ChatGPT for event planning: prompts, workflows and AI tools
ChatGPT, Claude and Microsoft Copilot can take a lot of writing and structuring off an event team’s plate. This guide shows what they do well, gives you prompts you can copy, and explains the difference between pasting AI output into your tools and connecting an assistant to your event software.
ChatGPT for event planning works best as a drafting partner: it turns rough notes into an event brief, an agenda, registration form fields, landing page copy, invitations and a reminder schedule in minutes. Give it the goal, audience, date, format and constraints, ask for a fixed structure, and check every fact before you use the draft. The guest list, sending and check-in stay in your event management software. If you want the assistant to build the event page and registration form directly, connect it through an MCP server, such as the AirLST MCP server, instead of copying and pasting.
What ChatGPT can help event planners do
AI for event planners is mostly about time: the first version of every document, email and page takes longer than the edits after it. A chat assistant is good at that first version, and at turning messy notes into a clean structure. It is not a source of facts about your venue, your budget or your guests.
| Task | Good fit for AI | Stays with your team |
|---|---|---|
| Event brief and concept | structure, goals, open questions, risks | budget, priorities, sign-off |
| Agenda and run of show | session flow, timings, breaks, speaker briefings | who speaks, final timings |
| Event page and registration form | page copy, sections, form fields and conditions | what you ask guests, approval before publishing |
| Invitations and reminders | subject lines, body copy, tone variants | guest list, sending, timing |
| Attendee communications | a schedule from save the date to follow-up | segments by guest status, final send |
| After the event | survey questions, recap email, report outline | numbers from your event data |
Event planning prompts you can copy
Every prompt below follows the same pattern: a role, the facts of your event, the constraints and an exact output format. Replace the brackets, paste the prompt into ChatGPT, Claude or Copilot and keep the chat open to refine the answer. The prompts work in any assistant; none of them needs a connection to your event software.
01Give it a roleFor example “experienced corporate event planner” or “B2B event marketer”. The answers get more specific.
02Give it the factsGoal, audience, date, place, format and the decisions already made. Anything you leave out, it will guess.
- 03Name the constraints
Budget range, brand voice, approvals, accessibility, languages. Say what it must not do.
04Fix the outputA table, a numbered list or a set of headings. Ask it to mark assumptions and list open questions.
You are an experienced corporate event planner. Draft the agenda for a one-day [event type, e.g. customer summit] on [date] in [city or venue].
Audience: [who, roughly how many]. Goal: [what the event must achieve]. Fixed points: [keynote, lunch, evening program, end time].
Output a table with start time, end time, session, format (keynote, panel, breakout, networking), owner and notes. Plan realistic buffers and room changes. Mark every assumption and list open questions at the end.
Work back from [event date] and create a planning timeline for [event type] with [number] guests.
Group the tasks by month and week: venue and hotel, speakers, event page and registration, invitations, reminders, catering, badges and check-in, follow-up.
For each task give an owner role and a deadline. Flag tasks that depend on another task. Do not invent prices or vendor names.
Review the text below as a critical guest who is short on time. List what is unclear, what is missing (date, time, place, dress code, deadline, contact) and what sounds like marketing. Then give me a revised version that keeps the facts exactly as they are.
[paste text]
Creating event briefs with ChatGPT
The event brief is where an AI assistant saves the most time, because it forces the questions everyone else will ask later. Paste your notes from the kickoff call, even if they are incomplete, and ask for a fixed structure. The value is in the open questions at the end: they become the agenda for your next stakeholder meeting.
You are an experienced corporate event planner. Turn my notes into a one-page event brief with these headings: goal and success criteria, audience, format, date and location, agenda outline, guest journey from invitation to follow-up, budget lines, risks, decisions still open.
Event: [name and type]. Goal: [e.g. qualified meetings with key accounts]. Audience: [who, internal or external, roughly how many]. Date and place: [date, city or venue, in person or hybrid]. Constraints: [budget range, approvals, brand rules].
Mark everything I did not give you as an assumption. End with the five questions I should answer first.
My notes: [paste]
Building event pages and registration forms
There are two ways to use AI for an event page. In the copy-paste workflow, ChatGPT writes the copy and proposes the form fields, and you build the page yourself in your event website builder. In the connected workflow, the assistant builds the page and the registration form directly in the event software, and you review and approve.
Propose the registration form for [event]. Guests are [audience]. We need the answers for: [catering, hotel, sessions, travel, badges].
For each field give: label, field type (text, single choice, multiple choice, date, checkbox), required or optional, why we need it, and which answers should show or hide other fields.
Keep the form as short as possible. Flag every field that collects personal data we may not need, and suggest the consent text we need to check with our privacy team.
