OnniCisco

CDA initiative · Onni + Cisco · Municipalities across Brazil

The smart networkof the SUS.

Data, AI agents, observability and security working together, so a municipality can see its whole health network and act on what matters to the citizen.

01 What the project is

Smart health network diagram: the SUS Command Center on top, AI agents serving UBS, UPA, regulation, admissions and audit, Splunk observability, the health data lakehouse in the center with network systems on the left and outputs on the right, and Cisco security at the base.

The smart health network: data, agents, observability and security integrated. Open at full size

A health network that sees itselfand acts in time.

Today, each point of care keeps its own records in its own system. The smart network connects them without replacing them. The citizen’s journey becomes one continuous story, the manager sees the whole network in one view, and professionals get AI support during care, always keeping the decision.

How to read the diagram

On top

Command Center

One center where the network is followed in real time.

At the point of care

AI agents

Support for UBS, UPA, regulation, admissions and auditors. Professionals validate clinical decisions.

Watching everything

Splunk

Detects what went off course and shows who needs to act.

In the middle

Integration bus and data

Connects existing systems and turns their records into reliable information.

Around it all

Cisco security

Protects people, equipment and services from end to end.

Proposed architecture. Products and integrations are defined in the scope of each municipality.

02 Command Center and AI agents

AI beside the professional.One center watching the network.

Clinical copilot

An AI assistant that follows the visit from start to finish: it listens, organizes the record and checks protocols, so the professional looks at the patient and not at the screen.

AI agents

Specialized programs, each with one task, like a support team behind the scenes. Everything reaches the professional ready for review.

Command Center

A center where AI agents, doctors and nurses manage the citizen’s journey in real time, from primary care to production.

From the visit to the network, in three moments.

1 · At the UBS

The record is born from the conversation.

  1. Day’s schedule
  2. Microphone on
  3. Agents prepare documents
  4. Protocols and quality rules
  5. Doctor reviews and signs
  6. e-SUS/PEC updated

The AI prepares and suggests. The professional reviews, decides and signs.

2 · At the UPA

The same logic, at the pace of urgent care.

  1. Check-in
  2. Triage with the nurse
  3. Consultation
  4. Exams
  5. Reassessment
  6. Outcome

The copilot organizes and flags. The nurse defines the risk classification.

3 · At the Command Center

See, anticipate, locate and act.

  1. See: how are we doing?
  2. Anticipate: what is getting worse?
  3. Locate: where to act?
  4. Act: who does what today?

Primary care, regulation, UPAs and production in one view. Doctors, nurses and agents act together.

03 What the bus is

One path connectingevery system in the network.

Think of a translator in the middle of the network. Each system keeps working as it does today, in its own format, and speaks only to the bus. The bus translates, organizes and delivers the information to whoever needs it.

Primary careUPAHospitalLaboratoryImagingRegulationPartner clinics
Integration bus
Single citizen historyCommand CenterNational network (RNDS)

Nothing is replaced

Current systems stay where they are and remain the official record of each point of care.

One history

What each unit records about a citizen comes together in one place, with the source and date of every item.

What happens vs. what should happen

The bus brings what is happening. Clinical intelligence says what should happen. The Command Center works on the difference.

04 Inside the bus

How a piece of informationtravels through the network.

  1. 01It enters through its own doorThe data arrives in the format of the system that produced it.
  2. 02It is translatedThe original is kept and the data is converted into the bus’s common model.
  3. 03It finds the personA single registry identifies the citizen, even with different records across systems.
  4. 04It joins the historyThe information is added to that citizen’s single history.
  5. 05It reaches whoever needs itProfessionals, management, the national network and AI agents consult it with the right permissions.

The whole journey in the municipal network

  1. Registration at the UBS
  2. UPA
  3. Regulation
  4. Admission
  5. Discharge
  6. Return to the UBS

Who gains what

Health professionals

The full history at the moment of care.

Regulation

The queue with the clinical context needed to decide.

Municipal management

The view of the network, and SUS production checked before it is sent.

IT and integration

One monitored integration point, with an error queue and reprocessing in one click.

Data protection and audit

Every access recorded, with purpose and justification.

Privacy applied at every reading

Care relationship

A professional reads the history only of those in their care.

Purpose

Every reading says what it is for. Without a purpose, it is refused.

Minimum necessary

Each role sees only what the work requires.

05 The Splunk layer

When something goes off course,the network notices.

Splunk connects the technical condition of the platform to the performance of the health service. It identifies deviations, links related events and shows the Command Center the impact, who is responsible and the result of the response.

