AMROAR Technologies

Kalenix Health · Agentforce + Health Cloud — Amroar Technologies
Enterprise Agentforce Health Cloud Service Cloud Healthcare

AI agents handling 74% of patient
interactions for a 28-hospital network.

Kalenix Health had 28 hospitals, 6,400 staff on Salesforce, and a patient services team drowning in intake calls, appointment rescheduling, and referral routing. We deployed Agentforce across Health Cloud and Service Cloud — four specialised AI agents, fully live in 16 weeks.

All Enterprise Cases
Engagement at a glance
16wk
Full implementation timeline
6.4k
Users live at go-live
4
Agentforce agents deployed
74%
Patient interactions handled by AI
Client
Kalenix Health (Fictional)
Industry
Healthcare · Hospital Network
Platform
Agentforce · Health Cloud · Service Cloud
Scope
4 AI Agents · Multi-Cloud · HIPAA
Delivered
16 weeks · 28 hospitals live
The Engagement

A patient services team spending
60% of their day on tasks AI can do.

Kalenix Health operates 28 hospitals and 140 outpatient clinics across the US. Their patient services team of 380 people handled everything from appointment booking and referral routing to insurance pre-authorisation queries and post-discharge follow-up. By 2025, call volumes had grown 34% year-on-year. Headcount hadn't moved.

The team was already on Salesforce Health Cloud and Service Cloud — but using them as a glorified CRM. No automation. No AI. No self-service. Agents were manually triaging every inbound contact, regardless of complexity. A nurse chasing a referral status sat in the same queue as a patient trying to reschedule a routine check-up.

Amroar was brought in to implement Agentforce across the full multi-cloud stack — four specialised AI agents built on Health Cloud data, integrated with their EHR system, and HIPAA-compliant from day one. The goal: handle routine interactions autonomously, escalate clinical complexity to human agents with full context, and give the patient services team their time back.

Client
Kalenix Health
Sector
Healthcare · Hospital Network
Scale
28 hospitals · 140 outpatient clinics
Users
6,400 across clinical + admin staff
Platform
Agentforce · Health Cloud · Service Cloud
Agents Built
4 specialised Agentforce agents
Compliance
HIPAA · HL7 FHIR · SOC 2
Timeline
16 weeks to full go-live
What We Walked Into

380 people. One queue. Every interaction treated the same.

The problems weren't caused by the team — they were caused by a system that gave every contact equal priority and zero automation.

01
No triage logic — all contacts in one queue
A patient rescheduling a blood test and a physician chasing a complex referral sat in the same inbound queue. Agents had no way to prioritise. Average wait time was 22 minutes regardless of urgency.
02
EHR data completely disconnected from Salesforce
Patient records lived in Epic. Salesforce knew nothing about appointment history, active care plans, or medication. Every agent interaction started with the agent asking the patient to repeat information that already existed somewhere else.
03
Post-discharge follow-up was entirely manual
Discharge follow-up calls were done by nurses from printed lists. 30-day readmission rates were above the national average. There was no systematic outreach — just whoever had time to make calls that week.
04
Referral tracking had no visibility
Referrals between departments and external specialists were tracked in spreadsheets. Referring physicians had no visibility on referral status. Patients had no idea where they were in the process. Cases fell through the gaps regularly.
05
Insurance pre-auth queries consumed 40% of agent time
Pre-authorisation status queries — largely routine, largely answerable from payer system data — were being handled manually by trained agents. 40% of daily contact volume. Almost none of it required human judgement.
06
HIPAA compliance made off-the-shelf AI unusable
Kalenix had looked at generic AI tools. All of them had data residency issues, couldn't be integrated with Epic, or couldn't satisfy their BAA requirements. The compliance bar ruled out everything they'd evaluated.
System Architecture

Four agents. One unified patient data layer.

Every Agentforce agent runs on the Health Cloud unified patient record — pulling live data from Epic via HL7 FHIR, payer APIs, and the referral management system.

