Building a competitive intel workflow with Claude Code

How I built a Claude Code competitive intelligence workflow that pulls pricing, positioning, and feature data from competitor sites and turns it into usable analysis.

· Page Sands

Every GTM team needs competitive intelligence. Most do it the same way: someone spends a few hours every quarter clicking through competitor websites, copying pricing pages into a spreadsheet, and writing a summary that’s outdated within weeks.

I wanted something faster. So I built a Claude Code competitive intelligence workflow that pulls competitor data, structures it, and produces analysis I can actually use for positioning and sales conversations.

What competitive intel actually needs to cover

Before building anything, I listed what I actually reference when making GTM decisions:

  • Pricing and packaging. What tiers exist, what’s included, how they gate features.
  • Positioning language. How competitors describe themselves on their homepage, product pages, and meta descriptions.
  • Feature sets. What capabilities they highlight and in what order.
  • Recent changes. New features, pricing shifts, messaging pivots.
  • Content themes. What they’re publishing about, which signals where they think the market is going.

Most competitive intel tools automate collection and spit out dashboards. The output is usually raw data without interpretation. A list of competitor blog titles doesn’t tell you anything until someone reads them and connects the dots.

Claude Code can pull the data and do that first pass of analysis in the same session.

The workflow: three steps

Step 1: Define your competitors in a structured file

I created a drafts/competitors.md file with a simple format:

## Competitors

### [Competitor A]
- URL: https://competitor-a.com
- Category: Direct competitor
- Key pages: /pricing, /product, /about

### [Competitor B]
- URL: https://competitor-b.com
- Category: Adjacent (overlapping use case)
- Key pages: /pricing, /features, /customers

This file serves as the input. Claude Code reads it at the start of every competitive intel session and knows which companies to analyze and where to look.

Five to eight competitors is the right range. More than that and the analysis gets shallow.

Step 2: Pull and structure the data

I run a Claude Code session that reads the competitor file, visits the key pages using web fetch, and extracts structured data into a consistent format. The prompt is straightforward: read each competitor’s pages, then fill in a template covering pricing tiers, positioning headline, key features, and any recent changes visible on the site.

The output goes into drafts/competitive-analysis.md as a structured comparison:

## [Competitor A] - Last updated: Mar 05 2026

### Pricing
- Starter: $X/mo (includes Y, Z)
- Pro: $X/mo (adds A, B, C)
- Enterprise: Custom

### Positioning
Homepage headline: "[Their headline]"
Product page lead: "[Their product description]"

### Feature emphasis (ordered by prominence)
1. Feature X
2. Feature Y
3. Feature Z

### Recent signals
- Blog post about [topic] published [date]
- New feature announced: [feature]

When every competitor’s data follows the same structure, patterns show up fast. Who’s leading with price versus product. Who’s pushing upmarket. Who just shipped something new.

Step 3: Generate the analysis

Structured data is useful on its own. But I also want the “so what.” I ask Claude Code to read the full competitive analysis file and answer specific questions:

  • Where are we differentiated that competitors don’t cover?
  • Which competitors changed their pricing or positioning in the last 30 days?
  • What themes are multiple competitors writing about?
  • Where is there a gap that nobody is addressing?

The answers go into a drafts/competitive-brief.md that I can share with the team or reference when writing content and updating positioning.

Why this beats a spreadsheet

Speed. The full cycle from pulling data to a usable brief takes about 20 minutes. The same coverage takes a half day manually.

Consistency. The structured template means every competitor gets the same treatment. No skipping sections because you ran out of energy on competitor number six.

Synthesis. Claude Code reads the full dataset and spots patterns across competitors. “Three of five competitors launched a free tier in the last 60 days” is the kind of observation that takes time to surface by hand.

What this doesn’t do

It doesn’t track changes over time automatically. Each run is a snapshot. I keep previous versions of the analysis file so I can diff them, but there’s no built-in change detection yet.

It also doesn’t cover everything. Win/loss data, customer reviews, sales call intel: those come from other sources. This workflow covers the public-facing competitive picture. The stuff anyone can see but few teams bother to systematically track.

And it requires judgment. Claude Code’s analysis is a starting point. Sometimes it overweights a positioning change that’s cosmetic. Sometimes it misses context you’d only know from being in the market. I treat the output as a first draft, not the final word.

How this connects to the rest of the content operation

The competitive brief feeds directly into content decisions. When I see a gap that competitors aren’t covering, that’s a potential post topic. When I see competitors converging on a message, that tells me where I need to differentiate.

Same pattern as how I used the X API and Claude Code to build a content plan. Social signals show what the market is talking about. Competitive signals show what the market is doing. Both inform what to write and how to position it.

Getting started with Claude Code competitive intelligence

If you want to build something like this, start small. Pick three competitors. Create the competitor file with their URLs and key pages. Run one session where Claude Code pulls and structures the data. See if the output is useful before adding complexity.

The workflow scales from there. Add more competitors. Refine the questions in the analysis step. Save briefs over time so you can track how the competitive picture shifts quarter to quarter.

Twenty minutes with Claude Code and a structured workflow gets you competitive context that most teams only update once a quarter.