Make helpful AI present wherever people need assistance.
Not only wherever they can open a browser, download an app or find the right website. For real people, in real places, in real situations.
ABOUT SIMULATINGMACHINES
AI belongs in the physical world too.
While most service AI is focused on resolving support online, we build it for the moment a person is standing inside a store, airport, venue or public space and needs help there.
CONCEPT VISUALIZATION · NO VENUE DEPLOYMENT SHOWNARTIFICIAL CONCIERGE INTELLIGENCE / ACI
A service layer for the place people are standing.
Artificial Concierge Intelligence is our name for a service architecture that connects an approachable interface to approved knowledge, authoritative venue state, supported actions and a clear human handoff. It guides the current Terminal software and the roadmap for future Terminal and Host deployments.
Every complex place becomes understandable.
Every visitor can ask in their own language, every answer ends in a route, a task or a person, and staff focus where judgment and human care matter most.
OUR ORIGINS / TWO BUILDERS
Two builders, one conviction.
SimulatingMachines started from a widening gap: online assistance is improving rapidly, while many place-dependent needs still end in a queue, a missing sign or a search for staff. Two builders began with Terminal software, a grounded venue-state layer and an operating architecture for closing that gap. The name is the method: simulate the machine and the place it will serve, rehearse the service there, then earn the right to stand it on a floor. The ambition is a repeatable physical support layer that venues can deploy and operate responsibly.
OUR PHILOSOPHY / WHAT WE HOLD TRUE
Machines that understand the place.
Six convictions decide what we build, what we promise and what we refuse.
Help is a physical problem.
Where, which way, is it in stock, can I make it in time: the useful answer is a route that holds, a shelf that is really there, a person reached in time.
A place speaks more than language.
Words carry meaning; places carry geometry, stock, motion and time. So our agents carry the venue itself: the map, the inventory, the routes.
Evidence before eloquence.
Agents speak naturally but never own the facts. Answers come from systems the venue approves, so precision is earned rather than performed.
Venues learn under human control.
Customers control supplied venue-state and service data under contract. Personal interaction records remain subject to lawful basis, notice, declared roles and data-subject rights; internal training, cross-venue learning and licensing each require separate permission.
People stay the point.
Agents absorb the repeated questions so staff keep the judgment, the care and the exceptions. Handing a moment to a person is a feature, not a failure.
Intelligence is a system, not a model.
The microphone, model, venue state, route, handoff and operating layer reach the visitor as one service. We design and test them by how they behave together.
THE PHYSICAL AI SYSTEM
Presence, intelligence and place.
Our product is not only the screen or the host. The value comes from connecting the complete service chain.
Terminal software
A fixed-position software experience for grounded search, deterministic routes and visual guidance. Its QR opens the public demo on a phone but transfers no route or session state. Production enclosure, stateful handoff and certification remain product work.
AI Hosts
A later embodied form intended to share venue state, approved tools, monitoring and handoff with the Terminal. No Host is deployed today.
Venue state
Today: an authoritative graph, spatial index and deterministic services, with a 3D presentation. Learned venue models and working twins are later research.
Experience intelligence
Eligible operational records can align request, source, route, correction and outcome under customer policy. No production result is claimed.
MAP THE CUSTOMER EXPERIENCE
The physical space is only half the map.
A useful interaction can become an operational signal when the deployment has a lawful purpose and the required permissions. Questions, routes, corrections and outcomes can show what visitors cannot find and where service still needs a person. Customer-facing analytics come first; internal training, cross-venue learning and licensable datasets each require separate rights and evidence.
What people need
Aggregate topic patterns reveal missing information, recurring concerns and demand by place and time.
Where journeys break
Route starts, completions and abandoned paths show where signage, layout or service design creates friction.
What remains unresolved
Handoffs, repeated questions and explicit satisfaction inputs help teams improve content and staffing.
What to improve next
Privacy-respectful trends can inform service placement, stock visibility, wayfinding and frontline support.
CUSTOMER-CONTROLLED DATADECLARED PURPOSESEPARATE TRAINING RIGHTSNO CROSS-VENUE IDENTITY
TRUST IS PART OF THE INTERFACE
Helpful, grounded and honest about its limits.
Approved systems own facts, routes, policies and supported actions. Private or complex situations have an explicit route to staff rather than a confident guess. A beautiful answer that is wrong is not a small failure; it is the failure we design against.
Clear consent
Voice and identification begin only after the person chooses them.
Grounded answers
Operational facts come from sources the venue approves and maintains.
Human escalation
Safety, emotion and judgment always retain a path to a person.
Service before commerce
Safety, accessibility and operational guidance stay separate from sponsored moments.
START WITH YOUR PLACE
Show us where people get stuck.
If your venue has repeated place-dependent questions, authoritative information, a named staff owner and an outcome we can measure, we would like to design a bounded paid Terminal pilot with you.
hello@simulatingmachines.comsimulatingmachines.comsimulatingmachines.ai