Weekly Research Brief
Once a week, what changed with your competitors and in your niche. One page, every line sourced.
What it is
A Claude Project that, on one day a week, looks at the competitors you name and writes one page on what changed — with a source on every line. It replaces the afternoon of tabs that never quite happens.
It looks
Their site, pricing page, news, social accounts, job listings, press. And anything new in your niche a buyer would notice.
It writes one page
Per competitor: what changed, why it matters to you, the source. Then the one thing worth opening yourself. "Nothing new" counts as a finding.
Your week with it
- Monday, 9:00
- The reminder fires. You paste one line into Claude. Two minutes later there's a page.
- Monday, 9:05
- You open the sources for the two most surprising claims. That's the whole quality check, and it's not optional.
- When something matters
- "Go deep on this." Three more sources, what it means for you, one thing to do about it this month.
Good for
Pricing you can't see
They changed a plan, added a tier, ran a discount. You'd have found out from a customer.
Launches and quiet removals
What they added and what they dropped. The second one is usually the more interesting.
Who they're hiring
Job listings are a roadmap nobody bothers to hide.
Where they showed up
A new market, a partnership, an event. Once a week is enough to catch it.
The honest part: it will sometimes be wrong, and it will sometimes say "nothing new" when there is. The two-minute source check is the price of the other 118 minutes it gives back. The tool it's adapted from says exactly this about its own reports.
What do you most want to know?
How often?
Where should the brief land?
Your brief. Do these in order.
What happens.
- Week 1
- The first brief is the most useful one — everything's new to you. Read it end to end. Check two sources. Fix anything wrong in the instructions.
- Week 4
- Most weeks say "nothing new" for someone. That's fine. You're now the person in the room who knew about the pricing change first.
- Month 3
- Run "month in review". The direction shows up. Add a competitor, drop one. Ten minutes a week, and you stopped guessing.
What it won't do
- See what isn't public. Private pricing, private roadmaps, private conversations stay private.
- Be right every time. Sources are how you catch it. Skip the check, and you'll act on a mistake eventually.
- Decide anything. It tells you what moved. What you do about it is still yours.
Prefer to read it? The whole kit in plain text, and what we changed in translation.
The kit, in plain sight
This is the general version. The tool above builds yours, with your competitors in it.
A weekly research brief is a Claude Project that, on one day each week, searches what changed at the competitors you name and in your niche, and writes one page with a source on every line. Fifteen minutes to set up. Two habits keep it honest: a calendar reminder so it actually runs, and opening the sources for the two most surprising claims before you act on them.
# Research Brief — how to run it I sell: [what you sell, to whom] Competitors I track: 1. [name] 2. [name] 3. [name] What I care about: pricing and offers · launches and removals · what they are saying · where they are showing up. When I say "run the brief", do this, every week: 1. Search the web for what changed at each competitor over the last 7 days. Look at their site, pricing page, blog or news, social accounts, job listings, and any press. 2. Search for anything new in my niche that a buyer of what I sell would notice. 3. Write the brief in the format below. One page. Every claim gets a source link. 4. If you are not sure something changed, say "unconfirmed" — never guess. If a competitor shows nothing new, write "nothing new". That is a finding, not a failure. 5. End with "Worth a look": the one thing I should open myself this week, and why. ## Brief format # Brief — [date] ## [Competitor] - What changed: … (source: link) - Why it matters to me: one line ## In the niche - … (source: link) ## Worth a look - One thing, and why. Rules: sources over confidence. Never invent a change. Short beats complete.
Run the brief. Today is Monday. Cover the last 7 days.
Go deep on [competitor]. Take the one change from this week's brief that matters most. Find three more sources on it. Tell me: what they're actually doing, what it means for someone who sells what I sell, and the one thing I could do about it this month. Sources for everything.
Month in review. Look back over the last four briefs. What are the two or three real patterns across my competitors — not this week's noise, the direction. What did I mark "worth a look" and never look at? One page.
What changed in translation
| In the source | Here | Why |
|---|---|---|
| A Python install, model API keys, a search API key, and somewhere to run it | A Claude Project with web search on | Same research loop, nothing to host. |
| A planner agent that writes research questions, then executor agents that answer them | One set of instructions that says what to look for | Your competitors don't change enough in a week to need a planner. |
| Any topic, any depth, on demand | Three named competitors, one cadence | A brief you actually read beats research you could run. |
| Reports with citations | Kept | This is the part that matters. |
| "Always review the final report" | Kept, as the two-minute check | The source's own advice. It's the habit that makes the rest safe. |
Adapted from GPT Researcher (assafelovic/gpt-researcher, Apache-2.0, ~29,000 stars, checked 6 September 2026).