The platform has opened its event data to the AI assistants planners already use. It is a bigger move than another copilot, and it shifts the hard questions from one event to the whole portfolio.
I recently caught up with Muhammad Younas, CEO of vFairs, and he ran through what the platform shipped in the first half of 2026. Five items. Most of it is category catch-up, sensible enough. One line, flagged “IMP” near the bottom, is the tell: vFairs has rolled out an MCP server, in beta with around 50 customers, with general availability due in August.
That is why this is worth your attention. Not because chatting to your data is new, but because of what vFairs is trying to build on top of it and what it says about how planners now work. Here is what it means, and where the rest of the list fits around it.
What is an MCP server, and why does it matter for event platforms?
An MCP server lets AI agents query a platform through a standard interface, so an assistant like Claude can pull your registration numbers or check a session’s attendance without anyone logging into a dashboard. MCP, the Model Context Protocol, is an open standard from Anthropic that runs on top of a platform’s existing API rather than replacing it.
Anthropic released MCP in late 2024, and it has since been taken up by Microsoft, Google and Stripe, among others. It is often described as a USB-C port for AI: one connector any model can use to reach any system. For events, the effect is plain. The place a planner goes to ask a question changes from the platform’s dashboard to whatever AI assistant they already have open.
Why did vFairs build one now?
vFairs says the trigger was behaviour. When it spoke to customers this year, a majority were already running their work inside Claude, analysing registration data, building run-of-shows and drafting copy. Rather than pull planners back to a dashboard, the aim was to bring event data to where they already work.
That is a specific and believable read of how planners are changing. If your team already lives in an AI assistant, a login to a separate dashboard is friction. The setup reflects that: add vFairs as a custom connector in your assistant, authenticate with your existing credentials, and access is read-only and scoped to the permissions you already hold. About two minutes.
The more interesting bet is cross-event. vFairs is pitching the server as an event portfolio advisor, not a single-event chatbot. Its retail and aviation customers run hundreds of events a year, and the questions that matter to them live across events, not inside one. Comparing revenue and performance across a programme. Tracking attendees as they move from one event to the next. Comparing session engagement across the whole calendar. That is exactly where a per-event dashboard falls down, and it is a sharper reason for the server to exist than being able to chat to your data.
What can it actually do today?
In beta the server surfaces attendees, registrations, tickets, payments, session and exhibitor details, and portfolio-level stats. Access is read-only. vFairs lists common queries like comparing this year’s registrations to last year’s, breaking registrations down by industry and revenue, and finding who registered but never checked in.
The read-only design is the right call, and it answers the first governance worry. An agent can ask, it cannot act, so nobody’s attendee list gets edited or emailed by a stray prompt. Access is authenticated and scoped to the permissions the user already holds in vFairs, which is the sensible shape for this: live event data out, no ability to change it.
How rare is a first-party MCP server in event tech right now?
Still uncommon. In a July 2026 review of a dozen leading event platforms, ETL found first-party MCP servers are the exception, not the norm. Most platforms reach AI assistants only through third-party bridges, and several have shipped AI copilots, which is a different thing.
This is the point worth holding onto. Plenty of platforms have bolted an AI copilot onto their dashboard over the past year. Bizzabo has its Event OS Copilot and the attendee-facing Bizzy, Swapcard has AI-recommended leads, Cvent has CventIQ. Those keep the intelligence inside the vendor’s own interface. A first-party MCP server does the opposite. It hands the data to whatever agent the customer chooses. Cvent data, for instance, is reachable through third-party connectors like CData, which is not the same as the platform building and standing behind its own server. vFairs building its own puts it among the early movers. One caveat on any rarity claim: a quiet beta is easy to miss, so treat this as the state of what is public, not a certified census.
What is coming next, and what is the catch?
vFairs’ next move is a skill library that turns the questions planners ask every event into one-click actions, so people stop re-typing prompts. The catch is not the read-only server itself. It is that your cross-event attendee data becomes queryable by whatever external model a planner connects, which is a consent and data-protection question for the organiser.
The skill library is a smart, unglamorous idea. vFairs noticed planners re-typing the same high-value questions every event, and some overthinking the prompts, so it is packaging the common ones, who registered but never attended, registrations by industry and revenue, highest-revenue registration days, as ready-made skills. That is the difference between a clever demo and a tool people use on a Tuesday.
