There's a particular kind of mental overhead that doesn't get talked about enough.

It's not the urgent stuff, like emails, deadlines, and the actual work. It's the ambient guilt of everything you're not doing. The stack of newsletters you've bookmarked with great intentions. The Friday afternoon nagging feeling that something important happened in your industry last week and you were too busy to notice.

Psychologists call this kind of thing attentional residue: the way unfinished tasks keep hogging mental bandwidth even when you're supposed to be focused on something else. It's a low-grade drain. Annoying enough to notice, not bad enough to fix.

Until now.

Scheduled AI reports exist, they're simple to set up, and they're very useful.

ChatGPT, Claude, Gemini, and Microsoft Copilot* all let you schedule a recurring prompt. You write it once, set when it runs, and it delivers a fresh report on autopilot - weekly, daily, monthly, whatever suits you.

The magic words are "schedule this" or "run this every Friday at 5pm." That drops it into a Schedules function (tucked inside Settings), and from there you can manage, tweak, or delete your schedules as needed. When it runs, you get an email notification and the report also lands in the original chat thread.

*Copilot works slightly differently. After running your prompt, click the three-dot menu and select "Schedule this." One quirk: it only runs 15 times before it stops. If you're running it daily, set a calendar reminder on day 14 to renew it. Mildly annoying. Very manageable.

The one thing that makes or breaks the whole thing.

A vague prompt produces a vague report. (Garbage in, garbage out. The oldest rule in tech, still being violated daily.)

When you write your prompt, be specific about four things:

• Your topic. What you actually care about. Don't specify a broad category like "business news". Instead, make it something you genuinely need or want to read about.
• Your sources. Name the publications, newsletters, or people you trust. Don't leave the AI to guess.
• Your time frame. Use relative language like "in the last seven days" rather than a fixed date. Since the prompt reruns automatically, relative timeframes keep working. Fixed ones don't.
• Your format. Tell it exactly what sections you want. Structure the report you'd actually want to read.

A real example (mine, since you asked)

I've programmed a daily briefing rather than a weekly roundup (because in AI, a week is basically a geological era) within Claude Cowork.

The prompt positions the AI as an intelligence analyst preparing an executive briefing specifically for someone running a corporate AI training firm. So instead of generic news, it filters everything through a specific lens: what does this mean for how people work, learn, and behave?

Each briefing covers five sections:

• The top developments of the last 24–48 hours, with a note on why each one actually matters (not just what happened).
• A behaviour and organisational psychology signals section, which covers research on how AI is changing thinking, creativity, decision-making, and collaboration.
• Implications for corporate AI training, so the insights are actually useful.
• Course design opportunities, which is essentially a standing ideation session for new programs and products.
• And a strategic signals section for the slow-burn trends worth watching before they become obvious.

The result is a daily briefing that arrives already interpreted, filtered for relevance, connected to implications, and ready to inform decisions at Inventium.ai.

(And that's the difference a well-crafted prompt makes. Same AI. Completely different output).

Where to start

Pick one topic that matters to your work. Write a prompt. Ask AI to improve your prompt. Tell it to run every Friday. See what lands in your inbox. Improve the prompt based on the output.

That's it. You don't need to overhaul anything. One scheduled report, one topic, and you've just outsourced a task that was eating your attention every week.

Your inbox might still be a disaster. But at least you'll be across the news.

Cheers

Amantha

Today's newsletter was brought to you by ELMO Software. And they have a gift for you.

If you work in HR, this free five-minute assessment is worth your time. ELMO Software's AI Maturity Assessment benchmarks you across two dimensions: whether you have the right foundations in place, and whether AI is actually delivering results. The assessment was built from research with 1,200+ HR leaders across Australia and New Zealand, so you're benchmarking against people navigating the same landscape as you. You'll get a personalised report with where you sit against ANZ peers, your biggest opportunities, and concrete next steps. Basically: it tells you whether your AI investment is paying off (or going nowhere).

​Take it here.​