Topic Scout · Ranked ideas
Fresh articles in.
Startable angles out.
Point Topic Scout at the publications your audience reads. It pulls the newest articles, scores every one against the voice and audience of the profile you run it for, and discards anything under a 0.75 relevance floor. What survives arrives as a post angle you can start, not a link you still have to crack open.
7-day full-access trial, $0.95 to start. Cancel anytime.
- Growth team hiring order matters more than budget sizeSearch Engine Journal87%
- Onboarding speed now predicts retention better than anything elseSaaS Mag84%
- Multi-product SaaS outgrowing single-product by 21% in 2026SaaS Mag82%
- CMO and CIO solve different AI agent problems; revenue gets lost between themSearch Engine Journal78%
A real run, replayed. The full cards are below.
Twenty tabs, a bookmarks folder, a note called ideas.txt. By posting day all of it is stale, and the box is still blank.
The scout reads your sources, ranks what is fresh against the voice, and writes the opening angle. You pick one and draft.
The output
One real run, shown unedited.
On Jun 9, 2026 at 22:40 UTC, the scout read one profile's sources, looking back 14 days. Four angles made the cut. These are the four cards exactly as the product rendered them.
- Search Engine Journal - SEO & Digital Marketing NewsJun 9, 202687%
Growth team hiring order matters more than budget size
From: “How To Build A Growth Marketing Team On A Startup Budget”
Direct GTM operator content. Cuts through hype about budgets and focuses on hiring sequence, which aligns with the contrarian, signal-over-playbook voice.
- SaaS MagMay 27, 202684%
Onboarding speed now predicts retention better than anything else
From: “Time-to-Value: Why Onboarding Is the New SaaS Retention Battleground in 2026”
Core to B2B SaaS retention and revenue. The benchmarks here are signal-rich for GTM operators who need to justify onboarding investment to their boards.
- SaaS MagJun 3, 202682%
Multi-product SaaS outgrowing single-product by 21% in 2026
From: “The Compound Startup Era: Why Multi-Product SaaS Is Outperforming Single-Product Peers in 2026”
This data on NRR, attach rates, and the compound startup playbook speaks to founder GTM decisions and how adjacent products compound growth.
- Search Engine Journal - SEO & Digital Marketing NewsJun 9, 202678%
CMO and CIO solve different AI agent problems; revenue gets lost between them
From: “The CMO And CIO Friction Point: Navigating The AI Agent And AEO Ecosystem”
Structural friction in enterprise GTM. Understanding where buyer alignment breaks down is critical as AI agents reshape how demand flows through organizations.
A real run from inside buyWords, Jun 9, 2026. The scores, angles, articles, and dates are the run’s own. Read article opens the published piece.
Three gates
An idea clears three gates, or you never see it.
Relevance, freshness, diversity are not adjectives here. They are constants in the ranking engine, and the run above is the receipt for all three.
Relevance
Every candidate is scored against the profile, and the prompt tells the model the bar exists: “the system filters out anything below 0.75, so don't pad scores.” Whatever lands under the line is discarded before you see it.
Freshness
Stale items are dropped before the model reads anything, so it cannot accidentally rank one. No publish date, no entry. This run looked back 14 days; here is how old its four survivors were when the scout caught them.
Diversity
No publication may own your list. This run sat exactly at the cap, and the rule cuts anything past it.
Relevance to what? This profile.
Before ranking, the model reads the profile: bio, topics, company description, offering, target audience, and voice notes. The same article scores differently for different voices, which is the point.
The reading list
Feeds, typed, validated, watched.
The five roster voices run 9 scout lists and 77 sources between them. This is the Sources tab on Michael Maximoff's real list: 18 feeds across four of the five source types.
- Jason Lemkin's SaaStrPopular blogshttps://www.saastr.com/feed/
- Lincoln Murphy's Sixteen VenturesPopular blogshttps://www.sixteenventures.com/feed
- Todd Caponi's Ethical Selling BlogPopular blogshttps://toddcaponi.com/feed/
- Neil Patel BlogPopular blogshttps://neilpatel.com/blog/feed/
- TechCrunchMedia & newshttps://techcrunch.com/feed/
- HubSpot BlogMedia & newshttps://blog.hubspot.com/marketing/feed
- RocketReach BlogCompetitorshttps://blog.rocketreach.co/feed/
- Google News - B2B SalesTopic mentionshttps://news.google.com/rss/search?q=B2B+sales+appointment+setting
- Google News - Sales EnablementTopic mentionshttps://news.google.com/rss/search?q=sales+enablement+marketing+automation
- + 9 more popular blogs
Add feeds one at a time, or let the suggestion engine fill the list: 13 of these 18 landed in one batch on May 14, 2026. The two Google News rows watch for topic mentions across the news.
Five source types
The suggestion engine hunts differently per type. This is the registry as it ships, labels and descriptions included.
- Media & newsJournalistic outlets and trade publications that cover the profile's topics.
- Popular blogsHigh-signal expert blogs and Substacks — individual voices, not newsrooms.
- CompetitorsBlogs and content feeds from companies in the same space as this profile.
- Topic mentionsbest effortFeeds that surface every mention of the profile's topics (alerts/aggregators). RSS isn't native to this type yet. Expect fewer hits.
- Own resourcesbest effortThe profile's own published content — their company blog or personal feed. RSS isn't native to this type yet. Expect fewer hits.
Two of them are flagged best effort in the product itself, not in the fine print.
Feeds rot. The scout notices.
A real dead feed, with the real error that killed it, caught May 14, 2026. Dead feeds get flagged in the list and skipped at run time, never silently counted.
Source discovery
AI-suggested feeds, fetch-tested before they save.
Ask for ten sources and the model proposes candidates. None of them touches your list until a live request proves the feed exists.
- STEP 1
Propose
The model names candidate publications for every open slot, with everything already in your list excluded so it cannot repeat itself.
ASKS FOR 3× MORE THAN IT NEEDS - STEP 2
Fetch
Every candidate gets fetched live and has to actually answer. A made-up address dies right here, and never reaches your list.
5 SECONDS TO ANSWER - STEP 3
Parse
Must be a working feed carrying at least one article. If the suggested address fails, nearby paths on the same site get tried before the candidate is dropped.
7 BACKUP PATHS TRIED - STEP 4
Save
Survivors insert after every attempt, so validated feeds appear in the list while the search is still running.
SAVED AS THEY PASS
“AI ran out of fresh ideas. Try again to keep searching.”
Eight attempts, a 240-second budget, then it stops and says so instead of padding your list with junk. A tool that flags its own limits is a tool you can trust on the rest.
These are the product’s real rules and its real words, not a brochure version of them.
Idea to draft
A score is not a post. One click starts one.
Every card carries its own exits, and the idea travels with its receipt: the angle, the article, the score.
Onboarding speed now predicts retention better than anything else
SaaS Mag · May 27, 2026A dead idea costs one click, goes nowhere, and stops cluttering the list.
Opens the Writing Assistant seeded with this idea. The draft comes back in the active profile's Voice DNA, not a house style, and the card remembers the conversation once a draft exists.
SEEDED WITH THE ANGLE AND THE ARTICLEOr skip the clicking entirely: set a list to weekly and the scout greets you with a fresh run. 3 of the 5 roster voices keep weekly auto-scout switched on.
THE FLOOR, ONE LAST TIME
Stop hunting for ideas. Start choosing between them.
Point the scout at ten publications and the next blank page shows up already read, already ranked, with the angle written.
7-day full-access trial, $0.95 to start. Cancel anytime.