My experience with optimization

Earlier in my career, I worked as a consultant leading operational, optimization, transformation, and implementation initiatives across a range of healthcare settings. Much of this work centered on improving patient access and strengthening clinic throughput through thoughtful template design. While industry best practices provided a strong foundation, the most effective recommendations always grew from a clear understanding of each clinic’s unique landscape: its patient demographics, provider and resource mix, and the specialty-specific scheduling and clinical needs that shape how a template performs day to day.

Through these months-long optimization projects, we often found that needs would change even during the short design process. The reasons were legion: a clinic had been acquired and folded into the system mid-year, a new clinic had opened up (and access needed to support a growing practice filled with new patients), a clinician retired, or a new group of residents began their program.

Assumptions vs. reality

The key assumptions underlying every engagement were that a hospital, once optimized, will stand still long enough for the work to matter, or that leaders supporting the optimization will have the same time to dedicate to continuing the work with the same thorough research and intention.

This is not the case.

These changes create a gap between what leaders think is happening, and what is actually happening in hospital hallways, patient rooms, and ERs. This created a barrier for leaders making informed decisions on staffing ratios, discharge processes, or resource allocation. In turn, this sows the seeds for inefficiencies, negatively impacting caregivers and patients.

While the causes varied, the result was the same: without ongoing intelligent monitoring, projects to optimize operations are too easily rendered obsolete, and no longer corresponded to daily reality. The root of the issue was not the quality of the work; instead, it was the assumption that a hospital, clinic, template, or workflow, once optimized, would remain static.

Conditions change constantly

Hospitals are dynamic organizations, and the data backs this up. A 2022 study published in the Journal of Hospital Medicine found that hospital physicians turn over at an average of 10.9% annually; across all staff categories, a 2026 NSI report found the turnover rate to be 20.7%. Throughout every specialty, a study from The Annals of Internal Medicine found that physician turnover increased by at least 35% between 2010 and 2018.

More frequent optimizations is not the solution

When leaders notice that optimizations become outdated, their first instinct may be to tighten the cycle, resorting to annual refreshes instead of biennial ones, standing retainers, or quarterly check-ins.

This instinct is understandable, but more frequent snapshots are still snapshots, and continue to capture static images of workflows, staff, and operations. More importantly, recurring updates are expensive and intensive; each one requires large teams to execute the work, and can still carry engagement fees that run into the hundreds of thousands of dollars. Continuously striving for “optimal,” yet consistently falling short of the ideal is a reflection of decisions being made in hindsight, rather than with real-time insight.

Why intelligence must be continuous

Currently, insights are treated as project outputs: work is completed, templates improved, and the engagement closes. Everyone moves on, even while the “optimized” template quickly ages.

In order to truly resolve the problem, hospitals need continuous, real-time intelligence that reflects on-the-ground realities. Instead of treating insights as project outputs, they should be seen as operational inputs: data that can be used to constantly generate recommendations to streamline operations and enhance care delivery.

That’s where agentic AI comes in. Rather than producing a correct answer to a question about conditions that existed during a period of time, AI generates recommendations continuously against current conditions, without requiring a new project to account for the clinic acquired last quarter or the cohort of physicians who joined last month.

With Access Agent, clinic administrators can see which rooms remain idle between visits; providers get templates that reflect their reality; and leaders can identify opportunities to improve scheduling and utilization without adding space or staff. That is what continuous intelligence looks like in practice: a system that makes the question of deliverable obsolescence irrelevant.

Rethinking the approach to hospitals

To be clear, AI is not replacing consultant teams, just adding a real-time dimension to hospital operations. Unlike consultant engagements which are reactive, AI will not expire and instead, will move and change with your hospital.

If the answer to the second question involves scheduling another engagement with a consultant, then leaders have only deferred their problem rather than solving it.