Now the part that needs a clear head. Read-only removes the obvious risk. It does not remove the data-governance one. Once your event data can be reached by an external AI assistant, your attendee records are moving to a model you do not control, across your whole portfolio. vFairs says it is being deliberate about how event data leaves the platform, which is the right instinct. The responsibility still sits with the organiser. If you are running this in the EU or UK, whether an attendee consented to their data being queried inside a third-party model is a question for your data protection lead before you connect anything.
The rest of the vFairs H1 list
Alongside the MCP server, vFairs shipped hosted buyer software, RFID and BLE smart badges through its partner VenuIQ, an expanded AI suite, and lead enrichment. None is unique to vFairs, and together they read as a consolidation play: one platform holding registration, matchmaking, onsite tracking and AI support.
Briefly, because these matter less than the MCP move. The hosted buyer software puts vFairs into a crowded matchmaking field alongside Grip, Swapcard, Brella and Jublia. The smart badges track dwell time and booth visits passively through partner VenuIQ, which is useful footfall data and a consent conversation you have to get right, since the badge works whether the attendee is thinking about it or not. The AI suite, matchmaking, an assistant, a writing tool, covers the same ground every major platform now covers, and the vendor’s efficiency figures, up to 70% of repetitive queries deflected and up to 80% off content production time, are vendor claims, unverified, and will vary by event. Lead enrichment turns a badge scan into a fuller profile through a third-party data source, which is handy for exhibitors and another provenance question worth asking.
The ETL agent-readiness questions
Whether it is vFairs or the next platform to ship one, an MCP server changes what you should ask a vendor. Five questions worth keeping in your back pocket. Call them the agent-readiness questions:
- Is the server read-only, or can an agent take actions? If it can act, who authorises it and can you audit what it did?
- Which credentials and permissions does it inherit, and can access be scoped per user and revoked cleanly?
- What data leaves the platform when an agent queries it, to which model, and under what retention terms?
- For passive tracking and enrichment, what did the attendee actually consent to, in plain language?
- What is the switching cost once your agents depend on one platform’s data? Convenience and lock-in are the same feature seen from two angles.
None of these are reasons not to connect. They are the difference between adopting this with your eyes open and adopting the demo.
Where this is heading
Expect first-party MCP servers to move from novelty to buyer criterion within a year. Once planners work through agents, the platform their agent can reach becomes the one they keep. Event platforms will start optimising to be readable and actionable by AI, the way websites optimised to be found by search.
vFairs shipping one is not a trend on its own, but it is the leading edge of one. A 2026 buyer’s guide already lists an MCP server as a forward-looking criterion when choosing a platform. The selection question is shifting from what a dashboard can do to what an agent can do with a platform’s data, and how safely. vFairs is building toward a portfolio advisor rather than a single-event chatbot, and the direction holds across the category: the platform stops being a place you visit and becomes a source your assistant reads. The ones that get the governance right, read-only first, clear on what leaves and where, will be the ones organisers trust to plug in.
FAQ
What is an MCP server in event technology?
An MCP server is a standard interface that lets AI agents query an event platform’s data. MCP, the Model Context Protocol, is an open standard from Anthropic that sits on top of a platform’s existing API. In events it lets an assistant pull registrations, check attendance or compare events without anyone logging into a dashboard.
Is vFairs the only event platform with an MCP server?
No. First-party MCP servers are still uncommon in event tech, and vFairs is an early mover with one, in beta with around 50 customers and general availability due in August. Many platforms currently reach AI assistants only through third-party bridges rather than their own server.
Is the vFairs MCP server safe to connect?
It is read-only and scoped to the permissions the user already holds, which removes the risk of an agent changing or sending data. The open question is governance: your event data becomes queryable by an external AI model, so confirm what leaves the platform, to which model, and whether attendee consent covers it before connecting.
How is an MCP server different from an AI copilot?
A copilot keeps the intelligence inside the vendor’s own dashboard. An MCP server exposes the data to whatever AI assistant the customer already uses, so the work happens where the planner is rather than inside the platform. It is the difference between a feature you visit and a source your agent reads.
Is vFairs a credible platform?
Yes. It was named a Leader in the 2026 Gartner Magic Quadrant for Event Marketing and Management Platforms for the second year running, and holds a 4.7 out of 5 from more than 1,700 reviewers on G2.