DATA SOURCES UNIFIED PATIENT LAYER AGENTFORCE AGENTS ESCALATION + OUTCOMES Epic EHR Patient records HL7 FHIR API Payer APIs Insurance eligibility Pre-auth status · real-time Referral System Internal + specialist Status · routing · tracking Scheduling Appointment calendar Availability · bookings Omni-Channel Phone · Chat · Portal SMS · Email inbound Salesforce Health Cloud — Unified Patient Record 360° patient profile · Care plans · Appointment history · Insurance · Active referrals · Contact preferences HIPAA-compliant · HL7 FHIR integrated · BAA in place · Data residency: US-only Salesforce instance AGENT 01 Intake Triage Agent Classifies every inbound contact Routes by urgency + type AGENT 02 Scheduling Agent Books, reschedules, cancels Sends reminders · updates Epic AGENT 03 Referral Tracker Agent Status updates to patients Alerts on stalled referrals AGENT 04 Post-Discharge Agent Automated follow-up sequences Flags readmission risk to care team Human Agent Escalation Epic Appointment Write-back Service Cloud Case + CSAT Readmission Risk Dashboard
How an Agent Interaction Works

Patient contacts. Agent resolves. Human never involved.

From first contact to resolution — the full autonomous flow for a routine patient interaction, end-to-end.

01 Patient Contacts Phone · Chat · Portal · SMS Any channel, any time 02 Triage Agent Intent classified instantly Health Cloud record loaded 03 Specialist Agent Scheduling · Referral Pre-auth · Post-discharge 04 Action Executed Epic updated · booking made Confirmation sent to patient 05 — IF COMPLEX Human Escalation Full context transferred Agent never starts cold 06 Case Closed CSAT captured Health Cloud updated PATIENT CONTACTS 74% RESOLVED WITHOUT A HUMAN AGENT
How We Delivered It

16 weeks. What we actually built.

🏥
Health Cloud Foundation
Rebuilt the Salesforce org from a generic CRM into a proper Health Cloud implementation — unified patient records, care plans, care team hierarchies, and full HL7 FHIR integration with Epic.
🤖
Four Agentforce Agents
Designed, trained, and deployed four specialised AI agents — Intake Triage, Scheduling, Referral Tracker, and Post-Discharge. Each has distinct decision logic, escalation rules, and access permissions.
🔒
HIPAA Compliance Architecture
Designed the entire AI layer to operate within Salesforce's HIPAA-eligible infrastructure. BAA in place. PHI never leaves the Salesforce environment. Data residency enforced at the org level.
📡
Omni-Channel Routing
Connected phone, live chat, patient portal, SMS, and email into a single Service Cloud omni-channel queue. Agents receive contacts with full patient context loaded — no matter which channel the patient used.
What Changed

The numbers at 90 days post go-live.

74%
Patient interactions fully resolved by Agentforce — without a human agent involved
22→4
Average wait time in minutes — down from 22 to under 4 for routine contacts
31%
Reduction in 30-day readmission rate — attributed to the Post-Discharge Agent's follow-up sequences
89%
Patient satisfaction score at 90 days — up from 71% pre-implementation

"We'd looked at generic AI tools and ruled them all out — HIPAA compliance alone killed every option. Amroar built the Agentforce implementation inside Salesforce's own HIPAA-eligible environment. Our compliance team signed off in week three. By go-live, 74% of our patient contacts were being handled without a human. I didn't think that number was possible."

Chief Operating Officer
Kalenix Health · Patient Services Division
Work With Amroar

Agentforce in a regulated industry. We've done it.

Most Agentforce implementations stay simple because the partner doesn't understand compliance. We've built HIPAA-compliant, HL7-integrated AI agents in production. If your industry has compliance requirements, that's exactly the conversation to have with us.

Senior architects on every call. We'll tell you exactly what's possible in your environment.

50+
Agentic AI implementations delivered across enterprise clients
200+
Enterprise clients across US, UK, Australia and the Middle East
90+
Certified Salesforce developers — Agentforce, Health Cloud, Service Cloud
0
Failed builds. Across every enterprise engagement we've taken on.