The landscape of medicine is constantly evolving. Not so long ago were the days of paper charts, tape recorders and long treks through the hospital to find where exactly that one test landed (Figure 1). Advancements in technology have ushered in a new era, one marked by electronic databases and multiple metrics of clinical efficiency. Young physicians cannot help but be grateful to train in these digital times, where the electronic medical record (EMR) provides extensive clinical histories, direct access to patient requests and immediate study results, all at the click of a button.

Figure 1: A shift in the medical landscape. From paper charts to EMR, physicians are spending more time facing their computer than their patients. Image originally created by author, edited and finalized by Gemini AI.

Young physicians also – in some ways – cannot help but regret training in the digital era. While the EMR provides extensive clinical histories, direct access to patient requests and immediate study results, it is an unfortunate reality that the demand to always be available has never been more apparent, and yet the focus has never been farther from the patient.

Medicine has become a balancing act, one where providers find that the scales are often tipped against them. Time once reserved for patient care is siphoned away for cumbersome, often unnecessary administrative burdens. Physicians are expected to juggle corporate tasks alongside wholesome treatment without the training or protected hours to do so. What has been perpetuated is this mentality of "do more, with less." Thankfully, new tools exist that may even the scales for health care professionals and return power into the hands of the physician.

Artificial intelligence (AI) offers an escape from the administrative burden driving provider burnout. Ambient scribes are widely studied, with numerous programs in development and implementation across the nation (Figure 2).¹ By passively listening to patient-doctor conversations to actively generate clinical notes, these AI tools improve nearly every measure of efficiency.² They provide a modest decrease in time spent within the note while improving the overall quality of documentation, which allows physicians to close their encounters more quickly and take less work home.³ The true advantage, however, is markedly increased time with face-to-face patient interaction.⁴

Figure 2: Ambient Scribes. There are a multitude of secure AI programs that are able to listen to patient-clinician conversations in real time and generate clinical notes using speech recognition with natural language processing. Image adapted from Elion's AI Scribe Market Map.

AI tools built directly within the EMR can further support the clinical work done by physicians. In the outpatient setting, AI will automatically review recent provider notes and provide summaries of interval visits to the user's specifications. Gone will be the need to pre-chart in the days leading up to appointments or hunt through the EMR to determine when a certain therapy was changed.

A similar function exists for inpatient stays, wherein important clinical changes over the last 24 hours will be highlighted in a list-like summary to review before the chart is even open. This could include modifications to medical treatments, vitals trend differences and new test results. AI can even pull data from signed notes to maintain an up-to-date clinical summary prior to patient discharge, which is especially helpful for busy services with rotating learners. Broader applications extend to addressing prior authorization rebuttals, organizing clinical schedules, and guiding billing.⁵ By existing within the EMR itself, these tools are especially appealing to physicians, as there is no need to switch between applications. The fewer the clicks, the more likely providers are to engage with the software.

Improving clinical workflow extends beyond the EMR to generative AI platforms. These large language models, like ChatGPT, Claude and OpenEvidence, scour online databases to answer clinical questions, create medical documents and guide patient care (Figure 3). These can be thought of as each having their own personality and work ethic.

Figure 3: Large-language models. Generative AI programs are able to receive open-ended prompts and utilize online databases to create new output. This highlights the flexibility of large-language models as they can not only respond to specific requests but also tailor the manner in which they respond to fit the user's needs. Image originally created by author, edited and finalized by Gemini AI.

OpenEvidence, given its partnership with high-powered scientific journals, is the Clydesdale of the bunch. It easily drafts patient instructions or follow-up plans based on evidence-based practice, with in-line citations providing easy access to primary literature. The undifferentiated patient is uniquely suitable for investigation using OpenEvidence or other programs. With the right prompt, AI models can provide a list of likely diagnoses, initial workup, potential treatment options and even red flags to avoid. Still, there is caution to be had when using this software. It is not uncommon for AI to misinterpret findings within papers or even generate citations to articles that do not exist.^6,7

Figure 4: Pitfalls with large language models. This diagram highlights the inaccurate output that may be generated from AI models when given a specific request on certain niche topics, like congenital heart disease. While often dependent on the user's prompts, many platforms recognize their limitations and include disclosures on their professional use. Image generated by author from unspecified AI generative platform.

We as providers must be careful not to replace our own reasoning with that of AI. As an example, a prompt requesting a diagram for a specific congenital heart disease following its index operation is often too complex a request and tends to fall short of the mark (Figure 4). The importance of verifying all output generated by any of the existing large language models cannot be understated, lest we lose not only our own critical-thinking skills but also the trust that anchors a patient-provider relationship.

Taken together, these three models – ambient scribes, large language platforms and EMR-integrated tools – can be coalesced into one large hybrid AI system. This represents the future of AI technology to support clinical workflow.

Hybrid AI combines the strengths of the other models into one unique tool that is able to learn from and grow with the provider.⁸ Imagine a personal AI assistant that completes encounters from start to finish, not only performing chart review and note generation, but suggesting a specific workup from current evidence depending on history or physical exam findings, filing for billing and scheduling appropriately timed follow-up (Figure 5). Hybrid systems would answer patient messages as easily as large language platforms and respond to insurance denials without any provider input. They have the capacity to revolutionize modern medicine and return balance to the physician's day-to-day life. Most importantly, they would allow focus to return to what matters most: the patient.

Figure 5: Hybrid AI Systems. By connecting and utilizing the three aforementioned models, hybrid systems represent the future application of AI – where smart clinical workflow is optimized and the provider is able to focus on the patient rather than the computer. Image originally created by author, edited and finalized by Gemini AI.

References

1. Lukac PJ, Turner W, Vangala S, et al. Ambient AI Scribes in Clinical practice: A randomized trial. NEM AI. 2025;2(12).
2. Pearlman K, Wan W, Shah S, Laiteerapong N. Use of an AI scribe and electronic health record efficiency. JAMA Network Open. 2025;8(10):e2537000-e.
3. Guo Y, Wang J, Hu D, et al. Evaluating ambient artificial intelligence documentation: effects on work efficiency, documentation burden, and patient-centered care. J Am Med Inform Assoc. 2026;33(2):273-82.
4. Schneider KR, Swann-Thomsen HE, Ribbens TG, et al. The impact of artificial intelligence scribes on physician and advanced practice provider cognitive load and well-being. J Am Med Inform Assoc. 2026;33(5):973-8.
5. Lavoie-Gagne O, Woo JJ, Williams RJ 3rd, et al. Artificial intelligence as a tool to mitigate administrative burden, optimize billing, reduce insurance- and credentialing-related expenses, and improve quality assurance within health care systems. Arthroscopy. 2025;41(8):3270-5.
6. Labenbacher S, Niederer M, Hammer S, et al. Performance of AI tools in citing retracted literature : content analysis. J Med Internet Res. 2026;28:e88766.
7. Liu H. Fabricated citations in the age of AI: A wake-up call for editors, reviewers, and authors. J Dent Sci. 2026;21(1):679-80.
8. Fahim YA, Hasani IW, Kabba S, Ragab WM. Artificial intelligence in healthcare and medicine: clinical applications, therapeutic advances, and future perspectives. Eur J Med Res. 2025;23;30(1):848.

Keywords: Cardiology Magazine, ACC Publications, CM-Sep-2026, Digital Technology, Artificial Intelligence