
Share
Before any algorithm touches a patient's care, Allina Health asks whether it improves safety, access, clinician support or experience. The 12-hospital system's COO explains how that discipline shapes real deployments, from surgery scheduling to discharge planning.
A surgeon waiting on a delayed operating room. A nurse scrambling to coordinate a discharge before a bed is needed elsewhere. A patient lying awake, unsure why housekeeping hasn't come to clean a room they vacated hours ago. These are the small, grinding frictions of hospital life, the kind most patients never think about until they're caught inside one. They're also exactly where Allina Health has decided artificial intelligence should earn its keep.
Allina Health runs 12 hospitals across Minnesota and Western Wisconsin. Like most health systems right now, it's fielding pitches from vendors promising that AI will fix everything from staffing shortages to burnout to bed turnover. The harder work isn't finding AI tools. It's deciding which problems actually deserve one.
"Does new technology improve quality and safety, make care more accessible, support the clinicians, or create a better patient experience?" said Dominica Tallarico, Allina Health's chief operating officer, describing the four questions her team asks before green-lighting any new system. If a proposed tool can't clearly answer yes to at least one of those, it doesn't move forward, regardless of how impressive the technology looks in a demo.
That discipline matters because hospitals are complicated ecosystems, not single workflows. A tool that saves time for one department can quietly create new burdens for another. Tallarico's job, as she describes it, is to make sure technology connects people, processes and care settings without one team's efficiency becoming another team's headache. "Enterprise alignment and local flexibility are not competing priorities," she said. "When we get the balance right, they reinforce each other."
That philosophy shows up in surprisingly small places. Every recovery room at Allina Health now has a sensor that alerts the environmental services team the moment a patient leaves. It sounds minor, almost mundane, but think of it like a smart doorbell for hospital rooms: it means housekeeping isn't guessing when to clean, and patients aren't woken by a knock while they're still resting. "It reduces intrusions when a patient is sleeping and optimizes when clean-up can be done," Tallarico explained. Allina is rolling this out in phases, using early lessons to refine each next step rather than flipping a switch system-wide.
Surgical scheduling is where Allina has seen some of its clearest returns, and the numbers are worth sitting with. A cancelled surgery doesn't just inconvenience a patient. It costs the hospital between $3,350 and $8,600 once you account for administrative rework, lost operating room time and idle surgeon hours, according to industry benchmarking cited by Allina. Multiply that across a 12-hospital system and the financial and human toll adds up fast, not to mention the toll on a patient whose procedure gets pushed back.
To tackle that, Allina partnered with Qventus, a company that builds AI and machine learning tools to automate hospital operations. The relationship isn't a simple vendor-client arrangement where Allina buys a finished product off the shelf. Instead, Tallarico describes it as collaborative product testing in real time: Allina brings forward operational pain points, Qventus builds early-stage tools around them, and Allina's teams pressure-test whether those tools actually hold up under real hospital conditions.

The results so far include 689 hours of unlocked robotics capacity and an annualized return on investment nearly six times the initial outlay, using Qventus's Surgical Growth Solution. Those freed-up hours translate directly into more patients getting access to care they need. Allina has set further goals: cutting surgery cancellations by 72 hours of lead time and improving estimates of same-day discharge timing, both aimed at keeping the surgical pipeline moving smoothly instead of stalling at unpredictable points.
Discharge planning is the newer frontier. Allina is in early stages of auto-populating discharge plans, building better sequencing queues so patients move through the system with less friction. Early feedback has been encouraging. Discharge times are moving earlier in the day, patient mobility is improving, and the broader trend across the organization shows length of stay ticking downward. None of those are dramatic headline numbers on their own, but together they represent real days returned to patients' lives and real capacity returned to hospital beds.
Underlying all of it is a clear philosophical line Tallarico insists on. AI, in her view, isn't a layer bolted onto existing software or the electronic health record. It has to be treated as foundational, something designed around the actual bottlenecks clinicians face rather than retrofitted onto old systems. "We ask our clinical teams questions, like what is contributing to cognitive load and where are the workflow bottlenecks?" she said. That question, more than any flashy capability, is what determines whether a tool gets built at all.
It also shapes which vendors Allina chooses to work with. "We seek partners who are devoted to healthcare and want to listen, learn and understand the core challenges that constrict care delivery," Tallarico said. Good technology alone isn't enough, in her telling. The teams behind it need to be willing to dig into messy operational details, and they need to treat data security and patient protection as non-negotiable, not an afterthought bolted on at launch.
What makes Allina's approach notable isn't the technology itself. Sensors that track empty rooms and software that predicts discharge timing aren't exotic ideas in 2026. What stands out is the restraint behind the strategy: a clear refusal to let AI creep into clinical judgment territory. "We do not see technology as a replacement to clinical acumen," Tallarico said, "and we are actively bolstering our systems to ensure we firewall automation that is unnecessary or overriding human judgement."
That distinction, between AI as an operational aid and AI as a decision-maker, is likely to define which health systems earn patient and clinician trust over the next several years. Administrative tasks like intake, scheduling and discharge coordination are, as Tallarico puts it, desperately in need of refinement across the entire industry. Fixing those quietly, without touching the sacred space of clinical decision-making, may turn out to be AI's most durable contribution to healthcare, even if it never makes for the flashiest pitch.
Tags
Original Sources
How Allina Health decides where hospital AI can help
↗ https://www.healthcareitnews.com/news/how-allina-health-decides-where-hospital-ai-can-help
About the author
Amara's entry point into AI was an epidemiology role at a London research hospital, where she spent five years studying how digital health tools reached — or conspicuously failed to reach — underserved communities. Watching early algorithmic systems in healthcare quietly entrench existing inequalities, she redirected her career toward the systemic consequences of AI at scale. She covers AI through an unflinching lens: who benefits, who bears the cost, and what evidence actually says versus what the press release claims. Her writing is calm and precise, but she doesn't mistake balance for neutrality.
More from The Steward →This Week's Edition
11 November 2026
14 articles
Related Articles

Tether Evo's New Research Tackles BCI's Calibration Problem with Cross-Subject Models
Models & Research · 5 min

Why Some Engineers Think the Ocean Floor Is the Future of Data Centers
Environment & Climate · 5 min

Europe's Digital ID Push Tests a New Privacy Tool: Zero-Knowledge Proofs
Policy & Regulation · 5 min
Related Articles

Tether Evo's New Research Tackles BCI's Calibration Problem with Cross-Subject Models
Models & Research · 5 min

Why Some Engineers Think the Ocean Floor Is the Future of Data Centers
Environment & Climate · 5 min

Europe's Digital ID Push Tests a New Privacy Tool: Zero-Knowledge Proofs
Policy & Regulation · 5 min
More Stories
© 2026 Cedar & Bloom. All rights reserved.