Hospitals and clinics

Nobody wants to get lost in a hospital. Now the way, the hours and the rules are on their own phone.

A hospital is a building people visit when they would rather not: clinics on different floors, a laboratory, radiology, registration, visiting hours, rules for companions, the car park. Patients and visitors ask the same questions at every desk and in every corridor, often in a language the staff do not speak. Today the answers sit on signs, at a busy information desk and with staff who have patients waiting. ToRun's AI assistant brings them to the visitor's own phone, starting from the spot where they are standing.

The software is live. We want to set up the first hospital together with its patient services team.
Main entrance
Where do I go for a blood test?
The laboratory is on the 1st floor. Take the lifts on your left, and register first at the desk opposite the lifts.
Until what time are samples taken?
From 7:30 a.m. to 4 p.m. on weekdays.
Illustrative example. The answers come from the documents the venue uploads.
Why it fits

Why a hospital needs it more than most buildings

A building nobody knows

Most patients come only a few times a year. Floors, wings and clinic names are new to them every time, and a sign cannot answer a question.

Every language, every age

Patients from abroad, visitors from out of town, elderly companions. They can type or speak, and the answer comes back in the language of the question.

Staff time belongs to patients

Every “where is radiology?” at the nurses' station is time taken from care. Questions like that are answered at the spot where they come up.

Rules that change

Visiting hours, companion rules, parking, a clinic that moved floors. When the document is updated, the answer changes with it; a printed sign stays as it is.

What patients and visitors ask

From the car park to the clinic door

The code at the main entrance and the code in the car park are different codes, so the assistant knows where the person is standing. The answers come from the clinic list, floor information, visiting rules and service information the hospital uploads.

Car park, level −2
How do I get to cardiology?
Take the lift to the 3rd floor and turn right as you step out. Registration is at the desk on your left.
Inpatient ward, 5th floor
Until what time can visitors stay?
Visiting hours on this ward are 1 p.m. to 8 p.m. One companion may stay overnight with a pass from the ward desk.
Main entrance
Is there a prayer room?
Yes, on the ground floor next to the cafeteria. There is also a quiet room on the 2nd floor.
Same visit, a moment later
And how do I get there from here?
The assistant remembers the previous question; nobody has to explain it again.

Illustrative examples. Answers are generated from the information the venue uploads; when the documents have no answer, the assistant says so and the question goes to the report.

Who gains

Four sides of the hospital, four different gains

The same setup gives each side something different: the patient finds the way, the information desk gets time back, the technical team hears about a fault from the person who saw it, and management sees where people get lost.

Patients and companions
Find the way without asking three times

Directions from the spot where they stand, the hours and the rules, in their own language, typed or spoken. No account, no name, no patient number.

Patient services and information
Spend their time on the people who need a person

Floors, clinics, hours and rules are answered at the spot. The desk keeps what only people can do: reassuring someone, an exception, a patient who needs to be accompanied.

Technical and support services
Hear about a fault from the person who saw it

“The lift in block C has stopped” or “the accessible restroom on the 2nd floor is locked” reaches the panel with its spot and time, and by e-mail. If nobody picks it up within the set time, it moves up to the next person in charge.

Management and quality
Sees where people get lost

Which spots collect the most questions, at which hours and in which languages, and what people looked for and did not find. Input for signs, staffing and interpreter planning.

Where the codes go

At every point where someone stops to ask

Main entrances and registration clinics, floors, registration, opening hours
Lift halls on every floor what is on this floor, the way to the next clinic
Car parks and drop-off points the way in, parking rules, the nearest entrance
Waiting areas and wards visiting hours, companion rules, the cafeteria
Cafeteria and lounges opening hours, nearby services, the way back to the clinic

A print-ready poster comes out of the system for each spot; the codes can also be printed on existing signs and directory boards.

What you keep

After a month, you know where people get lost

The report is made of questions: what was asked, at which spot, when, in which language, and whether it was found.

Illustrative report
Most-asked topics
The way to a clinic
3,140
Laboratory and blood tests
1,690
Visiting hours
1,210
Parking
830
Asked for, not found
Interpreter service 420 questions
Phone charging point 260 questions
Late-night pharmacy 140 questions
Questions per hour, an average weekday Busiest spot: lift hall, ground floor · Busiest hour: 09:00–10:00
Languages of the questions Local language 71% · English 11% · Arabic 8% · Russian 4% · German 2% · Other 4%

Illustrative report. The real figures come from your own visitors' questions.

