OccasionAI
Parents spend hours searching for venues, catering, and entertainment for children's parties. We built a planning app that takes a description of the occasion and returns a structured task list with real local vendor options — selected within the app.
Planning a birthday isn't hard because the tasks are hard. It's hard because nobody tells you what the tasks are.
A parent planning their child's birthday party faces a disorganised mix of tasks: find a venue, arrange catering, sort entertainment, send invitations, organise a cake. There's no standard order. Nothing surfaces what they've missed. Every booking is a separate search loop — a venue directory here, a review site there — each one producing a list of options with no guidance on which one is actually the right fit.
Generic search doesn't solve it. Aggregator sites exist for some categories but not for the full event. Chat-based AI assistants give general advice but can't surface real local options. The parent ends up with a dozen browser tabs, a note in their phone, and the persistent anxiety that they've forgotten something.
The problem isn't that parents don't know what kind of party they want. It's that nothing translates that into an organised plan with real, local options attached.
Describe the occasion.
Get a plan with real vendor options.
Stack
The flow is structured, not open-ended. No chat thread, no back-and-forth conversation. The parent answers a small set of questions — what they want to plan, location, date and time, rough budget, and any constraints like food allergies — then the app generates a proposed task list from that input.
Describe the occasion
One free-text box: "Noah's 6th birthday party in London." Then a short structured form — date, time, budget, number of guests, and any safety-critical details like allergies. No lengthy questionnaire. One screen, a few fields.
Review the proposed task list
The app generates a proposed checklist of tasks, each tagged as a vendor task (Book Party Venue, Book Entertainment) or a plain checklist item (Send Invitations, Order Cake). The parent reviews it, unticks anything they don't need, and adds anything that's missing. One AI call generates the full proposal — not one per task.
Get the full plan with real vendor options
Once confirmed, every task gets a real due date calculated back from the event. For each vendor task, the app surfaces a shortlist of real local options — each with a plain-language reason it's a good fit for this occasion. Location, capacity, pricing context, and any allergy or safety flags. The parent selects a vendor within the app; choosing a different option swaps the choice rather than adding a second one.
Return to the plan any time
The plan is persistent. A parent can return to it at any time and see exactly where things stand — what's done, what's due, what's still to book. No state lost between sessions.
Real options — not a directory.
Every vendor task in the plan comes with a shortlist of real local options. What makes this different from a search result list is the reason each option appears. Not a review score — a sentence tied to the actual occasion: location match, capacity for the number of guests, whether a nut-free policy is confirmed, pricing in range.
The parent evaluates the match, not the star rating. That's a meaningful difference when the constraint isn't generic preference but a safety requirement like an allergy.
Matched to this occasion
Vendor options are surfaced against the specific brief — location, guest count, budget, and any flags — not shown as a generic list filtered by category.
Plain-language reason
Each vendor card explains why it's a good fit in a sentence. The parent can evaluate it immediately without clicking through to a separate review page.
Select, don't stack
Choosing a different vendor swaps the selection rather than adding a second one. One venue per task, no accidental double-booking.
Safety flags surfaced
Allergy and dietary constraints noted at input are carried through to vendor matching — flagged on any option where the policy is unconfirmed, not buried in a review.
Birthdays first.
The architecture runs further.
The product currently supports birthday planning. The structured flow, vendor matching engine, and task generation are built for this occasion type — not retrofitted from a generic template.
The underlying architecture — description in, AI-generated task proposal, vendor shortlist with reasons, persistent plan — is not birthday-specific. Adding a new occasion type means configuring the task templates and vendor categories for that occasion, not rebuilding the flow. Weddings, christenings, and school events are in the roadmap.
Birthdays — live today
Children's birthday parties. Structured task generation, local vendor matching, due-date scheduling. Full flow from description to persistent plan.
Weddings
In roadmap. Same architecture, different task templates and vendor categories. Longer planning horizon.
School & community events
In roadmap. PTA events, school fairs, community gatherings. Group coordination layer being designed.
Related
Also relevant
Building a consumer app with real data behind it?
30 minutes. No pitch. An honest conversation about your product and whether this approach is the right fit.