Nobody needs AI to get started

2026

Sometimes all it takes is stopping doing by hand what a machine does better. And in the most challenging cases, hyperautomation opens up a new level of possibilities: most operational gains in healthcare don’t depend on artificial intelligence, but rather on automating what is still done manually.

In previous articles we’ve seen how healthcare wastes operational resources and how poorly implemented technology can exacerbate the problem. Today I want to talk about something different: simple, concrete automation with immediate impact, and how, in more complex processes, the integration of automation and artificial intelligence opens a new level of transformation.

There’s a very common misconception when it comes to automation in healthcare. People immediately imagine artificial intelligence, surgical robots, diagnostic algorithms—impressive, expensive things that are far removed from the reality of most organizations.

But the automation that has the greatest impact on the day-to-day operations of a hospital or healthcare unit is none of these. It is much simpler, more precise, and often forgotten:

• Appointment reminders are still being made via manual phone calls.
• Prior authorizations continue to be processed via email and even fax.
• The management reports that someone compiles in Excel every Friday during two hours.
• Appointments that depend on a phone call between two services could communicate automatically.

The invisible cost of repetitive manual labor.

It is estimated that 77% of healthcare professionals waste 45 minutes a day on inefficient workflows, equivalent to four weeks lost per person, per year. Not on complex work, but on repetitive tasks that could be automated.

Administrative costs represent about 25% of all healthcare spending, according to the American Medical Association. And approximately 24% of a hospital’s working budget is concentrated in administrative functions such as: processing emails, making calls, managing queues, updating records, and transcribing data between files and applications.

According to the 2024 CAQH Index, the healthcare sector spends $83 billion annually on staff time conducting routine administrative transactions between providers and funders, with providers bearing 97% of that cost.

The same report estimates that the sector has the potential to save more than $20 billion by migrating to automated workflows.

This is not a problem unique to the United States. It is a universal pattern that is replicated in any healthcare system that has not questioned its administrative processes in recent years.

Sources: CMS (ref. Intuition Labs AI Adoption Report, 2025) / AMA (ref. TopflightApps, 2023) / CAQH Index 2024 (ref. Notable Health, 2024)

Benefits of simple automation in healthcare

Portsmouth Hospitals in the UK have automated maternity appointment scheduling. The system now automatically suggests upcoming appointments based on the expected delivery date and necessary tests, with appointments confirmed within 24 to 48 hours instead of weeks, and automatic reminders to reduce no-shows.

The result? More patients treated, better continuity of care, less overburdened teams.

It wasn’t Artificial Intelligence (AI) automating a process that was still being done manually out of inertia.

And the financial return is real. Hospitals report a return of $3.20 for every dollar invested in workflow automation, with typical returns realized in just 14 months.

In 2024, 60% of organizations adopted automation solutions with an AI component, registering an average efficiency increase of 35%. A patient automation solution based on Natural Language Processing (NLP) saved 20 hours of manual work per day in a single healthcare organization.

Sources: Kissflow Healthcare AI Workflow ROI Report (2025) / Newo.ai, The Future of RPA: Key Trends and Growth in 2025

Because automation in healthcare frees up professionals, it doesn’t replace them.

There is a legitimate concern surrounding automation in healthcare: the fear that it will replace people. This concern is, to a large extent, misapplied, at least in this context. Automation is not replacement, it is liberation.

Automating appointment scheduling doesn’t replace professionals. It frees them up to do what only they can do. Automating a weekly report doesn’t eliminate a manager. It gives them back hours to analyze, decide, and improve.

Automation can reduce administrative tasks by up to 30%, helping professionals dedicate more time to patients. And this restored time has a value that goes far beyond the financial aspect; it has a direct impact on the quality of care and the well-being of those who provide it.

The question isn’t ‘should we automate?’. The question is ‘where do we start?’. The answer is almost always the same: with the most repetitive, most manual process, the one that generates the most silent frustration within teams.

Source: Master of Code. AI in Healthcare Statistics (2025)

Hyperautomation in healthcare: when and why to apply it.

However, there are processes in healthcare that traditional automation cannot solve on its own. Processes involving unstructured data, handwritten clinical notes, natural language reports, heterogeneous documentation from multiple systems, paper or telephone consultation requests. Processes that require context interpretation, reasoning about incomplete information, or adaptive decision-making.

This is where Hyperautomation comes in, a concept created by Gartner to describe the orchestration of multiple advanced technologies: RPA (Robotic Process Automation), Artificial Intelligence, Machine Learning, NLP (Natural Language Processing), and process mining, working together to automate end-to-end processes that were previously inaccessible to conventional automation.

A concrete example is the prior authorization process, one of the biggest administrative bottlenecks in healthcare. According to AMA data, doctors and their teams spend an average of 13 hours per week processing around 39 authorization requests per doctor.

93% of doctors report that this process causes delays in care. 94% say it harms clinical outcomes. 40% of clinical practices have staff dedicated exclusively to this task.

With Hyperautomation, NLP systems analyze clinical notes, histories, and exam reports in real time; ML algorithms automatically cross-reference with funder policies; and RPA bots submit requests to the appropriate portals, reducing approval time from days to minutes.

Pilot programs using NLP to interpret clinical documentation and automatically generate authorization requests have demonstrated not only a reduction in processing time, but also a decrease in errors and rejections.

The same principle applies to outpatient clinical documentation, where AI-based systems automatically transcribe consultations and generate structured clinical notes, giving doctors hours of time per week. Or to the management of surgical material inventory, where predictive algorithms anticipate needs and trigger orders without human intervention.

Hyperautomation isn’t for every process, nor should it be the starting point. Gartner recommends a progressive approach: begin with the simplest and most structured tasks using classic RPA, consolidate the gains, and only then move on to integrating AI into more complex and high-impact processes.

Sources: IDC Health Insights (2025). The US Healthcare Prior Authorization Crisis. / AMA Prior Authorization Physician Survey (2024) / Quazi & Raju (2026). HyperAutomation in Healthcare: Transforming Operations Through AI, RPA, and Intelligent Workflows. Springer. DOI: 10.1007/978-3-032-02853-2_33 / Gartner Strategic Automation Decision Framework (ref. Nividous, 2024)

A practical perspective for healthcare organizations.

The path is not binary; it’s not ‘simple automation’ or ‘hyper-automation’. It’s a logical progression that begins where the impact is most immediate and the risk is lowest.

First , identify the most repetitive, manual processes that generate the most friction for both teams and patients. These are natural candidates for initial automation using RPA or low-code tools. In other words, the first step will be to map these processes, measure them, and listen to professionals and users.

Second , develop automation that addresses the potential efficiency gains identified in the previous point.

Third , measure the impact. How many hours were freed up? How many errors were eliminated? What is the processing time before and after? Without measurement, there is no learning and no argument for the next investment.

Fourth , in more complex processes, where there is unstructured documentation, contextual decisions, or integration of multiple systems, explore the possibilities of Hyperautomation with a well-defined pilot project.

The operational transformation of healthcare will not happen all at once. It will happen process by process, team by team, organization by organization. And each step taken is a step that frees up human time for what truly matters.

Don’t miss the opportunity for a small improvement today, while waiting for the big improvement that may come later!

In the next article: why a culture of continuous improvement doesn’t originate from processes, but from people, and the type of leadership that involves them.

If you’d like to assess which processes in your organization have the greatest potential for automation, we can help.