There is a conversation happening in every health system boardroom right now, and it usually goes something like this: "We need AI." The executives nod. Someone mentions a clinical decision support tool. A pilot gets approved. Six months later, the pilot has stalled, the vendor is asking for more data, and nothing about the organization's day-to-day operations has actually changed.
Meanwhile, the real bottleneck — the one burning through millions of dollars in labor costs every year — has nothing to do with clinical decision-making. It is the operational infrastructure: the onboarding workflows, the authorization requests, the PDF generation, the email routing, the credential verification. The work that nobody puts on a conference slide but everyone does for eight hours a day.
This article presents two verified case studies of healthcare organizations that took a different approach. They did not start with clinical AI. They started with operational infrastructure modernization — and the results were measurable, significant, and replicable. I will walk through exactly what they automated, what it cost, and what it produced. Every number in this article is sourced and cited.
The Hidden Cost of "We've Always Done It This Way"
Before we get into the case studies, I need to make a distinction that most automation vendors deliberately blur. There are three fundamentally different categories of technology investment in healthcare, and conflating them leads to poor decisions:
Clinical AI refers to tools that support clinical decision-making — diagnostic imaging analysis, predictive patient risk models, AI-assisted documentation. These tools operate within the care delivery workflow and directly affect clinical outcomes. They are high-value, but they are also high-complexity, heavily regulated, and slow to deploy.
Process automation refers to robotic process automation (RPA), workflow bots, and rule-based systems that handle discrete, repetitive tasks — extracting data from one system and entering it into another, routing emails, generating standard documents. These are tactical tools. They solve specific problems. But they are not a strategy.
Operational infrastructure modernization is the strategic layer above both. It means systematically auditing, redesigning, and automating the end-to-end operational workflows that run a health system — from hire to retire, from referral to reimbursement. It combines process automation with task mining, intelligent document processing, computer vision, and API integration into a coordinated program that transforms how the organization operates. This is what Alberta Health Services did. This is what produces the largest returns.
The distinction matters because most organizations skip the strategic layer. They buy an RPA tool, automate two or three processes, declare victory, and wonder why the overall operational burden has barely moved. Automating individual tasks without redesigning the underlying workflow is like putting a turbo on a car with no transmission — the engine revs, but you do not go anywhere.
Consider the baseline. According to the American Medical Association, physicians spend approximately 5.8 hours of an 8-hour workday on electronic health record tasks (AMA, ama-assn.org). Industry estimates consistently place administrative costs at roughly 24% of total healthcare spending. McKinsey's 2025 report on agentic AI in revenue cycle management estimates that health systems spend more than $140 billion per year on revenue cycle management alone (McKinsey, "Agentic AI: The race to a touchless revenue cycle," 2025). For a deeper analysis of the scale of this administrative burden, see our breakdown of the $258 billion healthcare admin crisis. Those EHR hours directly reflect the quality of the underlying healthcare information management system — better HIM infrastructure means less manual navigation and rework.
These are not minor inefficiencies. This is structural waste embedded in the operational fabric of every health system. And the organizations that have addressed it systematically — not with a single bot, but with a comprehensive modernization program — have produced results that are difficult to ignore.
Let me show you two of them.
200 Work-Years Saved: The Alberta Health Services Story
Alberta Health Services (AHS) is Canada's first and largest province-wide, fully integrated health system. It operates 100 hospitals with 34,000 beds, employs approximately 10,000 managers, processes 17,000 new hires per year, manages 47,000 staff transfers annually, and routes 1.5 million emails per day. This is not a pilot site. This is an enterprise-scale operation — and the operational infrastructure that supports it was, until recently, overwhelmingly manual.
The details of this case come from a Public Sector Network case study published in March 2025, authored by Heather Dailey (publicsectornetwork.com).
AHS launched an Intelligent Automation program that ultimately automated 47 distinct processes across the organization. This was not a single RPA deployment. The technology stack included robotic process automation, task mining, chatbots, computer vision, and Azure AI APIs — selected and combined based on the specific requirements of each workflow. The program was led by Jesse Tutt, Program Director of Intelligent Automation.
What They Automated (and What They Didn't)
The 47 automated processes span IT, HR, finance, and clinical administration. Key categories include:
- HR onboarding and transfers: 17,000 new hires per year and 47,000 staff transfers per year now flow through automated workflows that handle credential verification, system access provisioning, orientation scheduling, and documentation generation.
- IT service desk operations: Chatbots and RPA handle tier-one requests — password resets, device provisioning, access requests — freeing IT staff for complex infrastructure work.
