Steal This Idea: AI Seating Arrangement App

Seating people at a wedding sounds simple until you actually have to do it.

Then you realize the problem isn’t:

Put 120 people into 15 tables.

The problem is:

These six people want to sit together.

Those two absolutely cannot sit together.

This couple needs to be near the front.

Grandma can’t walk far.

The kids should sit near their parents.

These three people know nobody.

Half this table speaks Portuguese and the other half doesn’t.

Table 8 only fits seven people.

Suddenly you’ve got a constraint-satisfaction problem disguised as wedding planning.

And somebody is trying to solve it by dragging names around in Canva.

That’s the idea.

The idea

Build an app where someone uploads their guest list, describes the relationships and constraints, and gets several suggested seating arrangements.

For each guest, you might know things like:

• Partner/spouse
• Family
• Friendship groups
• People they know
• People they don’t know
• People they should sit beside
• People they should not sit beside
• Age
• Children
• Language
• Accessibility requirements
• Wedding party / family importance
• Dietary requirements
• Table preferences

The user also defines the room:

• Number of tables
• Seats per table
• Table shape
• Head table
• Distance from exits
• Distance from dance floor
• Distance from speakers
• Accessibility
• Children/family areas

Then the system produces a seating plan.

Not just one.

Maybe three:

Keep existing groups together

Encourage guests to mingle

Minimize conflicts

And explain the tradeoffs.

The result being sold isn’t:

AI generates a seating chart.

It’s:

Stop spending three evenings arguing over where 120 people should sit.

Why I think there may be a problem here

This is one of those tasks that is objectively small compared with the entire wedding.

But it can consume a ridiculous amount of mental energy.

The couple probably knows things the wedding planner doesn’t.

The parents have opinions.

There are social politics.

The guest list keeps changing.

Someone cancels.

Someone adds a plus-one.

Suddenly one table has ten people and another has six.

You rearrange everything.

And unlike choosing napkin colors, there can be real consequences to getting it wrong.

Put the divorced parents beside each other and you’re going to hear about it.

Put the one person who knows nobody at the edge of a table full of childhood friends and they may have a terrible night.

So there is at least a plausible pain.

The question is whether it’s painful enough that people will pay to make it go away.

That’s what I’d test.

Don’t start by building an optimization engine

This is exactly the kind of idea I’d be tempted to over-engineer.

You could model every guest as a node in a graph.

Relationships get weights.

Conflicts get negative weights.

Tables have capacity constraints.

Then you build some beautiful optimization algorithm.

Maybe use simulated annealing.

Maybe integer programming.

Maybe an LLM converts natural-language descriptions into constraints.

That all sounds fun.

And none of it proves there is a business.

The first version should probably be manual.

Someone sends you:

• their guest spreadsheet;
• a photo or description of the venue layout;
• the table capacities;
• their seating rules.

Then you create three arrangements for them.

You could use:

• ChatGPT;
• a spreadsheet;
• some simple code;
• your own judgement.

The customer doesn’t care how clever the algorithm is.

They care whether the result saves them work.

Who I’d test first

Weddings are the obvious market.

I’d probably start there because:

• the problem is easy to understand;
• couples already spend money solving wedding problems;
• there is strong search intent;
• there are lots of online communities;
• people have a fixed deadline;
• the emotional cost of getting it wrong can be high.

But I wouldn’t assume weddings are the best long-term market.

The same problem appears in:

• corporate dinners;
• conferences;
• award ceremonies;
• fundraising galas;
• university events;
• formal dinners;
• cruise ships;
• networking events;
• large birthday parties.

Business events might eventually be more attractive because they happen repeatedly.

A wedding customer may use the product exactly once.

An event planner might use it 50 times per year.

But I’d still start with weddings because it is easier to understand the problem.

The first offer

I would create a simple landing page.

Something like:

We’ll create your wedding seating plan for you.

Send us your guest list, table layout and any awkward family politics.

We’ll send you three optimized seating arrangements within 24 hours.

Then charge for it.

Maybe:

€49 for up to 80 guests

€79 for up to 150 guests

€129 for larger weddings

I wouldn’t spend too much time deciding the exact price.

The initial goal is simply to find out:

Will anybody pay at all?

You can change the price later.

What I’d want to learn

I’d want to understand where the actual difficulty lies.

Is it:

• keeping friendship groups together?
• divorced families?
• plus-ones?
• language groups?
• kids?
• fitting everyone into table capacities?
• constant last-minute changes?
• getting parents to agree?
• figuring out who knows whom?
• designing the visual seating chart?

Those are very different problems.

Maybe the optimization isn’t actually the hard part.

Maybe the hardest part is collecting the relationship information.

The bride knows that Sarah knows Emma from university.

The groom has no idea.

Neither wants to manually classify 150 relationships.

That might lead to a completely different product.

For example:

Send every guest a link.

Ask them:

Who on this guest list do you know?

Now the system builds the social graph automatically.

That could be far more valuable than the seating algorithm itself.

But you won’t discover that by sitting at home designing features.

You discover it by helping people do the task.

How I’d test demand

I’d find people who are currently planning weddings.

Not people who got married last year.

