Skip to main content
For agencies

Become your clients’ AI department.

You build their websites. Now build the assistants that live on them — add value for your clients; add value to your services.

It’s not a matter of if

Your clients are asking about AI.

Website assistants are becoming mainstream. People will argue about usefulness and cost, but the conversation is no longer about if, it is about when. When will you be adding AI assistants to your clients’ websites?

The opportunity for your agency is to monetise more of what you do well — understanding your clients. The process of adding AI includes architecture decisions, agent definitions, training data, testing, implementation and ongoing maintenance. These consultant-led services are where your agency can generate new revenue streams.

copadre removes the complexity and risk from delivering public-facing website assistants. With a unique combination of defensive security features and conversation management tools, copadre answers all the questions your clients will ask. Instead of maintaining an in-house chat framework across multiple codebases, you can focus on service-led delivery and value.

Creating value

The jobs your clients need you for.

The specification.

Defining a client’s agents is consultancy. You sit with the client, work out what their visitors actually need — a specialist for support, one for taking contact details, a concierge routing between them. Making decisions about how many agents are needed and how to organise clusters is skilled, billable work.

The training.

Capturing and processing training data is a separate task that often requires technical skills and is best suited to an AI-fluent worker. As the agency, you are the natural supplier of these specialised skills.

The easy bit.

Going live is one line of code on a site you already manage, styled with CSS you already control. It slots straight into your build process and your deploy checklist.

The retainer.

Every conversation the agents have is recorded for reporting purposes. Each month you can review what visitors are asking, teach the agents what they were missing, and send the client a report with insights — you can generate reports using your favourite AI.

How clusters work

One copadre, multiple clients

Everything under one roof.

Your copadre account contains a separate workspace for each client — their agents, their conversations, their own API key (theirs or yours). You work across all of them from one login, and each client only sees their own bits.

Assign permissions via one of four roles, including a read-only Viewer role for anyone who just needs to interrogate conversation data and test new agents.

And when you’ve built something you’re proud of, export your work as a single file, rebrand it and import it as the starting point for your next client.

Master the day-to-day

Run every client from your desktop.

Connect Claude — or any AI tool that speaks MCP — to your copadre account, and the day-to-day is done in conversation. A client emails: can the assistant mention this month’s offer? You describe the change, test it in a private preview chat, and publish — live on their website in minutes, without leaving Claude. To your client it looks like an agency that’s always on top of things, because you are.

At month end, let your AI pull a client’s conversations and write the report: see what visitors keep asking, review this month against last, highlight novel enquiries and note where the agents went out of knowledge — and what you taught them as a result. Set it running on a schedule and the report is waiting before the meeting. Insight drawn from their own visitors’ words is something nobody else can sell them.

And everything you build this way is looked after: every definition is encrypted at rest with a full change history, each embed key is bound to a single agent and copadre’s guardrails protect every agent from prompt injection and time-wasters.

Add AI to your rate card.

Free* for six months — no card