- Email processing: 1.5 million emails per day are routed, categorized, and actioned through automated classification and response systems.
- Clinical data documentation: One of the highest-value automations involved PDF generation for clinical data. This single workflow — automating the creation, formatting, and distribution of clinical PDFs — saved $1.3 million.
What they did not automate is equally instructive. Clinical decision-making, patient-facing interactions, and complex exception handling remained with human staff. The automation program focused exclusively on the operational infrastructure — the high-volume, rule-based, repetitive work that consumed administrative capacity without requiring clinical judgment.
The Surprising Quick Wins
The $1.3 million saved on PDF generation is worth pausing on. This was not a glamorous process. Nobody would put "PDF generation" on a strategic roadmap slide. But across a system of 100 hospitals, clinical data PDFs were being generated, formatted, reviewed, and distributed manually — thousands of times per day. The cumulative labor cost was enormous, and the process was almost entirely rule-based. Automating it was technically straightforward and produced one of the program's largest single returns.
This is a pattern I see repeatedly: the highest-ROI automation targets are often the most overlooked. They are not the processes that executives discuss in strategy meetings. They are the processes that managers quietly absorb into their teams' daily workload because "that's just how we do it."
The cumulative result of the 47 automated processes: 200 work-years of manual labor eliminated. That is approximately 414,000 hours of human effort per year that no longer needs to happen. Those hours were redistributed to quality improvement initiatives, direct patient support, and other activities that require human judgment and empathy.
"Intelligent Automation has enabled us to streamline time-consuming administrative tasks, reducing manual work and freeing up healthcare staff to focus on delivering quality care." — Jesse Tutt, Program Director, Intelligent Automation, Alberta Health Services (Public Sector Network, March 2025)
50% Less Manual Work: How USA Health Doubled Authorizations
If the Alberta case demonstrates what operational infrastructure modernization looks like at enterprise scale, the USA Health case demonstrates what targeted process automation can achieve for a specific revenue cycle bottleneck — prior authorizations.
USA Health is an academic health system based in Mobile, Alabama. Their prior authorization workflow was a classic manual bottleneck: staff called payers, waited on hold, submitted faxes, tracked responses in spreadsheets, and re-submitted when information was missing. It was slow, error-prone, and completely unscalable.
The details of this case come from an Experian Health blog post published in October 2025 (experian.com/blogs/healthcare).
USA Health implemented the Experian Health prior authorization platform to automate their authorization workflow. The implementation was phased over 6 to 8 months, expanding from a single service line to six service lines without adding any new staff.
Before vs After: The Numbers
The transformation is best understood through direct comparison:
- Daily authorization volume: Before automation, each employee processed approximately 20 accounts per day. After: 40 to 50 accounts per employee per day — a 100% increase in throughput.
- Manual work reduction: 50% reduction in manual effort, driven primarily by automated status checks against payer portals. More than 50% of accounts now receive instant automated status responses without human intervention.
- Annual volume: The 28-person authorization staff manages 130,000 authorization requests per year.
- Service line expansion: The team expanded from 1 to 6 service lines — a 6x increase in operational scope — without hiring a single additional employee.
- Task organization: The platform uses 30 dynamic work queues to route tasks by priority, payer, and service line, ensuring staff work on the most time-sensitive authorizations first.
Why No New Hires Were Needed
This is the detail that matters most for healthcare leaders evaluating automation. USA Health did not reduce headcount. They did not lay anyone off. What they did was absorb a 6x increase in operational scope with the same team — because automation removed the repetitive, low-value tasks (phone calls to check status, manual data re-entry, duplicate submissions) that had consumed most of each employee's day.
The 28 staff members who previously spent their time on hold with payers now manage exceptions, handle complex cases, and ensure that the automated workflows are running correctly. Their roles shifted from data entry to oversight and problem-solving. That is not just an efficiency gain — it is a workforce transformation.
"We knew we needed to transform our authorization workflow processes." — Amy Grissett, Senior Director of Ambulatory Revenue Cycle, USA Health (Experian Health blog, October 2025)
Wondering how many hours your team spends on authorizations each week? We offer a free 30-minute operational process audit for healthcare practices. Book yours here.
The 5 Processes Every Clinic Should Automate First
The Alberta and USA Health cases involved large organizations with significant budgets. But the underlying principle applies at every scale: identify the processes that are repetitive, rule-based, and high-volume, then automate them in order of impact. Based on what I have seen work across healthcare organizations of all sizes, here are the five processes that consistently produce the fastest returns.
