copy & paste

You are a patient staff engineer. Explain the code below: what it does in plain English, the non-obvious assumptions it makes, where it's fragile, and the 2 things I should be careful about before I change it. Then give me a one-paragraph summary I could put in a README. [paste code]

Why: Turns 'nobody knows how this works' into a documented system.

▩ FOR THE COO  ·  operations & execution

The 2026 ops question isn't whether to automate but which steps to hand off. AI can scan a process and flag the automation candidates — the repetitive, rule-based, high-volume steps — while marking the judgment steps that should stay human.

⌘ Find the automation candidates in a process · copy & paste

You are an automation analyst. From the process below, classify each step: automate now (repetitive, rule-based, high-volume), keep human (judgment, relationship, exception-handling), or human-in-the-loop (AI drafts, human approves). For the top 3 automation candidates, give the expected time saved and the one risk to control. Flag any step where full automation would be a mistake. [describe the process step by step]

Why: Automating a judgment step causes the mess; this tells you where the safe wins are.

🔭 Where the signals converge

Where independent intelligence streams — macro podcasters, our email-intel desk, AI-lab feeds — line up on the SAME read, cross-checked against our own live trading algo. The one AI brief that shows you where unrelated operators agree.

AI as an INFLATIONARY / capex-supercycle force (not just deflationary productivity) [MEDIUM · email intel + podcasts agree]

AI capex & compute economics — hyperscaler buildout / 'the AI trade' [MEDIUM · email intel + podcasts agree]

The cross-source read: Cross-source tape today: 2 of 10 convergent themes lean defensive/structural-risk. Highest-conviction edge: 'Mega-cap tech distribution / late-cycle topping' (3-layer agreement, incl. our own algo model). Convergence = where independent operators agree; that's the higher-conviction read for brief/newsletter/algo than any single source.

Where they split: Direction tally across convergent themes: 2/10 lean defensive/structural-risk — the CONVERGENT sub-themes are more informative than aggregate inbox sentiment (which is typically split).

▶ Watch this week · stay an AI expert

a16z Podcast

with Andreessen Horowitz · YouTube

Cross-industry view of how AI is reshaping sectors from healthcare to legal to logistics — useful when you want to see what's coming to YOUR industry before it arrives.

Know someone who'd get better at AI with this?

The AI Rundown is free — every story, every play, the full archive and the audio edition, always. The best way to support it: forward it to one person who'd actually use it.

Always free · one email a day · unsubscribe anytime