THE PART YOUR ATTENDEES NEVER SEE.
Lineup isn't a chatbot with a landing page. It's the system an organiser runs a whole audience through — upload the programme, import the audience, ask everyone what they came for, match it, deliver it, and account for every credit spent.
SIX STAGES, START TO FINISH.
- app.lineup.events/programme
3 conflicts found — Hall B double-booked 13:00, 1 session with no end time, 1 duplicate title. Resolve before accepting.
Day: TueTrack: allNeeds review412 sessionsTime Session State 09:15 Opening plenary Parsed 10:30 Sovereign AI keynote Review 11:00 Model risk in banking Parsed 13:00 Data residency panel Review 15:00 Procurement for AI vendors Parsed 17:30 Exhibitor mixer Parsed 01PROGRAMME
Upload the programme in whatever form it already exists — PDF, Excel, CSV, Word. It's parsed into structured sessions with times, tracks, rooms, speakers and capacities, then handed straight back to you for review. Nothing is ever written silently: every parsed session lands flagged for confirmation, and the conflicts it finds — a room double-booked, a session with no end time, a duplicate title — are listed before you accept anything.
- app.lineup.events/audience/importStep 2 of 3Map your columnsFirst namefirst_nameEmail addressemailMobilephoneJob titleroleCompanycompany1,204 rows·1,187 valid·12 duplicates·5 unusable numbers02
AUDIENCE
Import your attendee list as a CSV and map the columns once. Phone numbers are normalised to international format, duplicates are caught by both email and number, and anything unusable is shown to you before a single credit is committed.
- app.lineup.events/campaigns/newChannelQuestionsHoursReviewChannelWhatsApp · 1 credit eachQuestions5 of 5Contact hours08:00–20:00, attendee local timeAudience1,187 attendeesCredits required1,187Balance before2,500Balance after1,313Blocked if the balance is short.03
CAMPAIGN
Choose WhatsApp or an AI voice call, edit the five questions, and set the hours you're willing to contact people — applied in each attendee's own timezone, not yours. The exact credit cost is shown before you launch, and the launch is blocked rather than overspent if your balance is short.
- app.lineup.events/campaigns/liveSent1,187Delivered1,164Replied812Completed735Opt-out rate0.8%ModeMock providersThreadWhich days are you on site?Tuesday and Wednesday.Anything already fixed in your diary?Extracted profileenterprise buyerssovereign cloudTue + Wedstand 12:00–14:3004
CONVERSATION
Watch it run. Every message in and out is recorded, with delivery state and cost. Open any attendee to read the actual conversation. And because every provider has a full mock mode, you can run the entire flow live in a room with a client — no WhatsApp account, no phone number, nothing spent.
- app.lineup.events/curationSarah N.Tomás R.Aoife K.Jonas W.09:15Opening plenary0.71
Broad framing before her 10:00.
15:00Procurement for AI vendors0.93She came to meet enterprise buyers.
17:30Exhibitor mixer0.84Unstructured time with buyers.
Kept free for meetings · 1 hrSwap 15:00Model risk in banking · 0.62Data residency panel · 0.5805CURATION
Every session is scored against what each person told you, then run through a scheduler that refuses to double-book them, respects the days and hours they gave you, caps how much of any day it fills, and protects an unbroken block for their own meetings. Every recommendation carries one sentence explaining itself, in the attendee's own words.
- app.lineup.events/deliverySchedule pageAI Festival 202609:15Opening plenary15:00Procurement for AI vendors17:30Exhibitor mixerKept free · 11:00–12:00Tue 3 Mar09:15Opening plenary10:00Client meeting15:00Procurement for AI vendors17:30Exhibitor mixer11:00Kept free — Lineup06
DELIVERY
They get a message, a link to their own schedule page, and calendar invites. Change the programme and re-send: the invites update in place rather than piling up duplicates.
BUILT FOR PEOPLE WHO GET ASKED HARD QUESTIONS.
Every table is tenant-scoped.
Row-level security on all of it, enforced in the database rather than the application.
The parser never writes silently.
Everything it extracts waits for a human to confirm it.
The credit ledger is append-only.
Charged on confirmed delivery, never on send. Corrections are new entries, so the history is intact.
The limits are in the database, not the prompt.
Five questions maximum, opt-out honoured immediately, follow-ups capped — enforced in code, so no model behaviour can exceed them.
Full mock mode.
Every external provider has a simulated implementation, so the whole system runs end to end with no accounts and no spend.
Calendar invites are done properly.
Correct timezones, stable identifiers, and updates that replace the original instead of duplicating it.
Attendee data stays the organiser's.
Processed on your instructions, opt-outs honoured immediately, personal data deleted after the event by default.
PROGRAMME IN, PERSONAL SCHEDULES OUT.
THE PROGRAMME BRIEF FOR NEXT YEAR, WRITTEN BY YOUR AUDIENCE.
DEMAND PER SESSION, BEFORE THE DOORS OPEN
You see which sessions are being matched to hardest while there's still time to move them into a bigger room.
EVERY ATTENDEE'S GOALS, IN THEIR OWN WORDS
Not a satisfaction score after the fact — what each person said they came for, before the event started.
THE SESSIONS NOBODY IS MATCHING TO
The dead slots surface early enough to re-cast them, re-title them, or merge them into something people want.
SEE IT RUN ON YOUR OWN PROGRAMME.
Send us your programme and a hundred attendees. You'll see the whole thing work before you pay for anything.