1. Insurance Eligibility Verification
The problem: Front-desk staff manually check patient insurance eligibility before appointments — often by logging into payer portals one at a time, or worse, calling payer phone lines. For a practice seeing 30 patients per day, this process can consume 2 to 3 hours of staff time daily.
The automation approach: Automated eligibility verification tools batch-check all scheduled patients against payer databases 24 to 48 hours before their appointments. Discrepancies are flagged for manual review; confirmed patients flow through without human touch.
Expected result: 70% to 90% reduction in manual eligibility checks. Fewer day-of cancellations due to coverage issues. Faster patient check-in. USA Health's experience with automated payer portal checks — where more than 50% of accounts receive instant status responses — illustrates the principle at smaller scale.
2. Prior Authorizations
The problem: Prior authorization is the single most cited administrative burden in outpatient care. Staff spend hours per day submitting, tracking, and resubmitting authorization requests. Missing or delayed authorizations lead to claim denials, delayed care, and patient frustration. For a deeper look at how authorization delays cascade into broader administrative waste, see our analysis of the healthcare admin crisis.
The automation approach: Platforms like the one USA Health implemented automate submission, status tracking, and follow-up. Dynamic work queues prioritize tasks by urgency. Automated status checks eliminate the need for staff to manually call or log into payer portals.
Expected result: 50% reduction in manual work, consistent with USA Health's documented outcomes. Throughput doubles without adding headcount. Authorization turnaround times drop from days to hours for straightforward cases.
3. Scheduling, Reminders, and Confirmations
The problem: No-show rates in outpatient practices typically range from 15% to 30%. Each no-show represents lost revenue and wasted clinical capacity. Manual reminder calls are time-consuming and inconsistent. For practices struggling with no-show rates, we have published a detailed guide on reducing patient no-shows with automation.
The automation approach: Automated multi-channel reminders (SMS, email, voice) sent at optimized intervals — typically 72 hours, 24 hours, and 2 hours before the appointment. Self-service rescheduling links allow patients to move appointments without calling the office. Waitlist automation fills cancellation slots automatically.
Expected result: 20% to 40% reduction in no-show rates. Significant reduction in front-desk phone volume. Improved patient satisfaction from convenient self-service options.
4. Claim Submission and Denial Follow-Up
The problem: Claims submitted with errors get denied. Denied claims require investigation, correction, and resubmission — a process that costs $25 to $118 per claim according to industry benchmarks. For the average practice, 5% to 10% of claims are denied on first submission, and a significant percentage of denied claims are never resubmitted. That is revenue walking out the door. For billing-specific automation strategies, see our guide on billing automation for small practices.
The automation approach: Automated claim scrubbing catches errors before submission — missing modifiers, mismatched diagnosis codes, incomplete demographic fields. Denial management workflows automatically categorize denials by reason code, generate corrected claims, and resubmit with minimal human intervention.
Expected result: First-pass claim acceptance rates improve from 85%-90% to 95%+. Denial follow-up time drops significantly. Recovered revenue from previously unworked denials adds directly to the bottom line.
5. Staff Onboarding and Credentialing
The problem: Healthcare onboarding involves license verification, background checks, credential validation, system access provisioning, compliance training tracking, and dozens of forms. Alberta Health Services processes 17,000 new hires and 47,000 transfers per year — and they found this was one of their highest-value automation targets.
The automation approach: Workflow automation handles document collection, verification against licensing databases, automated system access provisioning, training assignment, and status tracking. Managers receive dashboards showing onboarding progress rather than chasing individual tasks via email.
Expected result: Onboarding time reduced from weeks to days. Compliance gaps eliminated through automated tracking. HR staff freed from data entry to focus on employee experience and retention — critical in an industry with chronic staffing shortages.
How to Start: From Audit to Automation in 8 Weeks
Knowing what to automate is the first step. Knowing how to execute is what separates organizations that capture value from those that accumulate unused software licenses. Here is an 8-week framework based on what has worked in practice — including a reference timeline from Cleveland Clinic, which deployed production-ready RPA automations in 8 weeks with a $700,000 ROI over 3 years (Huron Consulting, huronconsultinggroup.com), and achieved 80% faster claim processing.
Weeks 1-2: Process Audit
Before automating anything, you need to understand your current state with precision. This means task mining — documenting every step in a workflow, who performs it, how long it takes, where errors occur, and how frequently the process runs. Time mapping captures actual hours spent, not estimates (which are always optimistic).
The deliverable from this phase is a process inventory: every administrative workflow in your organization, ranked by volume, time consumption, and error rate. Most organizations discover 3 to 5 processes they did not even realize were consuming significant labor — the "PDF generation" effect from the Alberta case.
