Ziva AI Agents · in practice
See it from your seat.
The same two agents change the day for everyone on the care team in different ways. Pick a role to see what lands on that person's desk, what no longer does, and one real case from start to finish.
One clinic, one day
Hundreds of changes come in.
A handful reach a person.
This is what a 300-patient panel might generate in a single day. Continuous Intelligence classifies every change. The Care Agent handles the follow-up, so only the exceptions reach a person.
Illustrative numbers for a hypothetical panel. They are not outcomes data.
Data points receivedReadings, labs, symptoms, events
1,240
Changes evaluatedAnything that moved
318
ClassifiedEvery change gets a state
No action 271
318
Handled by the Care AgentRepeat readings, check-ins, confirmations
31
Reach a personWith reason & suggested step
16
4 high priority12 needs review31 observation271 no action
By role
What changes for each person
on the care team.
Each role sees only the work that belongs to them.
DP
Dr. Priya Shah · internal medicine
1,400-patient panel · sees 22 patients a day
Before
×An inbox full of out-of-range readings, most of them noise
×Hears about ED visits days later
×Has to work out the context before deciding anything
With Ziva
✓A short review queue, each item with its reasoning
✓Adherence and symptoms already checked
✓Decides, approves or edits in one step
One case · Robert K.*, BP drift
Mon
IntelligenceNeeds review
Systolic BP ~160 for 9 days against a baseline of ~130. Still below the 180 threshold.
Mon
Care AgentChecks adherence before it reaches Dr. Shah
"Have you been taking your evening dose?" → "I forget it most nights."
Tue
Dr. ShahMoves the dose to the morning, in 40 seconds
2 wks
Care AgentRepeat readings collected · back near 132
ZivaDocPro · My review queueTuesday 7:45am
3Need you
27In observation
1,370No action
Robert K.*, 67Needs review
BP ~160 × 9 days (baseline ~130). Misses evening dose most nights, per Care Agent.
Suggested: move dose to morning; recheck 2 weeks.
ApproveEditMessage patient
Maria L.*, 72High priority
Day 7 after discharge: ankle swelling, weight +3.1 lb in 48h. Nurse Jenna already on it.
Linda P.*, 64Needs review
Cardiology note received and summarized: Holter ordered. No action needed from you yet.
No longer chasing
Normal readingsOne-off spikes"Did they take it?"Missing context
JR
Jenna R.*, RN · care manager
220 patients on monitoring · about 25 transitions a month
Before
×Calls every discharged patient herself, and half the calls go to voicemail
×Chases missing readings by hand
×Problems are found at day 14, not day 1
With Ziva
✓Ziva runs the routine check-ins
✓She gets the 3 that matter, with the trend attached
✓Her time goes to clinical conversations
One case · Maria L.*, day 7 after discharge
9:02am
Care AgentDay 7 symptom check-in
"Any swelling or shortness of breath?" → "My ankles are puffy since yesterday."
9:03am
IntelligenceHigh priority
Swelling plus weight +3.1 lb in 48h in a heart failure patient → sent to Jenna.
9:20am
JennaCalls Maria, then reaches Dr. Shah · diuretic adjusted
Days 8–12
Care AgentScale readings watched · reminder only if one is missed · stable by day 10
ZivaDocPro · Care managerToday
3Escalations
41Check-ins done by Ziva
6Awaiting reply
Maria L.*High priority
Ankle swelling · weight 168.2 → 171.3 lb · HF admission 7 days ago.
Call nowLoop in physician
Tom W.*Needs review
Day 3: unsure which of two inhalers to use. Isabella shared the discharge instructions, but he's still unsure.
Grace O.*Needs review
No response after 3 attempts over 2 days.
No longer chasing
Routine day 1 callsVoicemail loopsRe-requesting readings"How are you feeling?"
LM
Luis M.* · care coordinator
Referrals, prior authorizations, scheduling
Before
×Tracks referrals in a spreadsheet
×Phone tag with patients and offices
×Finds out about prior auth problems weeks late
With Ziva
✓A list of only the stuck items, each with the reason
✓The patient's own words attached
✓Ziva confirms when each item is resolved
One case · Linda P.*, referral wait
Day 3
Care AgentChecks cardiology scheduling
"Have you booked your appointment?" → "Earliest was two months out."
Day 3
IntelligenceNeeds review
The wait is longer than the referral's urgency allows → sent to Luis.
Day 5
LuisSecures a slot in 2 weeks with a partner group
Day 24
Care AgentVisit confirmed · specialist note summarized · loop closed
ZivaDocPro · BarriersOpen items
5Stuck
63On track
19Closed this week
Maria L.*Prior auth
"The pharmacy said it needs approval." New HF medication not started, 1 day after discharge.
James T.*Cost
"It's $480. I'm not sure I can keep paying that." GLP-1, day 2.
Linda P.*Access
Cardiology wait of 2 months, but the referral is marked within 4 weeks.
No longer chasing
Referral spreadsheets"Did you book it?"Missing specialist notes
KB
Keisha B.* · medical assistant
Device onboarding, reminders, outreach
Before
×Spends hours a week on reminder calls
×Doesn't know who has stopped sending readings or why
×Keeps trying numbers that have changed
With Ziva
✓Ziva sends the reminders
✓She gets device problems and patients Ziva can't reach
✓Once she fixes the problem, Ziva picks up where it left off
One case · Grace O.*, unreachable
Day 1–2
Care Agent3 attempts by text and in-app at different times
Day 2
IntelligenceNeeds review
The protocol limit was reached → routed to the MA queue.
