Good morning.

Last week, a founder I advise spent $4,200 on a freelance market researcher.

The typical deliverable was a 22-page competitive landscape report. Took two weeks. Half the data was already stale by the time he opened the PDF.

That same afternoon, I showed him how to set up Claude Cowork, point it at the same market, and had a more comprehensive analysis (with pricing breakdowns, positioning maps, and messaging teardowns) saved to his laptop in under 40 minutes.

This was not a ChatGPT-style wall of text. He got actual files. An Excel workbook with competitor feature matrices. A slide deck ready for his board meeting. A markdown brief with sourced citations he could hand to his marketing team (and for further discussion with Claude or ChatGPT).

Any AI can summarize a competitor's website. The difference was execution. Cowork doesn't advise. It does the work.

And when you configure it specifically for market research (for the kind of intelligence that drives your marketing, sharpens your sales, and validates your offers) it becomes something most operators and teams never had:

A dedicated research employee who works on your schedule, in your files, for $20 a month.

Here's how to set it up, what to ask for, and the workflows that are replacing five-figure research retainers.

— Sam

Let’s get into it.

If you and your team has been using Claude, ChatGPT, or Gemini for research, they've been working with one hand tied behind their back. Not because the AI is bad but because the chat interface limits what's possible.

Regular AI chat works like this: 

Ask a question, get an answer, copy it somewhere, ask another question, paste it into a different doc. The person is the glue holding the whole thing together: project manager, file organizer, and output formatter all at once.

For a quick question, that's fine. For actual market research, it falls apart fast.

Claude Cowork launched in January 2026 as a tab inside the Claude Desktop app. Anthropic built it because developers were already using Claude Code (their terminal-based coding tool) for non-coding tasks. They realized the same agentic power could be made accessible to everyone.

Here's what makes it fundamentally different for research:

This is a research capability you and your entire team can deploy.

The setup is simple enough that your marketing lead or ops person can have this running before lunch. You need a Claude Pro account ($20/month) and the desktop app.

What follows is the full setup. If you're the one configuring it, walk through it step by step. If you're handing this to someone on your team, send them this issue. Everything they need is here.

Go to claude.com/download and install the Claude Desktop app (available for macOS and Windows). Open it up and you'll see three tabs at the top: Chat, Cowork, and Code. Click Cowork.

This is critical, especially if multiple people on your team will be using this with proprietary data. Before anyone feeds Claude any competitor intel, customer lists, pricing strategies, or internal sales numbers, make sure conversations aren't being used to train Anthropic's models.

Go to File → Settings → Privacy and toggle off any options related to using your data for model improvement.

Do not skip this. If your team is researching competitors, analyzing customer behavior, or testing pricing, that's sensitive business intelligence. Protect it before anyone prompts. This applies to every AI tool, not just Claude, but it's especially important here because Cowork has access to actual files.

If you're rolling this out to your team: make this step part of the onboarding checklist. Every seat needs this configured independently.

This is the step most people skip, and it's the most important one for getting research-quality output.

Click File → Settings → Cowork → Edit next to "Global instructions." This is where you tell Claude who you are, what kind of business you run, and how you want research delivered.

Here's a template you can customize (and share with your team so everyone's instance is calibrated the same way):

On your computer (or a shared drive your team can access), create a folder called something like "Market Research, Cowork". Inside, create subfolders:

If your team shares a Google Drive or Dropbox, this folder structure works there too. Anyone running Cowork can point it at the same shared context.

Drop in everything you already have. Customer interviews. Sales call transcripts. Google Analytics exports. Product reviews. Old market research. The more context Cowork has, the sharper its analysis.

Create a simple aboutme.md file in the root of your folder. This is a text file (saved with the .md extension) that Cowork reads and uses as persistent context for everything it does.

Here's a market-research-specific version:

Open Cowork, point it at your research folder, and give it a task. Start simple:

If the output doesn't match your expectations (wrong format, wrong depth, too generic), go back and tighten your global instructions. This calibration step is how you train the system before handing it to the rest of your team.