Using the event brief below, write the copy for the event landing page: headline, one-sentence subline, three reasons to attend, agenda overview, speakers section placeholder, practical information (date, time, place, how to get there) and a short FAQ with four questions.
Tone: [brand voice]. Length: scannable on a phone. Do not add facts that are not in the brief.
[paste brief]
In AirLST, create a landing page for the event [name] using our brand template. Use the approved brief below for the copy.
Build the registration form from the existing guest fields: [fields]. Show the hotel question only if the guest selects “I need a room”. Ask me before you create any new guest field.
Render the page on mobile and desktop and tell me what you would improve. Do not publish.
[paste brief]
With the AirLST MCP server, the assistant works in Airstage, the AirLST builder for landing pages and registration forms. It can add and change page elements, create subpages with SEO settings, build the registration form from your guest fields with conditions and steps, apply a design template, check the result on mobile and desktop, and undo steps. Publishing creates a numbered version that you can switch back from right away. Details on the MCP server page.
Writing event invitations with AI
Invitations are short, so the details matter: subject line, the first sentence and a clear call to action. Ask for several variants and choose, rather than asking for one perfect version. Use placeholders for names and personal links and map them to the merge fields of your event software; do not paste your guest list into the chat.
Write an invitation email for [event] on [date] in [place] for [audience]. The guest should register by [deadline].
Give me five subject lines (under 50 characters), one preview text and one email body under 150 words. Start with what the guest gets, not with us. One call to action: “Register now”. Use [first name] and [registration link] as placeholders.
Tone: [formal, warm, executive]. Then write a second version for guests who already attended last year.
Sending stays in your event software: in AirLST, invitation emails go out in your design with personal registration links, and every reply updates the guest list. The MCP tools do not send email, so an assistant cannot send an invitation by mistake.
Planning attendee communications
The communication plan is the second big time saver. Ask the assistant for a schedule by guest status, not one email list for everyone: guests who have not replied need a reminder, registered guests need practical information, declined guests may get the recording.
Create an attendee communication plan for [event] from save the date to follow-up. Event date: [date]. Registration deadline: [date].
Output a table with: send date, audience by guest status (invited, no reply, registered, declined, attended, no-show), channel (email, SMS, push), purpose and subject line.
Keep it to the messages that are really needed. Then draft the reminder for guests who have not replied and the practical-information email for registered guests.
Draft a post-event survey for [event] with at most eight questions: one overall rating, questions on content, organization and venue, one open question. Then write the thank-you email that links to the survey, and a version for no-shows with the recording.
In the event software, the segments come from the guest list itself, so the plan turns into reminders and workflows that fire on guest status. The draft comes from the assistant, the timing and the audience from your data.
Using AI with event management software
A chat assistant does not know your event. Your event management software does: the guest list, registration answers, status and check-ins. How well the two work together decides how much time AI really saves.
| Copy-paste workflow | Connected through MCP | |
|---|---|---|
| How it works | ChatGPT drafts, you copy the text into your tools | the assistant works inside the event software through an MCP server |
| What you need | any assistant, any software | an assistant with MCP support and software with an MCP server |
| Event data | only what you paste into the chat | what the signed-in user may see, no export |
| Result | text and lists you build with yourself | drafts of pages and forms directly in the builder |
| Control | you check every paste | user permissions, every step logged, you approve and publish |
Start with the copy-paste workflow: it needs no setup and shows quickly where AI helps your team. Connect the assistant once the same kind of work repeats, for example a page and form for every stop of a roadshow. More on the topic in AI in events.
Connecting ChatGPT, Claude and Microsoft Copilot to AirLST
AirLST has no AI generator of its own: you bring the assistant your company already uses, and AirLST provides the tools and the controls. The standard connection is the AirLST MCP server in Airstage, for landing pages and registration forms. The rollout to the other modules of the suite follows module by module; order and timing are still open.
01Copy the addressIn the Airstage builder, open the dialog “Connect your agent” and copy the server address.
02Add it to your assistantAdd the address as a connector or MCP server. The dialog has guides for each assistant; in companies an admin may need to enable it.
03Sign in with your AirLST userAuthentication runs through OAuth, without a client key. The assistant gets exactly your permissions.