Not people who might get married someday.

People who currently have this problem.

I’d look in:

• wedding Facebook groups;
• Reddit wedding communities;
• local wedding groups;
• wedding-planning forums;
• Instagram;
• TikTok;
• bridal communities.

I’d ask:

Has anyone here reached the seating-plan stage yet? I’m testing a service where you send me your guest list and constraints and I generate a few seating arrangements for you.

Even better, offer the first few free in exchange for letting you observe how they currently do it.

Then start charging quickly.

The free tests teach you the process.

The paid tests tell you whether there’s a business.

Search could be a very strong distribution channel

This problem has something I like:

People know what to search for.

They search things like:

• wedding seating chart
• wedding seating planner
• how to arrange wedding tables
• where should divorced parents sit at wedding
• wedding table arrangement
• wedding seating chart template
• how to seat wedding guests
• wedding seating plan generator

That means you can create free tools around the problem.

For example:

Free wedding table capacity calculator

Input:

• 120 guests
• 10 round tables
• 10 people per table
• head table of 8

Get a basic layout.

Then:

Want us to arrange the actual guests for you?

Or:

Free random wedding seating generator

Useful enough to attract visitors.

Not good enough to solve the difficult version.

Upgrade to the paid product when you need constraints.

The free utility becomes distribution.

Pinterest could be unusually good

Wedding planning is extremely visual.

People already use Pinterest for:

• seating charts;
• table layouts;
• decorations;
• wedding planning;
• templates.

So you could create content like:

8 wedding seating arrangements for 100 guests

How to seat divorced parents at a wedding

Round tables vs rectangular tables

Wedding seating mistakes to avoid

Each image leads to the tool.

That’s a much more natural acquisition channel than buying random Meta ads.

Wedding planners could be even better

The obvious problem with B2C weddings is that each customer gets married once.

Hopefully.

So customer lifetime value is limited.

A wedding planner is different.

If they organize 30 weddings per year and the product saves them an hour on each one, they have recurring value.

I’d definitely test planners after the initial consumer test.

The pitch becomes:

Send us the guest list and we’ll give you three viable arrangements instantly.

Or eventually:

Upload your guest list, enter the constraints and regenerate the seating plan whenever someone cancels.

That last part could be extremely useful.

Wedding seating plans change.

A planner doesn’t want to manually rearrange 140 people because one family of four cancelled.

Software becomes much more valuable when it handles change.

Venues could be another distribution channel

Wedding venues already know:

• the room layout;
• table sizes;
• capacity;
• common configurations.

Imagine the venue gives every couple a link:

Plan your seating at Hotel X.

The venue layout is already configured.

The couple uploads the guest list.

That could create built-in distribution.

Instead of acquiring every wedding customer yourself, you acquire the venue once.

Then the venue sends you customers continuously.

You could potentially give venues the software for free and monetize the couples.

Or charge venues as SaaS.

Again, don’t decide upfront.

Test both.

What could make this difficult

The biggest issue is that seating arrangements are subjective.

There may be no objectively correct answer.

You could generate something mathematically optimal and the customer immediately says:

Absolutely not. Steven cannot sit with Mark.

Because of something that happened in Ibiza in 2017.

This means the product needs to be very easy to correct.

The experience should probably be:

1. Generate arrangement.
2. User says what they don’t like.
3. Lock certain guests/tables.
4. Regenerate everything else.

That iterative process may matter more than producing the perfect answer on the first attempt.

The interface might matter a lot

I normally wouldn’t obsess over UI during validation.

But eventually this product probably needs a good visual interface.

People should be able to drag guests around.

Lock tables.

See warnings.

Add constraints.

Maybe click on someone and say:

Must sit with:

Sarah
John

Prefer near:

Maria

Cannot sit with:

David

Then regenerate.

The combination of visual editing and optimization is probably where the final product gets interesting.

But again:

Don’t start there.

Could AI actually help?

Yes, especially with messy human input.

Someone could type:

My parents are divorced. Mum is remarried to Steve. Dad is coming alone. They are civil but I’d rather have them on separate tables. My sister should be near Mum. My grandparents need to be close to the entrance. The Portuguese side of the family mostly doesn’t speak English.

The system turns that into structured constraints.

That is genuinely useful.

AI could also explain decisions:

I placed your grandparents at Table 2 because it is closest to the entrance and away from the speakers.

Your parents were placed on separate tables while both remain near the head table.

Portuguese-speaking guests were distributed across three tables while keeping immediate families together.

That makes the result feel intentional rather than random.

The business model

I’d initially charge per wedding.

Something like:

Free

Basic seating planner.

€29

AI-generated seating plan.

€79

Advanced constraints + unlimited regeneration.

€149

Human-reviewed seating plan.

Maybe those prices are wrong.

That’s why I’d start manually.

For wedding planners, perhaps:

€49/month

or

€20 per event

For venues:

Maybe white-label pricing.

There are lots of possible models.

Don’t optimize them before proving demand.

The strongest version might not be a wedding app

This is something I’d keep in mind.

Suppose wedding consumers like it but don’t pay much.

Then you discover corporate event organizers desperately need it.