Day 3
KeishaReaches Grace's daughter; number had changed · updated
Day 3
Care AgentResumes the check-in · Grace replies within the hour
ZivaDocPro · OutreachToday
4Need you
212Reminders sent
188Readings in
Grace O.*Unreachable
3 attempts over 2 days, no reply. Last contact 9 days ago.
David R.*Device
"Scale battery died." Ziva will re-request the weight tomorrow.
Ahmed N.*Device
BP cuff hasn't synced in 6 days, even after a reconnect guide was sent.
No longer chasing
Reminder call listsUnanswered texts"Did you take your reading?"
AC
Alex C.* · practice leader
Operations, quality, staffing
Before
×No view of whether follow-ups actually happen
×The only answer to more patients is more staff
×Protocols live in people's heads
With Ziva
✓Every workflow is tracked from start to close
✓Staff time goes to exceptions
✓One written protocol that Ziva follows every time
Setting up a transitions protocol Configurable protocols · coming soon
1
Trigger
Discharge from a hospital, ED or SNFSent by the discharging facility · for patients enrolled in monitoring
2
Schedule
Day 1, 3, 7, 14, 30Your days, not Ziva's
3
Questions
Meds, instructions, follow-up, symptomsApproved wording
4
Routing
Symptoms → RN · barriers → coordinator · unreachable → MABy role
ZivaDocPro · Workflows overviewLast 30 days
38Transitions open
112Closed
1.4Staff touches each
Transitions of care94% reached by day 2
Referral loops closed81%
New meds confirmed started88%
Illustrative dashboard. These figures are not reported outcomes.
ML
Maria L.*, 72 · patient
Home after a heart failure admission
Before
×Sent home with a stack of paperwork
×Not sure who to call, or whether a symptom matters
×Waits for the clinic to call
With Ziva
✓Isabella checks in, in plain language
✓Problems reach the right person the same day
✓Feels looked after between visits
Maria never sees the four states or the routing. She simply hears from her care team, through Isabella, when it matters, and gets help when something's wrong.
✦
Isabella · Riverside Cardiology
Text messages
Day 1
Hi Maria, it's Isabella from your care team. How are you feeling since coming home?
Tired but ok. I couldn't get the new pill, the pharmacy needs approval.
Thanks for telling me. I've asked Luis on your care team to sort out the approval. You don't need to call anyone.
Day 2
Good news: your new medication is approved and ready at the pharmacy. Can you pick it up today?
Yes, this afternoon.
Day 3
Were you able to start it? Your follow-up visit is this Friday at 10am.
Started this morning. See you Friday.
Day 7
My ankles are puffy since yesterday.
Thank you, that's important. Jenna, your nurse, will call you in the next few minutes.
*Stories are inspired by real patient experiences and health patterns observed in care. Names, images, quotes, and certain details have been changed or combined to protect privacy. Individual experiences and outcomes may vary.
One discharge, every role
Who does what over 30 days.
Maria*'s transition from start to finish. Ziva fills most of the grid. People appear only where their judgment or action is needed.
Care AgentIntelligencePerson
Discharge
Day 1
Day 2–3
Day 7
Day 8–14
Day 30
Hospital / SNFDischarging facility
Discharges MariaSends discharge summary
Patient
Goes home
"Pharmacy needs approval"
Starts medication
"Ankles puffy"
Weighs at home
Stable at home
Ziva agentsIntelligence + Care Agent
Starts workflowSummarizes discharge docs
Day 1 check-inNeeds review
Reminds pickupConfirms meds & visit
Symptom checkHigh priority
Watches readings, reminds if missedBack to baseline
Hands off to ongoing care
Coordinator
Submits prior auth
Nurse
Calls Maria
Physician
Friday visit
Adjusts diuretic
MA / front desk
Hatched cells: nothing needed from that role. In this example, people stepped in 4 times over 30 days.
*Stories are inspired by real patient experiences and health patterns observed in care. Names, images, quotes, and certain details have been changed or combined to protect privacy. Individual experiences and outcomes may vary.
Routing map
Every exception has an owner.
Your organization decides who receives what. Here is a typical setup. Ziva sends each exception to that role with the context.
| Exception | State | Physician | Nurse | Coordinator | MA | Ziva continues |
|---|---|---|---|---|---|---|
| Serious symptoms after discharge | High | |||||
| Sustained change from baseline | Review | |||||
| Medication not obtained · cost or PA | Review | |||||
| Concerning side effects | High | |||||
| Referral stalled or note missing | Review | |||||
| Patient unreachable | Review | |||||
| Device not syncing | Review | |||||
| Single elevated reading, no symptoms | Observe | |||||
| Short-term volatility (e.g. one bad night) | None |
Primary ownerInformed
When patients don't answer Configurable protocols · coming soon
The Care Agent doesn't give up after one message.
An example of a no-response protocol. Ziva tries different channels and times before a person is involved. Your organization sets the number of attempts and the timing.
Attempt 1 · 9am
Text message
Isabella's check-in in plain language.
Reply → workflow continues
Attempt 2 · 2pm
In-app message
Push notification from the Ziva app.
Reply → workflow continues
Attempt 3 · next day
Voice call Soon
Isabella calls at the patient's usual response time.
Answer → workflow continues
Attempt 4 · day 2
Caregiver contact
If the patient has consented to it.
Reply → workflow continues
Protocol limit
Escalate to a person
Sent to the MA with every attempt logged. Higher-risk patients can escalate sooner.
See it with your own team.
We'll map your roles, protocols and routing, then walk through a week of your patients the way your team would see it.