Event Planning Checklist: Build One Per Event Type With AI

Keep one master task library with each task's lead time counted back from the event date, then ask AI to pick and adjust tasks for each event type. Corporate, private party and conference lists differ in approvals, lead times and suppliers. A date formula turns lead times into due dates, so the list re-dates itself for each booking.

AI drafts a plausible list from a clear prompt. What it cannot know is your suppliers' real lead times, your venue's rules, or which task last year's event proved you forgot. Treat its output as a first draft that your debriefs correct, and never as a compliance checklist: permits, insurance and safety duties depend on the event and the place, and the venue or the relevant authority is the source.

Follow me on Instagram@sagnikteaches

The examples use three illustrative events: a leadership away-day for 80 people, a 40th-birthday party for 60 and a two-day conference for 250 delegates. Lead times are starting points to replace with your own.

Connect on LinkedInSagnik Bhattacharya

Why one master checklist fails across event types

Most of the venue, catering and day-of tasks repeat from one event to the next, which is why a generic checklist feels adequate. The trouble is the rest: the tasks that exist for only one type, and the lead times that differ by months. A conference list copied onto a birthday party buries the host in speaker agreements. A party list copied onto a conference forgets the call for speakers until it is too late.

Subscribe on YouTube@sagnikteaches

Instead of three separate lists to maintain, keep one library and let the event type decide which tasks appear and when. That is a job AI does well, as long as the library underneath it is sound.

Build the task library once

The library is a spreadsheet with one row per task. The columns do the work: an ID so tasks can refer to each other, a lead time in days before the event, a dependency, the event types the task applies to, and a "done when" that names the evidence. Owners are roles, not people, so the library survives staff changes.

In the Types column, C is corporate, P is private party and K is conference. You do not have to type a hundred rows. Have the AI draft a first pass, then spend an hour deleting, merging and correcting it with your last three events beside you.

I run an event planning business that does corporate events, private parties and conferences.
Draft a master task library as CSV with the columns: ID, Task, Phase, Days before event,
Owner role, Depends on (ID), Event types (C, P, K), Done when (the evidence), Flag.
Phases: Brief, Venue, Suppliers, Guests, Programme, Compliance, Final week, Event day, After.
Use days before the event as the lead time, and make the dependencies consistent: no task may
be due before a task it depends on.
In the Flag column write VERIFY for any task involving a permit, licence, insurance, health and
safety or accessibility duty. Do not say what any rule is. Do not invent supplier names.
Aim for about 120 rows and tell me which rows are your guesses.

Two instructions in that prompt earn their keep. "Done when" forces each task to be checkable, so "sort catering" becomes "menu agreed in writing, dietary counts sent". The VERIFY flag keeps the AI from slipping legal claims into the list, and gives you a filter for the tasks that need a real source. Store the finished library and your prompts together: a Project in ChatGPT or Claude can hold the CSV and the instructions so every new event starts from the same reference.

Count back from the event date

Put the event date in one cell and a lead time beside each task, and let a formula produce the due date. Excel's WORKDAY function takes a start date, a number of working days and an optional list of holidays, and Microsoft's page says a negative number gives a past date. Google Sheets lists WORKDAY with the same arguments. To count back in calendar days and roll a weekend date to the previous Friday, use this:

Event date in B1, lead time in days in column D, due date in column F:
F5:  =WORKDAY($B$1-D5+1,-1,Holidays)

Holidays is a named range of your closure dates. Leave out the third argument if you have none.
Test one row that lands on a weekend before trusting the column.

The "+1" then "-1" is the trick: WORKDAY steps back one working day from the day after the target, so a working day returns itself and a Saturday or Sunday returns the Friday. Now try it on the away-day, with the event on Tuesday 9 March 2027.

Two things show up. First, every multiple of seven keeps the event's weekday, so a Tuesday event puts all of these on Tuesdays, which is handy for a weekly review. A Saturday party puts them all on Saturdays, and the formula rolls each to the Friday. Second, the venue contract lands on 29 December, in the end-of-year holiday period, when many offices are closed. No formula fixes that, and a holiday list only skips the dates you give it. Add a "Check" column that you fill in by eye for due dates in closure weeks, and ask the AI to list them.