Their constraints might include:

• put potential customers beside salespeople;
• distribute employees from different offices;
• keep competitors separate;
• seat sponsors prominently;
• maximize networking;
• mix departments;
• ensure VIP placement.

Now you’re solving a much more valuable optimization problem.

The same core engine could become:

Optimize who should sit next to whom at business events.

That’s potentially a much larger B2B product.

But I wouldn’t start there unless customer conversations point that way.

Start with the easiest problem to test.

The Next 7 Days

If I wanted to find out whether this idea deserves any more of my time, this is exactly what I’d do.

Day 1: Define the hypothesis

Write down:

People currently planning weddings find seating arrangements frustrating enough to pay someone €50+ to help them produce one.

Then list the assumptions underneath it.

For example:

• Seating arrangements take several hours.
• Couples find the process stressful.
• The problem happens often enough.
• People trust an external tool/service with the decision.
• People will provide their guest data.
• The result can be meaningfully better than a spreadsheet.
• €50+ is cheap compared with the time/stress saved.

Don’t research the market for six hours.

The goal is to identify what needs proving.

Day 2: Find 20 people with the problem

Go where people are actively planning weddings.

Search:

• Reddit
• Facebook wedding groups
• local bride groups
• wedding forums
• Instagram

Find 20 people who mention:

• seating plans;
• guest lists;
• table arrangements;
• family conflicts;
• wedding planning stress.

Message them.

Don’t pitch immediately.

Say something like:

I’m researching how couples handle wedding seating plans. Are you at that stage yet? I’d love to ask you five questions about how you’re doing it.

Get conversations.

Day 3: Talk to at least five

Ask about behaviour, not opinions.

Good questions:

• How many guests do you have?
• Have you started the seating plan?
• How are you doing it?
• How long have you spent on it so far?
• What’s the hardest part?
• Who else is involved in deciding?
• Have you had to redo it?
• What happens when someone cancels?
• What are the awkward constraints?
• Have you tried any software?
• Have you paid for anything that helps with this?

Then ask them to show you what they’re currently using.

Spreadsheet?

Piece of paper?

Canva?

Wedding planning software?

Seeing the process is much more valuable than hearing someone say it is annoying.

Day 4: Create the offer

Based on what you learned, create a one-page landing page.

Do not build the product.

Offer something like:

Wedding seating plan done for you

Send us your guest list, table sizes and all the awkward constraints.

We’ll send you three seating arrangements within 24 hours.

€49.

Include a payment button.

That’s enough.

Now send it to the people you interviewed.

Say:

I decided to test this properly. I’m doing the first few manually for €49. Want me to do yours?

Now you learn whether their problem survives contact with a price.

Day 5: Deliver one manually

Your goal for Day 5 is to get one real seating arrangement completed.

Use whatever tools make it easiest.

Spreadsheet.

LLM.

Python script.

Paper.

Doesn’t matter.

While doing it, document every step.

Write down:

• what information was missing;
• what took the most time;
• what decisions were difficult;
• what could easily be automated;
• what required human judgement.

Then send the customer the result.

Ask them to critique it aggressively.

Day 6: Test distribution

Now try three small distribution experiments.

Experiment 1: Wedding community

Make a useful post:

I helped someone optimize a 120-person wedding seating plan. Here are the five constraints that caused most of the problems.

Don’t spam the product.

Put the offer at the bottom.

Experiment 2: Search

Create one useful page around something people already search for:

How to Make a Wedding Seating Plan Without Losing Your Mind

Include the service.

Experiment 3: Wedding planners

Email 20 wedding planners.

Something like:

Quick question: how much time do you normally spend helping couples rearrange seating charts? I’m testing a service that takes a guest list + constraints and generates several viable arrangements.

See which channel gets responses.

Day 7: Look at the evidence

Don’t ask:

Do I still like this idea?

Ask:

What happened?

Count:

• People contacted
• Conversations
• People who said the problem was painful
• People who already had a solution
• Landing-page visits
• People who asked about the service
• People who paid
• Price objections
• Time required to fulfill
• Quality of the result
• Distribution channels that produced interest

Then decide.

If you spoke to 10 people and nobody really cared:

Kill it.

If people cared but nobody paid:

Change the offer, price or market.

If someone paid and immediately asked if they could change the guest list and regenerate it:

That’s very useful information.

If three wedding planners say:

I’d use this every week.

Forget the consumer product and investigate the planner market.

You are not trying to prove that your original idea was correct.

You are trying to find whatever version of the idea reality supports.

Why I like this one

It’s a small, understandable problem.

It happens at a moment when people are already spending money.

There is a deadline.

There are real constraints.

The output is obvious.

And you can test the entire business without building the software.

That’s exactly the kind of idea I like.

Maybe nobody will pay because existing wedding-planning tools are good enough.

Maybe couples will pay €100 just to make the problem disappear.

Maybe the real customers are wedding planners.

Maybe the real business is corporate event seating.

Maybe the interesting feature turns out to be collecting relationship data rather than generating the seating plan.

You don’t need to know any of that yet.

Find someone currently arranging wedding tables.

Offer to solve it for them.

Ask for money.

Then let reality decide.