Signs

The spots where people ask the way most are where a sign is missing or unclear, and the asked-but-not-found list says what they wanted. It downloads as Excel or CSV.

Guides and interpreters

The spread across hours, spots and languages shows where guides and interpreters are needed, and when.

Patient information

What patients and visitors ask is a record of what confused them: input for the next patient leaflet, the website and the signs.

Setup

How it fits the hospital

1Before it goes live

You upload the clinic list, floor and wing information, visiting and companion rules, parking and services, as Word, PDF, Excel or PowerPoint; no special format. Your staff mark the spots and places on the floor plan, and a print-ready poster comes out for each spot. The page goes live with the hospital's name, logo and color.

2Every day

Patients and visitors scan, ask and read. When a clinic moves or visiting hours change, the document is updated and the answers follow. Fault reports reach the panel and e-mail straight away.

3Month by month

The demand report is in the panel, and the asked-but-not-found list downloads as Excel or CSV: input for signs, staffing and patient information. The spots and codes stay where they are; what changes is the content you upload.

It sits next to your systems Your hospital information system, appointment line and patient app stay as they are; the assistant is not connected to them and never sees a patient record. It books no appointments: it tells people where to go and how, from your documents.
What it does not do It does not diagnose, does not give doses and does not replace a doctor or a nurse. For a health question it gives general, publicly documented information and points to a professional for the personal decision. It is not an emergency channel.
Principles

Three principles

No paid answers

No tenant or exhibitor can pay to be put first; the order of an answer cannot be bought.

It doesn't make things up

The assistant is set up to answer only from the documents the venue uploads. When there is no answer, it says so, and the question goes to the asked-but-not-found list.

No identity required

No account, no form, no name or phone number. The visitor scans the code and asks.

Privacy

We collect questions, not people

The record is made of the question: what was asked, at which spot, when, in which language, and whether an answer was found. The venue sees what its visitors are looking for without knowing who they are.

Minimal dataVisitors are never asked for an account, a name, a phone number, an e-mail address or an ID number. The page asks for neither location nor camera permission.
Recordings are not keptA spoken question is turned into text, and the recording is not stored.
No profilesQuestion records contain no identity, IP address or device information, and no tracking cookies are used. The browser holds only a temporary session key that keeps that visit's conversation going and is deleted when the tab closes.
Your data stays yoursOnly the people you authorize can see it. It is not sold and not used for marketing.
Roles set by contractThe venue is the data controller and ToRun is the data processor. Our Data Processing Agreement is published. Read the Data Processing Agreement
Hosted in the EUHosting is in the European Union (Germany). When the contract ends, data is returned or deleted as you choose; the sub-processors are listed in the agreement.
Protected against misuseAutomatic bot verification and a per-visit question limit; answers and fault reports pass automatic content checks.

Frequently asked

Will it answer medical questions?

It is set up to answer questions about the building and its services from your documents. For a health question it gives general, publicly documented information and one sentence pointing to a professional; it does not diagnose or give doses.

Does it see patient records?

No. It is not connected to your hospital information system, and the page asks for no name, patient number or ID. Question records contain no identity, IP address or device information.

Do we need to translate our documents?

No. You upload your documents in the language you work in; someone who asks in another language gets the answer in theirs.

We run more than one venue.

Each venue is set up with its own content, codes and reports, all managed from one account. Head office writes once and the venues below it answer from the same information; each manager sees only their own venue.

Can we get a surprise bill?

No. The number of questions is set in the contract and usage is visible in the panel at any time; when the number runs out, no extra charge is made without your approval.

Is the AI trained on our data?

No. The AI is not trained on your data; the documents are read at the moment a question is asked. That is why an update to a document shows up in the answers.

Let's set up the first hospital together

The software is ready and live. We want to do the first setup together with a hospital's patient services team: a limited number of spots to start with, and at the end we read the report of what patients and visitors were looking for together. Scope, duration and terms are agreed in a meeting.

Before we meet: we can build a trial page from the clinic and visitor information on your hospital's website, so you can try the assistant on your own phone with your own hospital's information.