04Describe what you needAsk for a page or form in plain language, watch the draft in the builder and approve before publishing.
| Assistant | How it connects | Status |
|---|---|---|
| Claude | Claude Desktop, Claude.ai (depending on plan; Team and Enterprise through the admin) and Claude Code, with an in-product guide | Standard |
| ChatGPT | developer mode on paid plans; connecting your own MCP servers is a beta feature in ChatGPT | Standard (beta in ChatGPT) |
| GitHub Copilot, Cursor, Codex CLI | GitHub Copilot in VS Code, Cursor and Codex CLI with in-product guides; any other client with MCP over HTTP and OAuth | Standard |
| Microsoft Copilot | Copilot Studio and Microsoft 365, set up for your company as part of the system configuration | Individual integration |
AI workflows for enterprise events
In large companies, the event is one step in a longer process: a campaign in the CRM, an invitation list from sales, attendance back to the account team. AI fits into that process when the data flows are in place first. Data moves through connectors and the REST API; the assistant works on top through MCP.
| System | How it connects | Status |
|---|---|---|
| Salesforce | connector: campaign members become guests, registration status goes back to the campaign member | Standard |
| HubSpot | connector for contacts and event data | Standard |
| Your own systems | AirLST REST API | Standard |
| SAP, Microsoft Dynamics, Microsoft 365, Power BI | REST API, connectors and MCP, set up for your company | Individual integration |
| Microsoft Copilot | Copilot Studio and Microsoft 365, see above | Individual integration |
- 01Campaign to event page
Marketing sets up the campaign in Salesforce; its members become guests in AirLST. The assistant drafts the event page and registration form in Airstage, the team approves.
02Invitations from the event softwareInvitations go out from AirLST with personal links. Registrations update the guest list and go back to Salesforce as campaign member status.
- 03Recurring formats
For roadshows, seminars or dealer events, the assistant creates one draft per date from the same brief and template, instead of copying and editing by hand.
04Reporting in your own toolsAttendance and status in Power BI or Microsoft 365 is an individual integration, set up with your IT.
Which systems connect, and how, is part of the setup for your company. An overview of the standard connectors is on the integrations page.
Data privacy, approvals and limitations
- Placeholders instead of guest data: draft in the chat with [first name] and [company]; the real guest list stays in the event software. Check your company’s AI policy and the data terms of your assistant’s plan before you paste anything internal.
- Permissions in a connected setup: with the AirLST MCP server, the assistant works only on event pages the signed-in user may see, and every tool call is stored. It cannot send email, export data or read out the guest list through these tools.
- Approvals: a person decides what is published. In Airstage, publishing creates a numbered version you can switch back from right away. Pages in the drag-and-drop module editor are read-only for the assistant.
- Where data lives: guest and event data stay with AirLST on servers in Germany, GDPR-compliant. Which language model you use, and what it processes, is governed by your contract with its provider.
- Facts: assistants make up venues, prices, speaker titles and legal wording with great confidence. Check every fact, and have terms, consent texts and accessibility information approved by the people responsible.
- Scope today: the AirLST MCP server works in Airstage (landing pages and registration forms). Everything beyond that runs through the standard connectors or an individual integration.
Your AI. Your systems. Your events.
Connect ChatGPT or Claude to AirLST and your assistant builds landing pages and registration forms in Airstage, with your permissions and every step logged. Salesforce and HubSpot connect as standard; Microsoft Copilot and your other systems as an individual integration.
- AirLST MCP server in Airstage, sign-in through OAuth
- Assistant works with the user’s permissions, every step logged
- No email sending or export through the assistant
- Data hosted in Germany
ChatGPT for event planning FAQ
Can ChatGPT plan an event?
It can plan on paper: draft the brief, agenda, timeline, form fields, page copy and emails. It cannot book a venue, check real availability or manage your guest list on its own. Treat its output as a first draft that a planner checks and decides on.
What are the best AI tools for event planning?
For writing and structuring, a general-purpose assistant such as ChatGPT, Claude or Microsoft Copilot is enough. For the event itself you still need event management software for the guest list, registration, invitations, check-in and reporting. The most useful setup connects the two, so the assistant can work inside the event software with your permissions.
What is a good ChatGPT prompt for event planning?
A good prompt gives the role, the event facts (goal, audience, date, place, format), the constraints (budget, brand, approvals) and the exact output structure you want. Ask the assistant to mark assumptions and to list open questions at the end. The prompts in this guide follow that pattern.
Can ChatGPT connect to event management software?
Yes, through the Model Context Protocol (MCP), if the software offers an MCP server. In ChatGPT, connecting your own MCP servers is currently a beta feature of paid plans and runs through developer mode. With AirLST, ChatGPT and Claude connect to Airstage, where they build landing pages and registration forms.
Does AirLST work with Microsoft Copilot?
Yes, as an individual integration. Microsoft Copilot through Copilot Studio and Microsoft 365 is set up for your company as part of the system configuration, by AirLST Integration Services or your IT. It is not a one-click integration. GitHub Copilot in VS Code connects to the AirLST MCP server with an in-product guide.
Is it safe to paste attendee data into ChatGPT?
Avoid it. Names, emails and dietary or accessibility needs are personal data, and what happens to them depends on your plan and your company’s AI policy. Draft with placeholders instead, and keep the real guest list in your event software. When an assistant works in AirLST through MCP, it works with the signed-in user’s permissions and cannot export the guest list.