TL;DR
We set up an AI assistant that knows our entire website and growth strategy, can spot where content has drifted from our positioning, and applies changes across every affected page in one go. What used to be a multi-hour manual review now happens in a single conversation.
Challenge
Our website has dozens of pages: work entries, service descriptions, solution pages, and insights. Each one needs to reflect our current positioning, tone, and strategic priorities. But strategy evolves faster than content. A service gets repositioned, a new capability gets added, the way we describe our value shifts based on what we learn from client conversations.
Keeping everything aligned meant opening pages one by one, comparing them against our growth strategy document, rewriting what needed updating, and committing changes file by file. For a site this size, a single strategic shift could mean touching twenty pages. Most of the time, the updates just did not happen because the effort was not worth it for each small change. Content drifted, and we lived with the inconsistency.
Solution
The setup has two parts that work together.
Part 1: Full website context
We attached our entire website content directory and growth strategy document as persistent knowledge for the AI assistant. Every conversation starts with the AI already knowing every page on the site, our positioning, tone guidelines, service descriptions, and strategic priorities.
This alone is useful. The AI can answer questions like "how do we currently describe our AI solutions service across all pages?" and give an accurate, cross-referenced answer.
Part 2: The ability to act on the codebase
We connected the AI to our website's code repository through an official integration. This gives it the ability to create a branch, update files, and open a pull request, the standard review-and-approve workflow developers use for any website change.
Combined, these two capabilities mean the AI can:
- Review any page against our growth strategy and tone guidelines
- Identify content that has drifted from our current positioning
- Recommend specific copy changes across multiple pages
- Apply all the changes in one pull request for us to review and approve
- Reference our strategic priorities when explaining why a change matters
How it works in practice
A typical session looks like this: I ask the AI to review how we describe a particular service across the site. Because it already has every page loaded, it can cross-reference the service page, related work entries, insights that mention the service, and our growth strategy's positioning guidelines, all in one pass.
When it finds content that could be sharper, more consistent, or better aligned with how we want to position ourselves, it drafts the changes. Then it creates a branch, commits the updated files, and opens a pull request. I review the changes, leave comments or approve, and merge.
What used to be a multi-hour manual process across dozens of files becomes a single conversation.
Setup
The technical setup takes about fifteen minutes. It uses <a href="https://github.com/github/github-mcp-server" target="_blank" rel="noopener noreferrer">GitHub's official MCP server</a> running locally through Docker, connected to Claude Desktop. The <a href="https://github.com/github/github-mcp-server/blob/main/docs/installation-guides/install-claude.md#claude-desktop" target="_blank" rel="noopener noreferrer">installation guide</a> walks through every step.
For the website knowledge side, you add your content files through Claude's Project settings. No code or configuration needed. The AI reads the files and carries that context into every conversation in the project.
What this means
Most teams treat website content reviews as periodic projects: someone opens a spreadsheet, audits pages one by one, and updates them over days or weeks. By the time the audit is done, new drift has already started.
This pattern turns strategic content alignment from an occasional project into something you can do in fifteen minutes. The AI has the full picture of your website and your strategy. It knows what should change, it can tell you why, and it can apply the changes for you to review. The result is a website that actually reflects your current thinking instead of last quarter's thinking.
Want AI that works directly with your content and strategy?
We build workflows where AI operates inside the systems your team already uses, not alongside them.