As you use Cowork, it automatically builds a memory of your preferences and context. To verify it's learning, go to File → Settings → Capabilities and check what it's retained.

This matters because the second research task will be sharper than the first. The tenth will feel like working with someone who's been on your team for months. The memory compounds. It learns your market, your competitors, your language preferences, and how you like deliverables structured.

From here, you scale up and scale out across your team.

This is where Cowork earns its keep. Manual competitor research is slow, tedious, and outdated the moment you finish. Cowork makes it continuous.

With the Claude in Chrome extension enabled, Cowork can browse competitor websites, read their pages, extract data from review sites, and compile it, all without anyone opening a browser tab.

Hand this prompt to your marketing lead. They run it once, share the Excel workbook with the team, and now everyone is working from the same competitive intelligence.

This slide goes straight into your next team meeting or board deck.

This is gold for you and your marketing team:

Your copywriter or marketing lead uses this to sharpen positioning without starting from scratch.

Cowork gives you what most audience research leaves out: the actual language your market uses.

Drop your customer reviews, testimonials, support tickets, and survey responses into the /audience folder. Then:

That spreadsheet becomes you and your team's copy bank. Every headline, every ad, every email can pull directly from the language your market actually uses. Share it with whoever writes your copy.

The executive summary goes to your sales team. The full document goes to marketing. Everyone works from the same ICP.

If you save social media posts, Reddit threads, or forum discussions into your folder, Cowork will mine them:

This is where market research meets revenue. Before your team spends weeks building a course, a service tier, or a product, let Cowork pressure-test the offer.

That deck becomes the starting point for your next launch planning meeting.

Your sales team (even if it's two people) can use Cowork to show up sharper on every call.

Run this before every sales call. Share the brief with whoever is on the call.

Print this. Pin it next to your sales team's monitors. Update it quarterly.

If you save post-sale notes, lost deal notes, or CRM exports:

After testing dozens of Cowork prompts for market research, here's what I've learned about getting the best output.

Bad prompt: "Do some market research on my competitors."

Good prompt: "Research [specific competitors]. Analyze [specific dimensions]. Format as [specific deliverable]. Save to [specific location]."

Cowork performs best when you tell it what to research, what to look for, how to format it, and where to put it. Share this principle with your team. It'll save them hours of iteration.

Don't ask for "an analysis." Ask for "an Excel workbook with three tabs" or "a Word document with an executive summary and detailed sections" or "a PowerPoint deck with one insight per slide."

The deliverable type forces Cowork to structure its thinking differently, and gives you something you can actually hand to someone or present in a meeting.

For large research projects, tell Cowork to use sub-agents:

Sub-agents run simultaneously, cutting research time dramatically. Useful when you need intelligence on multiple fronts before a planning session.

One of Cowork's newest features is scheduled tasks. Set up recurring research that runs automatically:

Your market intelligence stays fresh without anyone managing it. You and your team opens the /output folder Monday morning and the latest competitive update is already there.

The setup above gets one person running. Here's how to make it a team capability.

The compounding effect matters here. Week one, you have scattered research. By month three, Cowork has deep context on your market, your competitors, and your customers, and every new prompt builds on everything that came before.

That kind of institutional knowledge used to take a dedicated analyst years to build.

You don't need to overhaul your entire research process. Pick one workflow and deploy it.

By Friday, you and your team will have more actionable market intelligence than most companies generate in a quarter.

And you'll have a research capability that keeps compounding without adding headcount.

Claude Cowork handles the grunt work between the question and the answer.

The hours of tab-switching, copy-pasting, and reformatting used to separate "I wonder what our competitors charge" from "here's a pricing strategy backed by market data."

You and your team's judgment still drives every decision. Cowork just closes the gap between curiosity and intelligence.

The operators who deploy this across their team first will make better decisions at every level of the business: marketing, sales, product, pricing.

And when you and your team has continuous intelligence instead of quarterly reports, the quality of those decisions compounds.

Set it up this week. By next month, you'll wonder how your team operated without it.

Until next time,
Sam Woods
The Editor

.