  1. Detect
  2. Correlate
  3. Trigger the workflow
  4. Confirm the result
Splunk in event-driven SUS management: event sources on the left, the SUS application inside Splunk in the center with six steps from receiving events to generating occurrences, and the Onni platform and Command Center on the right. A simulated regulation example runs along the base.

Splunk in event-driven SUS management. The example uses simulated data. Diagram in Portuguese. Open at full size

Examples of rules applied to health

AreaWhat is detectedWhat happens
APSFollow-up pending past the agreed deadlineA task goes to the reference team
UPAReassessment pending, or a failure that affects the queueThe responsible team is notified
RegulationRequest with no progress beyond the deadlineQueue reviewed; bottleneck located
AIH and productionRejected file close to the operational deadlineCorrection opened and acceptance followed
ContractsObligation falling due with no evidenceDocuments requested; management informed

Rules are defined and approved by the health department and those clinically responsible. An alert does not replace a clinical decision.

06 The smart cabin at the UBS and the UPA

A cabin that bringscare closer to the citizen.

The cabin combines Cisco collaboration devices, Meraki connectivity and Onni software to offer assisted measurements, teleconsultation and an integrated record. At the UBS it supports follow-up and access to remote support. At the UPA it supports reception and risk classification by the nurse.

Smart Triage at the UPA: an AI-assisted glass booth where a patient and a nurse use remote clinical support, the Onni copilot and vital-sign equipment, with Cisco Meraki connectivity, risk classification by a trained nurse, and integration with the interoperability bus.

Smart Triage: the cabin in its UPA version. The infrastructure is the same at the UBS; the care flow adapts. Project concept. Open at full size

Three layers working together

Collaboration

Bring the local team closer to remote support, with video and audio.

Connectivity

Carries the unit’s traffic securely and applies the network’s policies.

Clinical application

Connects sensors, Onni agents and the medical record through the bus.

The Cisco devices

Cisco Desk Pro + Webex

Screen, camera and audio for assisted teleconsultation and case discussion.

Webex Control Hub

Manages devices remotely: setup, updates and diagnosis.

Cisco Meraki · Wi-Fi

Managed wireless network, with policies defined by the project.

Managed PoE switch

Organizes wired connections and makes diagnosis and expansion easier.

Cisco Meraki MX

Unit security and SD-WAN: connects the unit to central services.

Meraki Dashboard

One place to manage the Meraki equipment of every unit.

Blood pressure, temperature, pulse, oxygen, weight and height sensors are clinical equipment of their own. The network carries the data; it does not interpret its clinical meaning. Models, quantities and licenses depend on sizing.

07 Benefits and innovation in the SUS

What changes with this architecture

  1. From reacting to anticipating.
  2. From loose systems to one network.
  3. From typing to caring.
  4. From a late report to action today.
  5. One focus: the citizen.

Citizen

More access

Follow-up of measures, pending items and next steps, closer to home.

Professional

More agility

Less manual recording and the protocol in plain sight during care.

Manager

More quality in the data

The whole network in one view, anticipating disallowances and transfer problems.

IT

Remote infrastructure management

Several cabins supported without depending only on on-site visits.

Expected benefits. Efficiency results are assessed against each municipality’s baseline, with no prior guarantee.

08 The Cisco security layer

Protection for people,equipment and services.

Security protects the connection between professionals, cabins, legacy systems and platform services. The goal is to reduce exposure and limit the reach of incidents, keeping controlled access to care and management.

  1. IdentityCisco DuoConfirms who is accessing the solution.
  2. AccessCisco Secure AccessGrants access only to authorized applications, including remotely.
  3. WorkstationsCisco Secure EndpointProtects the computers used in consultation, triage and administration.
  4. ServicesCisco Secure WorkloadRestricts communication between services, limiting how far an incident can spread.
  5. Security operationsCisco XDR + Splunk ESInvestigates occurrences and links signals from different sources.
  6. ResponseXDR Automation · Splunk SOARRuns defined playbooks, with approval when needed.

Example: an incident at the cabin

  1. Suspicious activity
  2. Detection
  3. Impact on the unit
  4. Approved playbook
  5. Care in contingency
  6. Verified recovery

These controls are complementary architectural options, chosen by risk, existing environment and licenses. Who may access each medical record is decided in the Onni application and the clinical systems, by role, unit and purpose.

Bring the smart networkto your municipality.

An initiative with Onni and Cisco, working with municipalities across Brazil.

Talk to Onni

Let's talk about your AI.

Tell us who you are and where you want to start. We reply within one business day.