Skip to main content
BRIEF · AI briefs

Ask the wire a question, get a sourced answer

Ask in plain language. It searches the live wire and the archive, opens the key releases in full, shows you each step as it works, and writes an answer in which every claim links to its release.

Free plan · Plus CA$29 or Pro CA$99/mo, cancel anytime
ppn.world/app/feed?tab=brief
What it is

The brief — PWAI in the terminal — is a chat window wired straight into the release store. You ask a question — what moved in Canadian mining this week, biotech approvals in the last 24 hours — and it plans the search, reads the releases that match, opens the most relevant ones at their source, and can search again mid-answer, on the live wire or in the archive, before it writes a word. You watch each step as it happens, and its reasoning while it thinks. The answer is built from those releases and nothing else, and every factual sentence ends with a numbered link to the release it came from.

It exists because the two alternatives both fail. Reading the wire yourself costs a morning and you still miss the item that ran at 04:00; a general chatbot writes a confident paragraph about news it never read. Here the retrieval runs first, the model is instructed never to add a release, a number or an event that is not in what it found, and when nothing matches the reply says so rather than padding.

It is for the person who has to be current by a deadline: an editor opening a shift, a comms or investor-relations team asked what a sector did overnight, an analyst writing the Monday note. If the alternative is skimming two hundred headlines to keep six, this is the shortcut that still shows its work.

What it does
LIVE

It shows its work

The reply starts at once: what it is looking for, how many releases matched, which ones it opened at the source, each extra search, and its reasoning while it thinks. When it starts writing, all of it folds into one line you can open again.

SOURCES

Every line traces back

Every factual sentence ends with a numbered marker that opens its release; hover it to see the publisher, country, date and headline first. The releases the answer cited are listed underneath, and one button copies the answer with its sources.

RESEARCH

Deep thinking researches

Turn on Deep thinking and it works like a research desk: several rounds of searches across the live wire and the archive, the decisive releases read in full, then a report with a summary, a dated timeline and a table of figures. It takes a few minutes and counts as three questions.

FORMAT

Tables, timelines, next questions

Comparisons come back as tables and dated events as a timeline. Every answer ends with three follow-up questions to pick from; regenerate an answer in one click, and stop one mid-way without losing what it wrote.

THREAD

Follow-ups keep the thread

Conversations sit in a sidebar and the last eight turns replay, so “and in Germany?” works, and a rewrite — “in French”, “shorter” — reuses the releases the last answer used. Attach any release, or paste its link, to make it the subject. The answer comes back in your interface language.

YOU

It knows who is asking

Tell it once who you are and how you want answers — “business reporter in Montréal, lead with the figures” — and every answer follows it. Ask from any screen with ⌘K, or dictate the question in Chrome, Edge or Safari.

How it works
01 · Open PPNWORLD AI

Type PWAI in the command bar, or type your question in ⌘K and pick “Ask PPNWORLD AI”.

02 · Ask in your words

Name the sector, the country and the window — “German energy deals this week” is already a complete question. For a researched report, turn on Deep thinking.

03 · Check the sources

Click the numbered markers, or open the releases listed under the answer, and read them yourself before you publish anything.

What it does not do

It answers only from what the wire and its archive carry: a subject no release mentions gets a plain “nothing matched”, not a guess. It opens releases at their source where the publisher's page allows it — otherwise it works from the headline and excerpt — and it paraphrases and links rather than reproducing a licensed text. On Plus it answers 10 questions a day, a Deep thinking answer counting as three; Pro is unlimited within fair use; the Free plan does not include it.

Frequently asked
Does it read the whole archive?+
When the question needs it. It answers from the live wire first and reaches into the archive when the question goes further back or the live wire has too little; with Deep thinking it searches the archive on purpose, for an issuer's history and earlier announcements.
Can it invent a release?+
The model only sees the releases found for that turn, up front or by searching along the way, and is instructed never to add one — or a number, or an event — that is not in them, and to say plainly when the set is thin. Every answer is labelled AI-generated and each claim links to its release, so a wrong line is checkable in one click.
Can I get it as an email?+
Yes, on Pro. Ask for it in words — “send me the news from Canada every day at 9” — or press “Every morning” under an answer for 7:00: PPNWORLD AI answers that question again each day at that hour, your time, and emails it with its sources and a link to stop, up to three questions per account. Separately, the daily digest of your saved searches goes out at 11:00 UTC.
The rest of the terminal
FEED
Live feed
GLOBE
World map
CALENDAR
World Agenda
WORKSPACE
Cockpit
WATCH
Watchlist
ALERTS
Alerts
Read next

One seat, the whole wire.

Every surface on this page is in the same seat, at CA$99 a month.