From mass email to conversation: AI in event communication

You have sent the confirmation, the programme and the reminder. Yet on the day of the event someone asks where their ticket is, and another person does not know their transfer was leaving from a different hotel. The information was in the email, but it has not resolved what each person needed.

That is the problem we want to address with EVA: the gap between sending a message and helping someone act on it. An AI assistant can reply in the same conversation, use that person’s context and continue the exchange when a question comes up. It changes both the channel and what happens after the message is sent.

Why sending more information does not fix communication

A general message leaves the attendee to find the part that applies to them. At an event with several ticket types, hotels or transfers, that can mean checking a programme, finding a document and deciding whether an instruction also applies to their situation.

Email also competes with other work messages and campaigns. Another reminder can bring the date closer, but it may repeat information the person already has. The difficulty appears when the organiser can only see that a message has been sent, while the attendee still does not know what to do.

Suppose Elena receives instructions for a conference. The email includes departures for all the coaches; she has a hotel booking and wants to know whether she has time for breakfast before hers. She needs to put two pieces of information together and perhaps contact the organising team. A message she can reply to has a chance to help her at that moment.

Email remains useful for documents or communications you want to keep there. What deserves a review is treating each send as the end of the work and judging its effectiveness only by how many people appear to have opened it.

An open or a click does not always mean someone read your message

Email metrics can include automated activity. Mailchimp documents how antivirus tools, security filters and preview generators can open messages or visit links, inflating opens and clicks. Its filtering options aim to separate that activity from human interaction. Mailchimp’s documentation on bot activity.

A separate mechanism is Apple Mail’s privacy protection, which can preload the pixel used to track opens even when the recipient has not opened the email. Mailchimp explains how this preloading affects its open reports. Apple Mail Privacy Protection in Mailchimp.

A recorded open therefore does not prove Elena has read the instructions for her transfer. A click alone does not show that she has decided to attend either. Before comparing campaigns, check what the platform measures and how it filters automated activity. Even SMS clicks can need filtering, as Mailchimp’s documentation recognises.

These limitations do not prove that real opens have fallen continuously for ten years. That claim would require comparable data and a method that accounts for changes in measurement. For your event, it is more useful to observe what happens after the message.

What changes when attendees can talk to EVA

EVA opens a conversation through WhatsApp, SMS or a chat embedded in your website. Depending on the channel, attendees receive messages or access the chat without installing a dedicated event app. They can ask questions in their own words, without having to guess which section contains the answer.

The language model interprets the question, and EVA uses the information provided to answer. A question such as “Does the coach collect me from my hotel?” needs both the transfer plan and that person’s hotel. If you have connected those details, the reply can specify their departure point. Without them, more information or help from the team is needed.

The tone can be natural and friendly while making clear that an AI assistant is writing. You do not need to pretend there is a person behind every message for the conversation to be useful.

The change is easier to understand by following Elena. Instead of receiving a document covering all transfers and finding hers, she gets an instruction prepared for her group and can resolve the question she still has.

EVA: Hello, Elena. I’m the conference’s AI assistant. Your transfer leaves tomorrow at 8:15 from the entrance to Hotel Norte. You can ask me here about your arrival and the programme.

Elena: Do I need to go to the venue first to collect my badge?

EVA: You can collect your badge when you arrive at the venue after the transfer. Your ticket with its QR code is in this conversation.

This example needs three things prepared: Elena’s transfer allocation, the badge collection procedure and her ticket. The value comes from that connection between information and conversation, as well as easy access to the message.

The channel, the context and the reply contribute different things

Switching from email to WhatsApp and sending everyone the same text leaves part of the problem in place. A familiar channel can make access easier; relevance depends on what you send, and conversation resolves what the message did not anticipate.

What you prepare	What changes for the attendee	What you need ready
Communication channel	They find the message in WhatsApp, SMS or web chat	A configured channel and planned contact with attendees
Segmented message	They receive the instruction for their ticket or group	Reliable registration data and segmentation criteria
Contextual reply	They ask about their hotel or transfer and receive a relevant instruction	Event information linked to that person
Follow-up conversation	They clarify a question while planning their arrival	Prepared answers and a way to check missing information

The organiser’s work changes too. In EVA’s panel you can prepare communications and schedule sends, while the assistant answers questions using the available knowledge. The team remains responsible for keeping event times, access instructions and attendee details up to date. If a transfer changes, review both its information and the scheduled messages.

Measure what attendees achieve after the message

Evaluate the process against a specific action: confirming a transfer, finding a ticket or resolving a question needed to attend. A human reply within the conversation provides a different signal from an open, although it does not guarantee attendance either.

You can review who answered a question, which enquiries remain unresolved and which actions the relevant system recorded. To confirm attendance, use access control data; sending a QR code or someone asking about it does not mean they entered.

If you want to test the effect of a new approach, compare similar events and record what else changed: the date, the audience, the price or the appeal of the programme. Do not attribute the whole difference in attendance to a channel simply because you added it that year.

You could start with a message that currently generates repeated questions. Prepare the recipients’ context, open the conversation and observe whether they can resolve those questions. You can explore EVA’s communication and attendee support capabilities and turn that approach into a communication schedule before, during and after your event.

A useful test ends with answers: what EVA could resolve, what information was missing and what the team still had to do. That gives you a basis for improving the next message beyond looking at its open rate.