The prompt that picks and adjusts tasks for one event

With the library saved, each new event needs only a short brief and a strict instruction to use the library, not to write a new list.

Here is my master task library as CSV: [paste or attach].
Event type: corporate away-day. Date: 2027-03-09. Guests: 80.
Venue: hotel meeting room, not yet booked. Budget band: mid.
Special factors: 3 wheelchair users; no sponsors; no overnight stay.
Select every task whose Event types include C. Keep each task's ID and lead time exactly as in
the library unless a special factor gives a reason to change it, and state the reason.
Remove tasks that cannot apply. List any task you add in a separate section marked NEW, with the
reason from the brief. Do not drop any task flagged VERIFY.
Output CSV: ID, Task, Days before, Owner role, Source (LIBRARY, ADJUSTED or NEW), Reason.

An illustrative reply, trimmed to four rows, shows what to look for:

T008,Collect three venue quotes,84,Planner,LIBRARY,
T012,Sign venue contract and pay deposit,28,Planner and budget holder,ADJUSTED,small venue
T030,Request dietary and access counts,28,Planner,LIBRARY,
NEW-1,Confirm step-free route to every room at each shortlisted venue,84,Planner,NEW,3 wheelchair users

The NEW row is the good surprise: the brief mentioned wheelchair users and the AI turned that into a task, which is the kind of adjustment this step is for. The ADJUSTED row is the problem. The AI cut the venue contract's lead time from 70 days to 28 on the strength of "small venue", a reason that is not in the brief. Worse, the invitations are due at 49 days, so they now go out before the venue is signed. Sorting the output by due date exposes that kind of dependency error immediately. Revert the lead time, and add one line to the prompt: "Do not change a lead time unless the brief gives a reason."

Corporate away-day: 80 people on Tuesday 9 March 2027

This is the dated list the library and the prompt above would produce, condensed to milestones. The rows marked with a note are the ones the date check flagged.

What makes this list corporate is mostly about people, not suppliers: a purchase order before a deposit can move, a sponsor who must approve the agenda, and attendees who need travel and calendar information long before they need a menu. Approval delays are a common cause of slipped dates, so the list names the approver on every task that needs one. The tutorial on intake forms shows how to capture that approval chain before the first task starts.

Private party: a 40th birthday for 60 guests on Saturday 26 June 2027

A private list is shorter and more social. Because the event falls on a Saturday and the lead times are multiples of seven, the formula rolls every date back to the Friday before, which suits a weekly Friday check-in with the host.

• Ten to eight weeks out (Fri 16 Apr to Fri 30 Apr). Agree the budget and guest range, name one decision-maker, choose the venue or setting and pay any hold, shortlist caterer and entertainment.
• Seven to five weeks (Fri 7 May to Fri 21 May). Invitations out by 7 May with an RSVP link and deadline. Caterer, music and cake booked by 14 May. Rentals such as a marquee and tables reserved by 21 May.
• Four weeks (Fri 28 May). RSVP deadline. Message or phone anyone who has not answered.
• Three weeks (Fri 4 Jun). Dietary and access counts to the caterer, seating draft, toasts and speeches agreed.
• Two weeks (Fri 11 Jun). Final numbers to caterer and venue, timings confirmed with entertainment, payments scheduled.
• One week (Fri 18 Jun). Final timeline to every supplier, delivery and set-up access confirmed, weather decision rule agreed.
• Event week. Weather call on Wed 23 Jun, set-up on Fri 25 Jun, party on Sat 26 Jun, thank-you messages and supplier payments the week after.

The special-factor line in the prompt does more work here than on the corporate event. Suppose the brief says the party is outdoors with four children among the guests. The AI should add tasks you would otherwise remember late: a weather decision rule with a named decider and a time, power and toilets for the space, a plan for supervising children, and a check on noise limits for the music. Mark the noise and any licensing question VERIFY, and get the answer from the venue or the owner of the space, not from the AI.

Conference: 250 delegates, 12 and 13 October 2027

A conference list is longest and starts earliest, because speakers, sponsors and delegates all have diaries to fit around. Dates below count back from day one.

