Find it. Route it. Keep moving.
A visualization of the interaction demonstrated in software today: ask naturally, ground the result and turn it into a deterministic route.
We are building walk-up AI service for places where people need a clear answer, an exact route or the right person, starting with Terminal software you can try today.
EXPANSION HYPOTHESES · ONE FIRST WEDGE WILL COME FIRST
ONE HELPFUL INTERFACE / EVERY PHYSICAL MOMENT
SimulatingMachines is building a service that turns questions on the floor into grounded answers, exact routes, supported actions and a clearer operating picture for the teams behind the experience. A person speaks, the place answers with evidence, the person acts, and the venue can improve under human control.
THE CLIENT-FACING OPERATIONS PLATFORM
The current Operations preview shows how service, experience and technology teams could review conversations, evidence, handoffs, routes, deployment health and release quality in one governed workspace.
Two complete synthetic Operations views, presented at a scale where the evidence stays legible.

A decision-ready view of customer conversations, self-service resolution, hidden demand and the next operating actions.

A terminal view of 37,280 traveler conversations, assistance peaks, route completion and urgent human handoffs.
Synthetic demonstration data · No live customer systems
PHYSICAL AI / IN MOTION
Three short product films show the system at work across retail, live events and the next generation of Host Agents.
A visualization of the interaction demonstrated in software today: ask naturally, ground the result and turn it into a deterministic route.
A concept for one approved interface guiding a guest from the first question to the route home.
A human-scale volumetric host can bring the same conversation, knowledge and spatial guidance into the room.
Product visualizations. Demo data and venue environments are illustrative.
TWO SURFACES / ONE INTELLIGENCE LAYER
Terminal Agents are where the service begins. Host Agents are the intended embodied form. simLabs owns the integration, evaluation and behavior across models, venue state, interfaces and the learning loop.
Terminal software for walk-up service: grounded questions, curated deterministic venue facts and routes, plus clearly fictional adaptive results for unfamiliar questions. The public demo is software today; production hardware and live venue integrations are the next proof.
Explore Terminal AgentsThe intended embodied form of the same service: approachable voices and visual identities grounded in approved knowledge, house rules and supported actions.
Explore Host AgentsResearch and engineering across venue state, simulation, evaluation, physical interfaces and governed learning. simLabs owns how the complete service is integrated and judged, including models we adopt or train.
Explore simLabsOUR OPERATING THESIS / ARTIFICIAL CONCIERGE INTELLIGENCE
Artificial Concierge Intelligence is our term for the target architecture: a conversational interface, authoritative venue state, approved tools, spatial guidance and a clear handoff to people. We integrate and evaluate the complete behavior, including models we adopt or train. The agent can speak naturally, but it never owns the facts: prices, availability and routes come from systems the venue approves.
The target Terminal combines enclosure, display, audio and compute for the environment it serves. Production hardware remains product work.
The current demo uses adopted speech models, deterministic services for curated venue facts and routes, and clearly fictional adaptive results for unfamiliar questions. Future models may be adopted, adapted or trained.
The current demo binds synthetic products and services to a versioned navigation graph, then presents the computed route in 3D.
Authorized outcomes and corrections are intended to become governed evaluations, scenarios and, where rights permit, training assets.
PRODUCT VISUALIZATION / DEMO INTERACTIONFROM QUESTION TO DESTINATION
A useful answer is not a paragraph. It is a route that holds and a shelf that is actually there. The current demo connects synthetic products and services to exact locations on a deterministic graph, then animates the route from where the visitor is standing.
EXPANSION HYPOTHESES / ONE WEDGE FIRST
These environments illustrate where the service could transfer. The first paid wedge will be selected by repeated place-dependent needs, authoritative state, a named operating owner and measurable completion.
See all use cases ↗
Airports, train stations, metros and bus terminals: departures, platforms, services and animated routes.

An in-store rep for every aisle: exact items, product questions, returns and loyalty-aware service.

Entrances, schedules, merchandise, artist moments and an official host for every guest.

A calm front door for appointments, practice questions, check-in and sensitive handoff to the care team.

Private guidance, service triage, document preparation and a direct route to the right specialist.

Campaigns, sponsored worlds and product stories that answer back and turn attention into useful action.

A front desk for every tenant: visitor check-in, meeting rooms, deliveries and the path to the right team.
OUR MISSION
Most AI service lives behind a browser: brilliant at answering, absent from the aisle, the concourse and the queue. We build Artificial Concierge Intelligence for the physical moment, where a traveler is hunting a gate, a shopper is holding a list someone else wrote and a guest needs help now. SimulatingMachines was founded to close that gap: for real people, in real places, in real situations.
Read our story ↗START WITH YOUR PLACE
We are looking for a first paid Terminal pilot: one venue, one repeated place-dependent problem, authoritative state and a result we can measure against the current service.
Talk to us