AI Website Automation
MCP is the standard that lets an AI model work directly with your own software. We build MCP servers for WordPress and for systems that have no connector available, so your team can carry out technical work by describing what they need.
Pricing from
Overview
Model Context Protocol is an open standard for connecting AI models to tools and data. It is what allows a model to query your systems, make changes and report back, and it is the foundation of most practical AI automation being built today.
Our Forerunner® WordPress AI Assistant is our own MCP implementation for WordPress. Built on WP-CLI, it lets you manage a single site or an entire network through instruction. Update plugins across a portfolio, publish a post to the client sites in one sector, or audit which installations need attention. Work that once meant logging into thirty dashboards is handled in a single request.
The wider opportunity is everything that is not WordPress. Where you run a proprietary system, a legacy database or an in-house application, no connector exists and none is likely to. We build custom MCP servers for exactly those situations, and the requirements are modest: an API or database access, a way to execute commands, and reasonably structured data underneath.
Permissions are central to this work, so we agree them before development begins. Everything starts read-only, write access is granted per system and per action, and every command is logged. We build these for agencies and in-house teams across the UK, and we run the same approach across our own client portfolio.
MCP is still early enough that adopting it now puts you comfortably ahead of the field, which is worth something while the position lasts.
What’s included
Custom MCP Servers
We build MCP servers for WordPress, CRMs, databases and the in-house systems that no off-the-shelf connector supports, so your own software becomes available to AI.
WordPress at Portfolio Scale
Update plugins across fifty sites, publish to a chosen group or audit versions network-wide. Our Forerunner® WordPress AI Assistant handles it through instruction rather than admin panels.
WP-CLI Foundation
Built on WP-CLI, so the AI produces commands WordPress already understands. That gives you a dependable foundation rather than a layer of automation on top of the interface.
Permissions You Set
Read-only until you decide otherwise, write access granted per system and per action, and a complete log of everything that runs.
Your Team Can Take It On
Handover includes the server code, documentation and a walkthrough, so your own developers can maintain and extend it.
How we deliver
Our AI Website Automation Process
Step 1: Identify the Right System
We identify which systems your team spends most time inside and which of them offer an API, which tells us where a server delivers the greatest return.
Step 2: Agree the Permissions
We agree what the AI may read, what it may change and which actions require approval, all documented before development begins.
Step 3: Build and Test in Staging
The MCP server is built and tested against a staging copy of your system, so everything is proven before it touches live data.
Step 4: Live with Full Logging
It goes live with logging enabled and write access limited to the actions you have approved, then extends as the logs build confidence.
Step 5: Handover to Your Team
We hand over the code, documentation and a walkthrough with your developers, so your team can maintain and extend it independently.
On this page
Client work
What this looks like in practice
We manage our client WordPress portfolio through our own MCP server. It reads and writes posts, pages, custom fields, WooCommerce data and SEO fields across the whole network from natural language instruction, which turns work that would take a full day of dashboard visits into a single conversation. Running it daily across live client sites is what gives us confidence building the same thing for others.
Pricing from
Questions
AI Website Automation FAQs
What does an MCP build cost?
A custom MCP server for one system starts at £2,000, which represents around three days of work at £670 per day. If WordPress management is all you need, the Forerunner® WordPress AI Assistant is £850 as a one-off purchase and you can install it yourself.
What is MCP?
Model Context Protocol is an open standard for connecting AI models to external tools and data. An MCP server is the component that makes your system available to the model safely, acting as a translator with a permissions layer built in.
What does our system need for this to work?
An API or database access, a means of running commands or queries, and data with enough structure for a model to work with. Most modern systems meet this comfortably, and we will confirm yours before quoting.
How do you control what the AI can do?
Everything begins read-only. Write access is granted per system and per action, every command is logged, and on client work we typically keep the more significant operations behind manual approval permanently.
Which businesses benefit most?
It suits businesses managing a substantial number of sites or running teams doing repetitive work inside a system with an API. Where you manage two or three websites, the arithmetic works out differently and we will tell you that.
Can our developers maintain it?
Yes, and we design for that. Handover includes the server code, full documentation and a walkthrough with your developers, so your team can extend it as your needs develop.
Latest news
Claude Fable 5: What Anthropic’s New AI Model Means for Your Business
The Best AI Models so far in 2026