If you do not have the internal bandwidth for this, our team can run the audit for you. See our services for details on operational process audits.
Weeks 3-4: Prioritization
Not every process should be automated, and not every automatable process should be automated first. Use an impact-versus-complexity matrix to rank your process inventory:
- High impact, low complexity: Automate first. These are your quick wins — the equivalent of Alberta's PDF generation or USA Health's automated status checks.
- High impact, high complexity: Plan for phase two. These require more integration work but produce significant returns.
- Low impact, low complexity: Automate opportunistically. Nice to have, but do not prioritize over higher-impact targets.
- Low impact, high complexity: Skip entirely. The effort-to-return ratio does not justify the investment.
Weeks 5-6: First Workflow Implementation
Deploy your first quick-win automation. This should be a process from the "high impact, low complexity" quadrant — one that produces measurable results within days of going live. The goal is not perfection. The goal is a working automation that demonstrates value to stakeholders and builds organizational confidence for subsequent deployments.
Cleveland Clinic's 8-week timeline included production deployment within this window, validating that meaningful automation can go live in under two months when the process is well-scoped and the team is aligned.
Weeks 7-8: Measurement and Expansion
Measure the first automation rigorously: time saved, error reduction, throughput improvement, cost impact. Compare against your baseline from the audit phase. Use these validated results to build the business case for the next wave of automations.
This is also where you plan phase two — the next 3 to 5 processes on your prioritized list. Each subsequent automation builds on the infrastructure and organizational knowledge from the previous one, accelerating deployment cycles.
What This Looks Like in Practice: A Simulation
Note: The following is a hypothetical projection, not a case study. It is provided to illustrate the methodology, not to represent actual results.
Consider a 5-physician outpatient clinic seeing 30 patients per day. Current state: the authorization team spends approximately 12 hours per week on prior authorization tasks — submitting requests, checking status, following up on denials, and resubmitting. Based on USA Health's documented 50% reduction in manual work, automation could reduce this to approximately 6 hours per week — a savings of 6 hours per week, or 312 hours per year.
At a fully loaded administrative staff cost of $25 per hour (salary, benefits, overhead), that translates to approximately $7,800 per year per employee performing authorization work. For a team of 2 to 3 authorization staff, the annualized savings would range from $15,600 to $23,400. This does not account for the revenue acceleration from faster authorization turnaround, which could be substantially larger depending on payer mix and service volume.
These figures are illustrative. Actual results depend on current process efficiency, payer mix, technology selection, and implementation quality. The point is the methodology: measure your baseline, apply documented improvement ratios conservatively, and calculate the financial case before committing to a specific tool or vendor.
Need help identifying which processes to automate first? Our team maps your workflows and delivers a prioritized automation roadmap — no commitment. Get started.
The Real Cost of Waiting Another Year
I want to close with a number that should concern every healthcare executive who is still in "wait and see" mode. According to McKinsey's 2025 analysis, more than 30% of healthcare providers have already prioritized AI and automation for their revenue cycle operations (McKinsey, "Agentic AI: The race to a touchless revenue cycle," 2025). That is not a forecast. That is a current state. Nearly a third of your market is already moving.
The Forrester Total Economic Impact study commissioned by Pega reported 186% ROI over three years for RPA implementations (Pega/Forrester TEI study). Important caveat: that figure is based on cross-industry data and is not healthcare-specific. Healthcare ROI may differ due to regulatory complexity, integration requirements, and workflow variability. However, the Alberta and USA Health cases — which are healthcare-specific — suggest that the directional magnitude is credible for well-executed programs.
The cost of waiting is not abstract. It is the number of hours your staff will spend this year on tasks that could be automated, multiplied by the fully loaded cost of those hours. For Alberta Health Services, that number was 200 work-years — approximately 414,000 hours — per year. For USA Health, it was the difference between processing 20 authorizations per employee per day and processing 50.
Every month you delay, you are paying the full cost of manual operations while your competitors are paying a fraction of it. That gap compounds. Organizations that automate first build institutional knowledge, optimized workflows, and operational flexibility that late adopters must pay more to replicate.
The thesis of this article is simple: the biggest gains in healthcare do not come only from clinical AI. They come from modernizing operational infrastructure. The case studies prove it. The question is no longer whether operational infrastructure modernization works — it is whether your organization will capture those gains this year or continue absorbing the cost of manual operations while the rest of the industry moves forward.
If you are ready to find out what operational infrastructure modernization could mean for your organization, we are ready to show you. It starts with a process audit — not a sales pitch. Let us start the conversation.
