The short answer
You can run fan engagement from an AI agent by connecting it to your platform account over MCP, the Model Context Protocol, an open standard for connecting agents to software. FANDEMiQ runs its own MCP server, so an agent such as Claude can set up events, build fan walls and FanMoments and read your stats under your login and your permissions, with anything fan-facing waiting for your approval by default.
What MCP changes
Without MCP, an AI assistant can talk about your fan platform but cannot touch it. You copy things in, copy answers out, and paste screenshots. With MCP, the agent works with your account directly, in the same terms a colleague would.
That is a practical difference. "What did fans say at Saturday's match?" becomes a question you ask in plain English and get an answer to, instead of a report you go looking for.
What an agent can do
The agent is not a chatbot bolted on to a corner of the product. It reaches the whole Manager. It can create events from your templates, make them live, change them mid-event and close them down. It can set the questions and campaigns that run at each event, build and preview FanMoments, lay out and publish fan walls, and send fans their content by email.
It can also read a brand from a website, draw images in your colours, write wording in your tone, and report on uploads, answers and insights. For anything without a dedicated tool, it can search the Manager's full API and call what it finds, under the same rules as you.
Three ways in
The first is to ask an agent in a chat. Connect Claude through claude.ai, the desktop app or Claude Code, or another agent that supports MCP, and talk to it. Anything fan-facing comes back to you as an approval link.
The second is to let it run hands-free on a schedule. With hands-free switched on for an account, an agent can create Saturday's event from a template, take it live, send fans their content and close it down afterwards. The third is to plug in your own systems, such as a nightly sync or a CRM, using a server key in place of a person signing in.
The guardrails that matter
The agent signs in as you, with two-factor, and only to the accounts you choose. It can never do more than your own login can, and you pick the scope: read, set things up, publish and send, or administer. Anything that reaches fans or costs money waits for a person, with a summary written by FANDEMiQ and not by the AI.
Hands-free is opt-in. It is off on every account until FANDEMiQ switches it on at your request, and an agent cannot switch it on for itself. AI work comes out of a monthly AI budget that only FANDEMiQ can raise, every action is recorded, and fans' email addresses and phone numbers stay hidden unless you grant the publish permission. You can disconnect an app or revoke a key and it stops at once.
An honest caution
With hands-free on, nobody sees a fan email or a live change before it happens. An agent that reads something misleading, such as a web page, a PDF or a fan's own answer, could act on it. So use hands-free for routine, templated work, and keep approvals on for everything else.
Think of it as the difference between delegating the Saturday setup and delegating the apology email. One of those you hand over.
What a good brief looks like
Agents do best with a clear, bounded job. "Set up a campaign for this weekend" is a fine start, and the agent will ask for what it needs, such as dates and times, the sponsor and where the images come from, before it creates anything.
Dates and times are local to your venue, so 7pm means 7pm where the event happens. Give the agent a template to start from where you can, because templated, repeatable work is exactly what suits an agent best.
How to start
If you already use Claude, add a custom connector and enter the FANDEMiQ address. Sign in with your usual login, tick the accounts and permissions, and ask your first question the way you would ask a colleague. If you would rather see it first, ask for a demo and we will show it running on a real account.
Quick answers
What is MCP?
The Model Context Protocol, an open standard for connecting AI agents to software. It lets an agent work with your account directly, instead of through screenshots or copy and paste.
Can an AI agent send fan emails on its own?
Not by default. Anything that reaches fans waits for your approval. Only on an account where FANDEMiQ has switched on hands-free, at your request, can an agent act without approval.
Which AI agents work with FANDEMiQ?
Claude, in claude.ai, the desktop app and Claude Code, plus other agents that support MCP with sign-in, such as Codex. Your own systems can use a server key instead of a sign-in.
Keep reading
- How to run fan engagement without a team pressing buttonsFan engagement usually stalls on staff time, not ideas. Here is how to run it for every game without anyone pressing a button, and what keeps it safe.
- Fan engagement platforms compared: what to ask before you buyChoosing a fan engagement platform? Use this checklist on data ownership, staff time, approvals, branding and exit, and ask every vendor the same questions.
