Reclaiming Time and Reducing Burnout With AI: A Guide for NPs

More than half (54%) of nurse practitioners (NPs) report moderate to high burnout. The root causes of burnout, like inbox overflow and the burden of electronic health record (EHR) documentation, are often perceived as problems without viable solutions. 

At the American Association of Nurse Practitioners (AANP) annual meeting, held from June 23 to 27, 2026, in Las Vegas, Nevada, Kaneez R. Odgers, DNP, APN, FNP-BC, presented ways in which artificial intelligence (AI) can be part of the solution to widespread burnout among NPs. Dr Odgers shared 4 use cases for AI in primary care: ambient documentation, inbox triage and drafting, prior authorization support, and clinical decision support.

“There are a lot of burdens that are not about seeing the patient: it’s documentation, it’s prior authorizations, it’s working within electronic health records, and it’s trying to access information,” she told The Clinical Advisor. “And with AI being increasingly used, we know that there are tools that we can use that can give us back time in our day and do administrative tasks. I liken it to having an assistant.” 

With AI-assisted care, Dr Odgers explained, NPs can break the cycle that occurs when burnout drives NPs to leave the field, the patient load grows for remaining NPs, administrative burden increases, there is less time for each patient, and then more NPs experience burnout. She noted that 1 in 5 NPs are considering leaving practice within 2 years due to burnout. 

Listen to a Conversation with AANP President, Valerie Fuller

“Most of us went into this — whether physicians, nurse practitioners, or physician assistants — to spend time with patients,” she said. “And what we’re finding is that we’re spending the majority of our time on these other tasks, and not with patients.”

To reclaim time and reduce cognitive load spent on repetitive, low-value tasks, Dr Odgers recommends that NPs use AI tools alongside their clinical judgment. 

Use Cases for AI in Primary Care

Ambient Documentation

Subjective, objective, assessment, plan (SOAP) notes take about 20 minutes to create without AI assistance, but when AI listens to the patient encounter in real time, it can create a structured draft for an NP to review, edit, and sign, cutting the total time down to about 8 minutes per note. Dr Odgers noted that this can help NPs reclaim between 90 and 120 minutes per day, with a 40% reduction in after-hours charting. 

“If you use AI to create a SOAP note, you have to read it thoroughly, vet it, and make sure that it is what you want it to say,” Dr Odgers said. “If it makes a suggestion on treatment, it’s not a hard and fast rule. That decision remains yours to make, and you just have to use it as an assistant.”

Inbox Triage and Drafting

By using AI tools to manage their inbox, NPs can achieve an approximate 40% reduction in time spent sorting and answering messages without sacrificing accuracy or oversight, according to Dr Odgers. She explained that NPs receive 47 or more messages per day on average, and the majority of these messages are low-complexity issues that require formulaic responses.

Triage functions of AI-assisted tools include flagging urgent versus routine messages, prioritizing medication refills by risk, routing laboratory follow-ups to the appropriate NP action, and reducing decision fatigue on low-stakes items. 

For drafting messages, AI can create templated responses to patients about refills or laboratory results, generate automatic after-visit summaries, and suggest boilerplate text for common instructions. Dr Odgers emphasized that even when AI drafts the message, it should never automatically send without NP review. 

Prior Authorization Support

On average, NPs can reclaim 2 work weeks per year if they use AI for prior authorization support, according to Dr Odgers. There are 4 ways AI can shorten the prior authorization process: criteria lookup, drafting evidence-based justification letters with relevant clinical data, predicting likely-to-deny requests, and tracking pending authorizations. 

Most of us went into this — whether physicians, nurse practitioners, or physician assistants — to spend time with patients. And what we’re finding is that we’re spending the majority of our time on these other tasks, and not with patients.

Clinical Decision Support

Dr Odgers explained that while AI does not replace an NP’s clinical judgment, it can support clinical judgment by surfacing details clinicians may miss. 

“It is not replacing your clinical judgment. It should never tell you what to do. It is only an assistant. The final, you know, work has to remain yours. The clinical judgment is yours,” she said.

For example, AI can flag allergy conflicts and dosing concerns, and recommend evidence-based order bundles to reduce omission errors. It can also extract up-to-date guidelines specific to the patient interaction. Further, it can suggest consideration of differential diagnoses consistent with the clinical picture.

Recommendations for Incorporating AI Into Primary Care

Dr Odgers recommends starting with ambient documentation; incorporating AI in this way has the highest time savings and lowest ongoing cognitive effort once adopted relative to other use cases for AI, she says.

The AANP’s position on health information technology “affirms that NPs should leverage technology to enhance care delivery, improve outcomes, and support the sustainability of the NP workforce while maintaining the NP as the accountable clinical decision-maker.” 

