VANTRA
Loading this page
Loading this page
Staff ask in plain language and get an answer from the clinic's own written knowledge, its live receptionist configuration, its results, its schedule and its enquiry pipeline - with the sources listed, and a plain answer when nothing covers the question.
Drawn from the assistant's own screen. Every name, time and figure below is invented sample data.
A patient wants to move Thursday's appointment. What's our cancellation policy, and how full is Thursday?
Based on your clinic's knowledge base: a move made more than 24 hours ahead carries no fee. Inside 24 hours, the deposit is held against the next booking.
Thursday 14 May has six booked appointments: 09:30 A. Mwangi (cleaning), 11:00 T. Bekele (review), 14:15 J. Okonkwo (whitening) and three more. From 16:00 the day is free.
My recommendation would be to offer 16:00 before quoting a fee.
Sources used as context
I don't have that information in your clinic's knowledge sources yet.
Add this to Internal Knowledge so the assistant can answer next timeShown to managers only.
I can only help with questions about your clinic - its services, pricing, policies, team, and operations.
Both are fixed sentences rather than the model improvising a refusal, and a question that is plainly about the clinic can never be pushed into the second one.
Staff ask the same kinds of questions they would ask a senior team member and receive either a source-backed answer or clearly labelled in-domain inference.
Source-backed answers use your SOPs, playbooks, handbooks, policies and operational documents rather than browsing the public internet. Any in-domain inference is labelled when no exact source is found.
When an answer uses retrieved material, Central Command shows the underlying sources. When it extrapolates on an in-domain business question, it states that no exact internal source was found.
New team members get up to speed without pulling senior staff into the same conversations over and over again.
When connected source content is refreshed, Central Command answers from the newly indexed material.
Answers are limited to the knowledge and staff access configured for that clinic, rather than exposing the whole tenant by default.
The assistant reads what the clinic gave it, at the moment the question is asked.
| Source | What it holds | How current |
|---|---|---|
| Internal Knowledge | Articles your managers write in the dashboard. Archived ones are left out. | Read when the question is asked |
| Agent Context | The live receptionist setup: hours, services, pricing approach, FAQs, providers, and what your booking and notification settings actually do. | Read when the question is asked |
| Monthly results | Aggregate counters: bookings created, appointment outcomes, calls handled, outbound, reactivation, payments recovered. | This month and the two before it |
| Upcoming schedule | Booked appointments by day, with time, patient, service and status. Cancelled rows are left out. | The next 16 days |
| Leads & enquiries | The recent pipeline by status. When the list is cut, the total is said out loud. | Most recent first, ten named |
| Document | Text extracted from files a manager uploads - up to fifty per clinic. | From upload; the original file is not kept |
| Your website | A cached read of your own public site. Ranked below everything above, because it can be out of date. | Cached, refreshed in the background |
At most ten of these reach any one answer; anything that does not match the question is dropped before the model sees it. There is no web search - the only site it reads is yours.
Every member of the clinic workspace with context access asks in the same chat and keeps their own history.
Writing an article, uploading a document and applying a proposed configuration change all need the manager grant.
Conversations stay in the asking staff member's own history for five days and are then deleted; they are not shared with the rest of the team.
Supply the SOPs, handbooks, clinical pathways, operational documents and internal guides that staff should be able to query.
Team members ask how things work, from shift handover to refund policy to clinical pathway, in plain language.
Central Command links the material used for a grounded answer and distinguishes source-backed facts from clearly labelled business reasoning.
Gaps and unclear processes are surfaced, so leadership can see where documentation needs to be written or tightened up.
The same question can have two correct answers depending on who is asking. This assistant only ever has one audience: the people who work here.
What it can reach
What it cannot reach
Ask it to change something - close Saturdays in August - and it writes nothing. It hands back the change spelled out, and nothing happens until a manager presses Apply, at which point the server validates that change again on its own.
A team member without that grant is told plainly that they cannot change the configuration, rather than being quietly ignored.
Apollo
Schools
Vantra Apollo connects attendance, academic records, fees, family communication and country-specific school administration in one daily work...
Learn moreConcierge
Vantra Concierge is an AI front desk for phone, SMS, WhatsApp, Instagram, Messenger, signature-verified inbound email and persistent webchat...
Learn moreScribe
Medical
Vantra Scribe captures a consultation and prepares structured draft notes and referral letters for the clinician to review, edit and sign of...
Learn moreHMIS
Hospital Core · Private rollout
Vantra HMIS, also known as Hospital Core or HIMS, is a private-rollout hospital platform for multi-facility teams. It brings patient registr...
Learn moreAtlas
Agricultural trade · Live demo
Vantra Atlas is a governed operating system for agricultural exporters. It connects commercial orders, lots, shipments, documents, certifica...
Learn moreProject X
Insurance broker operations · Working title
Vantra Project X is the working title for a production-capable insurance broker operations product. It connects applications, membership, po...
Learn moreLIS
HMIS module · Private rollout
Vantra LIS is the laboratory information system module in Vantra HMIS. It connects the EMR order worklist to specimen accession, structured ...
Learn moreRIS + PACS
HMIS module · Private rollout
Vantra RIS manages imaging worklists, accession and StudyInstanceUID association, radiology reports, finalisation, amendments and clinical c...
Learn moreBook a 30-minute demo and we'll show you what Vantra Central Command would do for your team this week.