Beacon Blog
What Your Selection Committee Should Know About MCP Before Talking to AMS Vendors

What Your Selection Committee Should Know About MCP Before Talking to AMS Vendors
If your association is choosing a new AMS in the next year, someone on a vendor call will probably mention MCP.
It stands for Model Context Protocol, an open standard that lets AI tools connect to systems like your AMS in a consistent way.
Asking whether a platform supports it gives your selection committee a quick read on one question: whether your association can add new AI tools later without paying someone to build a fresh integration each time.
What MCP is, in plain words
Your AMS already talks to other software through an API, which is a set of rules that lets one program ask another for data. Your event platform might use the API to check whether a registrant is a current member, and your email tool might use it to pull a renewal list. Each of those connections is an integration, and someone had to build it and has to keep it working.
AI tools need the same kind of access to be useful. An AI assistant that can read your member records can answer a staff question like “which lapsed members attended last year’s conference?” in a few seconds. One that cannot see your data can only give general advice. Before MCP, each AI product needed its own custom connector to each system, so connecting three AI tools to your AMS meant three separate projects.
MCP gives those connections a shared format. A vendor builds one MCP server for its AMS, which is a small piece of software that describes what an AI tool may see and do in that system. Any AI tool that supports MCP can then connect to the server without new custom work. People often compare it to a USB port: a device built to the standard plugs into any computer built to the same standard.
Because MCP is an open standard, no single company controls who can use it. It was first released in late 2024, and a large number of AI products and software vendors have adopted it since.
Why your selection committee should ask about it
An AMS contract usually runs several years, and your association will probably keep the system longer than the first contract term. The AI tools your staff use today will not be the ones they use in three years.
If your AMS supports MCP, swapping one AI tool for another is closer to a settings change than a development project. If it does not, each new tool brings the same questions back: who builds the connector, what it costs, and how long your staff wait for it. If your association has no technology team, that cost often decides whether staff get to use a new tool at all.
MCP support also tells you something about the vendor. A platform that has built an MCP server, or has a dated plan for one, has thought about how your staff will work with AI. A vendor who has not heard the term is giving you useful information as well.
What a “yes” on MCP does not tell you
MCP support is a starting point for the conversation. It does not mean the vendor’s AI features are good, and it does not mean the connection is set up the way your association needs it.
Start with scope. An MCP server only exposes what the vendor chose to include. One vendor might let AI tools read member profiles and nothing else, while another also lets them update records or create invoices. Read-only access is the safer place to begin, and write access saves more staff time, so you want to know which one you are getting.
Then ask about permissions. Your staff have different access levels in your AMS today, and the AI connection should respect them. If a part-time coordinator cannot see payment history in the AMS, an AI tool working on that coordinator’s behalf should not see it either.
Cost and maturity come next. Some vendors include MCP access in the subscription, and some price it as an add-on or count it against API usage limits. Some have a working server that customers use every day, and some have a line on a roadmap. Either answer is fine to hear, as long as you hear it before you sign.
Finally, ask whether the AMS keeps a record of what an AI tool viewed or changed. When a record is updated and nobody on staff remembers doing it, that log is how your team finds out what happened.
A quick example from the staff side
Say your association has 2,500 members and a staff of six. Your membership director wants to ask an AI assistant every Monday which members registered for the spring conference and also have a renewal coming due in the next 60 days, so she can send each of them a personal note.
That one request needs data from your AMS and your event platform, and the notes go out through your email tool.
If all three systems support MCP, the assistant can reach each one through the same standard, and your membership director gets her list without anyone exporting spreadsheets. If the event platform and email tool support it and the AMS does not, the AMS becomes the gap, and the workaround is the same Monday spreadsheet she builds by hand today.
This is why we suggest your committee list every system your staff would want AI to reach before vendor meetings begin. The AMS usually sits at the center of that list, so its answer carries the most weight.
Questions to bring to your next vendor meeting
These are written so any committee member can ask them as is.
Do you offer an MCP server for your AMS today? If not, is one on your roadmap, and for which quarter?
Which records and actions does it expose, and can AI tools only read data or also create and update records?
Does the AI connection follow the same user roles and permissions our staff have in the AMS?
Is MCP access included in our subscription, or is it priced separately or limited by API usage?
Which AI tools have your customers connected to it so far?
Does the AMS keep a log of what an AI tool viewed or changed?
Record each vendor’s answers in the same format so your committee can compare them side by side after the demos. If an answer to the first question is vague, ask for a date in writing.
Where MCP fits in your selection
MCP belongs on your requirements list as one line among many. It ranks below how well a platform handles your membership model, your events, and your finances, and below the quality of the vendor’s implementation support, because a well-connected system that your staff struggle to use day to day will not pay off. What MCP adds is a simple way to check whether the platform you pick today will still work with the AI tools your staff want to use later.