In addition to retaining clinical judgment, it is important for NPs to understand that AI hallucinations are a real risk, Dr Odgers explained. AI can fabricate drug doses or interactions, cite outdated or non-existent guidelines, or misstate clinical criteria among other errors. 

Recent reports from Stanford researchers measured how often AI medical advice could potentially cause medical errors, specifically errors of omission. The study by Wu et al, called NOHARM (Numerous Options Harm Assessment for Risk in Medicine), which covered 10 specialties with 12,747 annotations, evaluated 1100 tasks from primary care-to-specialist consultation to measure the frequency and severity of potentially harmful errors from LLM-generated medical consultation. Potentially severe errors were found “in up to 24.6% of cases” across 20 LLMs and 4 widely used retrieval-augmented generation (RAG) clinical AI tools, wrote the authors. More than 80% of severe errors were omissions, that is, recommendations that never appear in the studies. And a “do nothing” control produced potential severe harm in 37% of cases, worse than every AI tested.

“These results show that despite strong performance on medical knowledge benchmarks, widely used AI tools can produce medical consultation advice with the potential for severe harm, and highlight the need for explicit measurement of clinical safety,” concluded the authors. 

The safeguard against AI’s confident production of incorrect clinical information, said Dr Odgers, is critical appraisal. Dr Odgers recommends treating AI output as a first draft, and never bypassing clinical review of any AI-generated content. 

To protect patient privacy, NPs should verify whether their practice has a business associate agreement (BAA) before clinical use of any AI tool. Enterprise clinical AI platforms that operate under a signed BAA are designed to prevent privacy violations. Consumer AI tools may still be appropriate for educational purposes, de-identified clinical questions, or drafting content that does not include protected health information. 

The Importance of Transparency

Transparency about AI use builds trust with patients, according to Dr Odgers. “One of the strategies I have suggested, and is supported by the literature, is to let patients know that you’re using it in order to increase the quality of the visit, not to do the visit,” she said. 

For example, an NP may say to a patient: “I use an AI assistant to help me document our visit more accurately, so I can focus on listening to you instead of typing.” 

In addition to transparency with patients, transparency in how AI tools are developed is a key component of health equity, Dr Odgers explained. If an AI tool is trained on biased data, the tool’s output will be biased and can cause harm. 

This has already played out in pulse oximetry, where AI-assisted technology is less accurate in patients with darker skin tones, as well as in sepsis prediction models that underperform in Black patients, Dr Odgers noted.

She recommends that NPs use the following equity checklist to prevent algorithmic bias: 

1. Ask: “Was this AI validated in patients like mine?”
2. Flag recommendations that seem to contradict clinical judgment.
3. Report disparities to the system.

How to Evaluate AI Tools for Your Practice

While many physicians, almost 81% according to an American Medical Association (AMA) survey, report using augmented intelligence tools in their daily work, in Dr Odgers’ experience, NPs have mixed thoughts on incorporating these tools into their practice. “I think people are excited, but they’re not sure how to integrate it in a way that is safe,” she said. “The information availability has been a little slow on the rollout, so unless you work for a large institution that provides training and information, private providers are kind of left to source this information by themselves.”

Common concerns about incorporating AI tools include patient privacy, depersonalizing care, AI making clinical errors, job security, and not knowing how to evaluate AI tools. 

“I think we’re always hesitant to use things that have the potential to not protect information, and I think we’re also afraid to create more work for ourselves. When electronic health records came out, they were sold to us with this idea that they’re going to decrease our workload, but I think most of us have found that they increased it,” Dr Odgers said. 

Table. Questions to Ask When Evaluating AI Tools for Your Practice

Question	Considerations
Does a BAA exist?	If not, the tool is not usable for protected health information.
Was the tool validated in clinical settings?	Is there peer-reviewed evidence or a rigorous pilot, or does the only evidence come from the vendor?
Who reviews before it reaches the patient?	Any patient-facing output requires NP review.
Does it work for my patient population?	Was it tested in patients with similar demographics and comorbidities?
Can I explain it to my patient?	If the clinician cannot explain how an AI tool influenced care, they should not use it for that purpose.

Overall, Dr Odgers is optimistic about how AI can help NPs reclaim time in their workday and reduce burnout. “I think it’s an exciting frontier that we are coming upon, and I think it has the ability to make our jobs easier,” Dr Odgers said. “The landscape in healthcare changes very, very fast. Technologies are advancing very quickly, and this is another tool in our arsenal.”

Visit our conference section for more coverage of AANP 2026.

This article originally appeared on Clinical Advisor