The 6am Clips, Rebuilt

For as long as campaigns have had media advisers, someone has been getting up in the dark to read the papers. The morning clips serve a real purpose — the day's first decisions depend on knowing what was said overnight — but the way most teams produce them has barely changed in twenty years: one person, three hours, forty browser tabs, and a long email that everyone skims and nobody quite trusts to be complete.

The grind in that ritual is automatable. The judgment in it is not. Rebuilding the clips well means being precise about which is which, and refusing to let the machine near the second part.

Draw the boundary of what you watch

Before any automation, the watch list. This is an editorial decision, not a technical one: it defines what your organisation considers coverage, and everything downstream inherits its blind spots. Write it down, argue about it, and revisit it each quarter. For most Australian campaigns and advocacy organisations, the list looks something like this:

• 01The metro dailies in every market you operate in
• 02The national broadcaster, online and on air
• 03Talkback radio, where a political day is often decided before 9am
• 04Community-language press and radio in the languages your electorate actually speaks
• 05Hansard, for what was said about you in the chamber
• 06Council minutes, for what was decided about you closer to the ground
• 07The specialist press for your issue, which writes today what the metros will write next week

Let the machine carry the buckets

Once the list exists, intake is plumbing. An automation platform can pull feeds, alerts and transcripts into one place on a schedule, overnight, without complaint. The point is not sophistication — most of this is fetching and filing — but reliability. A pipeline that runs at 4am every day, including the days when the person who built it is on leave, is worth more than a clever one that needs tending.

The first genuinely intelligent step is collapsing the pile. Twenty items about the same announcement are not twenty stories; they are one story wearing twenty mastheads. An assistant is good at recognising that the wire copy, the metro rewrite and the radio segment all describe a single event, and grouping them under one line with the variants listed beneath. The digest shrinks from eighty items to perhaps a dozen stories, and suddenly it is possible to actually read it.

Sort by what it demands of you

A clustered digest still needs a hierarchy, and the useful one is not topic but obligation. Three tiers cover almost everything: stories that need a response today, stories worth knowing before the first meeting, and background that feeds the longer arc. The assistant can propose the tier; a person confirms it, because the difference between a shrug and a crisis is frequently invisible in the text and obvious to anyone who knows the campaign.

Then hold the format still. Same sections, same order, same headers, every single morning. Consistency is what makes the report scannable — a campaign director should be able to absorb the day in four minutes because their eyes already know where everything lives. A report that reinvents itself daily costs each reader a small toll of reorientation, and those tolls compound across a team.

Ask the machines what they say about you

One newer habit belongs in the weekly rhythm rather than the daily one. A growing share of voters no longer search for your candidate or your organisation — they ask an assistant, and take the answer at face value. So once a week, put the questions to the major assistants yourself: who is this candidate, what does this organisation stand for, what is their record. Note what comes back, and where it is wrong or stale, treat it the way you would treat an error in a newspaper — correct the public record the models draw on, through your website, your published material and the coverage you generate.

The judgment is the product

None of this removes the editor. It restores them. The hours once spent copying links were never the job; the job was always the thin layer of judgment at the end — this one matters, that one can wait, this one needs the candidate's eyes before 8am. Automation returns those hours, and the digest becomes trustworthy precisely because a person the team knows has read it and signed it. The machine assembles the morning. The editor still decides what it means.