Six months to 20 weeks out: 13 April to 25 May 2027

The venue contract is signed on Tue 13 April (182 days out). The call for speakers opens on Tue 27 April, the programme skeleton of tracks and time slots is drafted by Tue 11 May, and registration opens on Tue 25 May. Conference-only tasks start here: ticket types and prices, the refund rule, the registration platform and the sponsor prospectus.

18 to 12 weeks out: 8 June to 20 July 2027

Sponsor and exhibitor deals close by Tue 8 June so their deliverables can be scheduled. Speaker confirmations are due on Tue 22 June, the AV and streaming specification on Tue 6 July, and the programme is published on Tue 20 July.

Ten weeks to three weeks out: 3 August to 21 September 2027

Catering estimates go to the venue on Tue 3 August, speaker agreements and headshots come in by Tue 17 August, and the final programme is fixed on Tue 31 August. Signage and printing are ordered by Tue 14 September, and registration closes with final numbers on Tue 21 September.

The last three weeks and the two days

The speaker rehearsal schedule goes out on Tue 28 September, run sheets are final on Tue 5 October, load-in is on Mon 11 October, and the event runs on Tue 12 and Wed 13 October. Debrief and sponsor reporting follow within the week.

The feature unique to conferences is the multiplier. One library row, "Confirm each speaker's AV needs", becomes forty rows when you have forty speakers. If each speaker also carries an agreement, a headshot and bio, and a travel or access check, that is 160 per-speaker tasks, and sponsors and exhibitors multiply the same way. Do not type them. Paste your programme CSV and ask the AI to expand the per-person tasks, one row per speaker per task, keeping the lead times from the library. Then filter by owner so the producer sees only speaker tasks and the sponsor lead sees only deliverables.

When the date is closer than your lead times

Sooner or later a client books the away-day with three weeks' notice. Say the booking lands on Tuesday 16 February 2027 for the event on 9 March: 21 days, against a library that wants 84. The answer is not to squeeze every task, but to sort them: anything with a lead time over 21 days is run in parallel, merged, replaced, dropped or accepted as a risk. Ask the AI to do the sorting and state its reason for each row.

The event is in 21 days. For every task in this list with a lead time over 21 days, choose one:
RUN IN PARALLEL, SHORTEN TO N DAYS, MERGE WITH (ID), REPLACE WITH, DROP, or ACCEPT RISK.
Give a one-line reason. Do not assume any supplier is available. Flag anything where the
supplier's availability, not our effort, is the limit.

The point of the exercise is to make the risks explicit and let the client choose. A rushed event also costs more of your time, so price it that way; the tutorial on event planner fees shows how to build the extra hours into a quote.

Check the list before the event, update it after

A checklist needs two reviews: one before an event relies on it, and one after the event has tested it.

Test the list before an event depends on it

A checklist can look complete and still miss the thing that goes wrong. Before you commit a new event's list, run a pre-mortem: tell the AI the event failed, and ask it to work backwards.

Here is the checklist and the brief for a corporate away-day for 80 people on 2027-03-09.
Assume the event went badly. List the ten most likely reasons, based only on the brief and the
checklist. For each, name the task that should have prevented it, or write "NO TASK COVERS THIS".
Also list any task that is due before a task it depends on.

An illustrative result has four lines worth acting on. The venue releases the hold because the deposit is late: covered by the contract task, but nothing confirms the budget holder will be reachable in the closure week, so add a task. Wheelchair users cannot reach a breakout room: no task checks the route between the main room and the breakouts, so add one. A microphone fails in the second session: the technical check covers it, but nothing provides a spare, so add that. A severe allergy is mishandled on the day: counts are collected, but nothing confirms how the caterer will label dishes, so add a task for that too.

Treat the AI's list as a prompt for your own judgement, not a verdict. Some of its ten reasons will be generic, and some will be wrong for your venue. Accept the ones that match a real risk, add the tasks to the event's list, and decide later whether they belong in the library. Two more checks cost little. Sort the dated list by due date and read it top to bottom, which exposes dependency errors that no prompt catches reliably. And compare it with the last event of the same type: any task on the old list that is missing from the new one should have a reason. If you want a sense of how far to trust the AI on facts such as capacities and rules, the tutorial on where ChatGPT invents event facts is worth ten minutes before you rely on any suggested figure.

