MAICE turns a conference into a live, machine-readable operations model and then plans, schedules, monitors and learns over it. You negotiate. It executes.
Serious MICE events still run on scattered tools and heroic manual effort: guest lists in spreadsheets, supplier chases over email, B2B meetings scheduled by intuition, volunteers coordinated by phone. The work gets done — but it does not scale, and nothing learns from it.
disconnected tools stitched together for a single event
of manual work to build one conference agenda and B2B schedule
systematic learning — each event starts from a blank page
MAICE represents each event as an Event Operations Graph — a typed, time-aware model where every task, person, room, budget line and deadline is a node, and every dependency is an edge the system actually understands.
Generic project tools have lists and boards. They do not know what a load-in is, why the AV supplier matters, or that a keynote speaker cancelling at 18:40 touches four rooms, two shuttles and ninety B2B meetings. MAICE does — because the knowledge of how events really run is encoded in the model itself.
The foundation. Delays propagate, conflicts surface, and nothing silently falls through the cracks — the whole event is one connected, always-current picture.
Specialised AI agents draft communication, chase deadlines, watch suppliers and keep the plan current — and every critical action waits for a human to approve it. Deterministic state machines guard contracts and payments; a full audit log records every move.
Agendas, room assignments, staff shifts and B2B matchmaking computed as a constraint problem: minutes instead of days, zero hard conflicts, and every placement explainable.
Compact predictive models, fitted to data from real events, flag risk before it lands: late suppliers, likely no-shows, overloaded teammates, budget lines about to break. After each event, the system runs a structured post-mortem — and gets measurably better.
One graph. Rooms, agenda, suppliers, guests, budget — connected, typed, time-aware.
Agents propose. Amber means a human decides.
Hard constraints never break. Soft preferences are optimized — and explainable.
RISK · 9 DAYS AHEADStage supplier has a pattern of late confirmations on multi-day builds. Suggested action: move confirmation deadline up by 4 days.
FORECASTExpected no-show rate for invited buyers: within planned range. Catering order can be finalized.
Models are small, fast, and fitted to real operational event data — they learn from every event that runs on MAICE.
MAICE is not autopilot. It is an operations core with a hard, designed-in boundary: agents prepare and execute routine work, while every consequential action stops at an amber gate and waits for you. Every decision — human or machine — lands in an immutable audit log. Try it:
Interactive concept demo — no message is actually sent.
MAICE is being developed inside Montaste, a working event company from Bar, Montenegro — and validated on the operational data of its own international events. Our events are the living lab: every module has to survive a real conference before it ships.
The Luxury Travel Summit Montenegro — our flagship international B2B event.
International buyers and exhibitors whose meetings, agendas and logistics MAICE learns from.
LTSM and the Destination Excellence Forum, run with national tourism institutions and industry partners.
Early access opens with a small group of event organizers, DMCs and PCOs, convention bureaus, and hotels with conference capacity. Join the waiting list and we will reach out as pilots open up.
Thank you — we will be in touch as early access opens.
We use your details only to contact you about MAICE early access. No newsletters, no sharing.