This article, written by Meena Mallipeddi, first appeared in Becker’s Hospital Review.
Hospitals have more digital tools than ever, yet throughput continues to suffer. Since 2017, emergency department encounters lasting over eight hours have nearly doubled, inpatient units are strained by staffing shortages, and specialty consults remain difficult to coordinate.
The industry has turned to AI to address these challenges. In many cases, AI delivers meaningful value, especially when data is structured and workflows are well defined.
However, many throughput delays stem from a more fundamental issue: critical work still remains manual, and systems remain disconnected across people and processes. Gaps between technology solutions, handoffs between teams, and variability in daily workflows introduce friction that slows care. Studies consistently show that these transitions and communication failures account for a significant share of throughput delays.
AI can only drive improvements when data is structured and visible—but most hospital bottlenecks occur in disconnected systems and outside of structured workflows.
Why AI Struggles in Unstructured Environments
Consider a typical TeleStroke activation. In many health systems, a consult request is coordinated through a series of manual handoffs across multiple systems before a neurologist can even begin clinical assessment.
An ED clinician calls a triage or call center, where a coordinator uses a scheduling system to identify the on-call neurologist. That provider is paged and must return the call. On average, it takes five minutes just to reach the neurologist through this process.
The call center then connects the neurologist to the ED, if they aren’t able to connect directly themselves. The neurologist logs into the PACS and EHR systems to review imaging and clinical data, but critical information is also relayed manually: there is no automated data flow, so typically the neurologist must rely on a verbal summary from the ED attending. Too much time is spent on manual handoffs and information sharing before clinical decision-making can even begin.
For a stroke patient, every minute matters, but when information moves through such manual, disconnected workflows, AI cannot make a meaningful difference.
By contrast, when a TeleStroke platform orchestrates the entire encounter, automation replaces manual coordination across the workflow—even as ED teams remain native in their EHR:
- No call center intermediary
- No paging delays
- Automatic escalations
- No separate PACS or EHR logins for the consulting neurologist
- One continuous, end-to-end workflow across systems
Because the encounter is orchestrated as a single workflow, every step is automatically captured, timestamped, and structured so AI can effectively use the data.
At AmplifyMD, we’ve seen this orchestration translate into:
- 52% reduction in program operational costs
- 35% reduction in ED team time per TeleStroke encounter
These results are part of a broader shift toward tech-enabled care models that optimize patient throughput by synchronizing workflows across the entire care continuum.
What Orchestration Changes
What the TeleStroke example illustrates is a broader reality across hospital operations: while clinical care has been digitized, most organizations still lack an operational layer that allows systems and teams to function as a coordinated whole. Orchestration is needed to manage how work moves across people, systems, and time.
Critically, this is not what the EHR was designed to do. EHRs document care; they do not manage the flow of work. An orchestration layer unifies steps scattered across systems, assembles the information needed for decisions, routes requests based on real-time criteria, and automates tasks that don’t require human judgment.
High-stakes sectors like aviation, logistics, and energy have long relied on orchestration to coordinate complex operations at scale. Healthcare is now reaching that same inflection point.
Once workflows are orchestrated, AI becomes a true multiplier. In this environment, AI can forecast demand spikes, prioritize consults by acuity, and surface bottlenecks that remain invisible in fragmented workflows. AI shifts from a standalone tool to an operational engine—but only after work moves through a reliable, repeatable process.
Conclusion: Throughput Is an Orchestration Challenge
Hospitals have spent two decades digitizing clinical information. The next gains in throughput will not come from faster decisions alone, but from synchronizing the people, data, and processes that move patients through care.
AI will play a transformative role in hospital operations, but throughput improves only when an orchestration layer is in place. The organizations that succeed will be those that build the operational foundation that allows AI, clinicians, and data to work in concert. With that foundation, hospitals can finally unlock the throughput gains they have been chasing for years.
See how AmplifyMD’s virtual care platform orchestrates the entire encounter end-to-end to solve the throughput crisis.
Meena Mallipeddi
Meena Mallipeddi is the Co-founder and CEO of AmplifyMD, where she champions the transformative potential of virtual care to enhance health outcomes, particularly in underserved communities. She also serves on the American Telemedicine Association’s Leadership Council. She was recently named to Inc. Magazine’s Female Founders 250 list and Healthcare Technology Report’s Top 50 Women Leaders in Healthcare Technology list.