Feed what happened back into the library

The library improves only when the debrief changes it. After each event, list every problem and near miss, and for each ask three things. Which task would have prevented it? Does that task exist, and was its lead time right? Does it belong to one event type or to all of them? Then edit the library: add the task, move its lead time, or widen its Types column. Version the file ("library v8, after the March away-day") and keep the previous version, so you can tell later whether a change helped.

Two sibling jobs follow the same loop. The run sheet for the day itself is built from the checklist's final-week tasks, and the budgets and run sheets tutorial covers it. Supplier-side checklists, such as the one in the tutorial on catering events and run sheets, are worth reading to see what your suppliers expect from you and when. For the after-event half, the tutorial on feedback, sponsor reports and debriefs shows how survey comments and debrief notes become the edits this section describes.

Questions about event checklists and AI

Should the checklist live in a spreadsheet or a project tool?

Keep the task library in a spreadsheet, because it is the source of truth and the AI can read it. Then import each event's dated list into a project tool if your team prefers one, provided that tool accepts a CSV import. Check your tool's import rules before you commit; the spreadsheet stays the master either way.

How many tasks should an event checklist have?

Enough that every task has an owner and a date, and few enough that people still read it. If a list for a 60-person party passes about a hundred tasks, it has probably absorbed supplier-level detail that belongs on the run sheet. Split the list by phase and keep the weekly view short.

Can AI tell me which permits and licences I need?

No. Rules depend on the event, the venue and where it takes place, and a model can state them confidently and wrongly. Have the AI add a task that says to verify permits, insurance and safety duties, then get the answer from the venue or the relevant authority and record it as evidence.

Where should the library and prompts be kept?

A shared spreadsheet for the library, plus a Project in ChatGPT or Claude holding your prompt and a CSV copy of the library as reference. As of October 2026, Claude's Free plan allows up to five projects and ChatGPT limits the number of files per project by plan, so one CSV is the safe format.

Further reads

• Which Event Planning Tasks Should You Hand to AI First? — Which planning tasks to hand to AI first, and which to keep.
• A Wedding Planner's AI Workflow From Enquiry to Final Timeline — One full workflow from enquiry to the final timeline.
• DJ Event Checklist: Gear, Timeline and Backup Plan Built With AI — A supplier-side checklist: gear, timeline and backup plan.
• Best AI Tools for Small Event Planning Businesses — Tools that can hold checklists, tasks and client portals.
• AI Software for Event Planning: What HoneyBook AI Covers — What event software covers, and what still needs a prompt.
• Event Venue Contract: Deposit, Cancellation and Damage Terms With AI — The contract terms that your venue deadlines depend on.
• AI for Marketing an Event: A Six-Week Promotion Plan for Venues — A week-by-week promotion calendar for a one-off 120-guest event, with prompts, partner emails, tagged links and a weekly check on ticket sales.
• AI for Meeting Planners: Venue Shortlists and RFP Replies — From client brief to venue shortlist to proposal reply: prompts, a scoring grid and a figure-check table for independent planners and small agencies.
• Bookshop Event Ideas: Author Nights, Book Clubs and Signings With AI — A filled-in event plan, a publisher outreach email, ticket-page copy and a promotion run for author nights, signings and book clubs.
• ChatGPT Meeting Minutes Generator: A Client-Call Recap for Event Planners — A six-stage routine and prompt that turns your client-call notes into decisions, owners, dates and a confirmation email.
• ChatGPT Prompts for Wedding Planners: Vendor Emails, FAQs, Thank-Yous — Nine prompts in three chains: vendor briefs and emails, guest FAQs that stay inside your fact sheet, and thank-you notes that name one real thing.
• Event Feedback Survey: Write It and Summarise the Answers With AI — Ten event survey questions, a frozen theme codebook, counts you verify in the sheet and a one-page report, with the AI doing the reading.
• Event Sponsorship Proposal: Tiers and Pitch Emails Drafted With AI — Price four sponsor tiers from your budget gap, then have AI draft the proposal and a four-email pitch sequence. One worked event, with the sums.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Microsoft Support's WORKDAY function page and Google's Sheets help for WORKDAY (syntax, negative day counts, holidays), plus Anthropic's and OpenAI's plan information on Projects, all checked in October 2026. Dates, lead times and tasks in the examples are illustrative starting points, not rules.