Best AI for Medical Residents

Disclaimer: nothing here is clinical advice. AI tools assist with documentation, research compilation, and presentation preparation. Clinical decisions belong to the treating physician with appropriate supervision.

Residency is unusual in how much written work it generates at the worst possible times. You're post-call, you have fourteen notes to finish before afternoon conference, and you're expected to present a case you saw for the first time twelve hours ago. The documentation demands don't care about your clinical load, your fatigue level, or how complicated the patients were that day.

Most residents learn to write fast out of necessity. But fast and good aren't the same thing, and there's a particular kind of cognitive drain that comes from spending the last ninety minutes of a thirty-hour shift writing admission notes. AI tools don't change the clinical work. They do change the documentation work in ways that are real and measurable.

The tools on this list are ones that fit into how residents actually work, not how residency might work in ideal conditions. They help with note drafting, literature review, and case preparation. None of them are EHR integrations or specialty-specific clinical decision support tools. They're general-purpose AI tools that happen to be well-suited to the writing and research tasks that take up a disproportionate share of resident time.

The documentation problem in residency

The inefficiency isn't that residents don't know what to write. It's that translating clinical thinking into structured documentation is slow work that happens at the end of already-long days.

A resident who's just completed a thorough assessment and plan in their head still has to produce a note that's organized correctly, uses appropriate clinical language, covers the required elements for billing and documentation purposes, and does it for each patient on the list. That translation from thinking to writing is where time disappears.

AI tools help most at exactly that translation layer. They don't do the clinical thinking. They help move from clinical observations and assessments to organized, appropriately written documentation faster.

1. Claude (claude.ai)

Claude is the most useful general AI tool for the documentation and reasoning parts of residency. The core use case is converting clinical bullet points into structured note text.

The workflow that works: you take your clinical notes from a patient encounter, the history, exam findings, relevant labs, and your assessment and plan, organize them as bullet points, and give Claude a clear prompt about what kind of note you're drafting and for what specialty context. Claude produces structured note text that you review, edit, and transfer into your EHR. The drafting time drops significantly.

For case presentation preparation, Claude is particularly useful. Give it the clinical details of your case and ask it to help you structure a problem-oriented presentation. It will suggest how to organize the information, flag the questions you should be prepared to answer, and help you think through the differential in a structured way. The differential isn't a substitute for your own reasoning, but having Claude's structured list in front of you while you prepare gives you something to push back against.

Discharge summaries are another area where Claude saves real time. These are documents with a fixed structure, a dense information requirement, and a strong tendency to pile up at the end of a rotation. Claude handles the structure and the language efficiently from clinical bullet points.

The HIPAA caveat is important for residents: don't paste patient identifiers, medical record numbers, or other PHI into Claude's consumer interface. The standard plan doesn't include a BAA. Keep patient identifiers out of AI-generated drafts and add them in your EHR after transferring the text.

Claude Pro is $20/month, which most residents can absorb personally. Check whether your program or institution has an enterprise licensing arrangement before paying individually.

Best for: Note drafting from clinical bullet points, case presentation structuring, discharge summary generation, assessment and plan organization. Pricing: Free tier available; Claude Pro at $20/month.

2. Perplexity

Perplexity is the most useful quick research tool for residents who need to know the current evidence on a clinical topic without spending thirty minutes doing a proper literature search.

Pre-rounding is the clearest use case. Before morning rounds, you have a patient with a presentation you haven't seen before, or a management question you're not sure about, and you have fifteen minutes before you're expected to present. Perplexity searches the web and returns cited summaries. Ask it about current first-line treatment for new diagnoses, recent guideline updates, or the evidence base for a management decision. It pulls recent sources and shows you where each claim comes from.

The key limitation is that Perplexity searches public web sources, not curated clinical databases. It won't give you the deep, structured clinical guidance that UpToDate or a PubMed search with appropriate filters provides. It gives you a starting point with sources you can check, which is valuable for time-constrained research but not a substitute for thorough clinical literature review.

For preparing conference presentations or M&M discussions, Perplexity helps with rapid background research on the relevant literature. You'll still need to go to primary sources for citations and verify the accuracy of what Perplexity returns, but the initial sweep is faster than starting from scratch.

At $20/month for Perplexity Pro, the upgrade from the free tier is worth it for the higher query limits and access to more thorough searches. Residents doing meaningful research use during a heavy rotation will hit the free tier limits.

Best for: Pre-rounding rapid research, background on unfamiliar diagnoses, quick guideline lookups, literature background for conference presentations. Pricing: Free tier available; Perplexity Pro at $20/month.

3. Glean

Glean is the tool to know about if your institution has deployed it. Glean connects to an organization's internal knowledge systems and makes everything searchable in plain language. For a hospital or health system, that means clinical protocols, department policies, clinical pathways, and institutional guidelines.

The practical value for residents is real. A resident rotating through a new service at an institution that uses Glean can search for the department's specific protocol for a clinical scenario, the institution's preferred antibiotic regimen for a given infection, or the documentation requirements for a particular procedure, without having to track down the right person to ask. That information exists somewhere in the institution's systems; Glean makes it findable.

Glean is enterprise software with custom pricing. Whether it's available to you depends entirely on whether your institution has deployed it. If your program has a subscription, ask your program coordinator or the informatics team about accessing it. It's not something you purchase individually.

The other Glean use case for residents in teaching programs is searching institutional knowledge from previous residency conferences, grand rounds archives, or educational materials your program has stored in internal systems. If that content is indexed in Glean, searching it is dramatically faster than digging through shared drives.

Best for: Searching institutional protocols and clinical guidelines, navigating department-specific policies, finding institutional knowledge indexed in internal systems. Pricing: Enterprise only; check with your institution.

Board preparation and self-study

Outside of clinical documentation, Claude and Perplexity both help with the self-directed learning that residency requires.

Claude is useful for explaining clinical concepts you're working through. Ask it to explain the pathophysiology of a condition you saw in clinic, to walk through the mechanism of action of a drug class, or to help you build a framework for approaching a clinical problem. Its explanations are often clearer and more interactive than a textbook because you can ask follow-up questions.

Perplexity helps with staying current on your specialty. Clinical guidelines update, new evidence changes practice, and it's hard to keep up during residency. Perplexity searches for recent publications and guideline changes with citations, which is faster than setting up journal alerts for everything relevant.

Neither tool is a substitute for board review resources, Amboss, UWorld, or whatever your program recommends. They're supplements for the learning that happens around clinical work, not replacements for structured board prep.

Practical suggestions for residents using AI tools

Start with one tool and one task. The residents who get the most value from AI tools aren't trying to use them for everything at once. They start with the specific task that takes the most time relative to its importance, often discharge summaries or admission notes, get good at the workflow for that task, then expand.

Develop a consistent prompt structure. The first week of using Claude for note drafting, the prompts are messy and the output is inconsistent. After two weeks of doing it regularly and refining your approach, the prompts are cleaner and the output requires less editing. The investment in learning the workflow pays off.

Keep PHI out of consumer tools. This isn't optional. If your program or institution has guidance on AI tool use for clinical documentation, follow it. If there's no guidance yet, the conservative approach is to use AI tools for structuring and language with placeholder patient identifiers, and add the actual identifiers inside your EHR.

Use Perplexity for quick lookup, not for final answers. Treat everything Perplexity returns as a starting point that you verify. The citation model is helpful, but citations need to be checked and primary sources need to be read.

Frequently asked questions

Do attending physicians care if residents use AI to draft notes?

Practices vary by program and by attending. The consistent standard is that the note reflects the resident's actual clinical assessment, is reviewed and accurate, and is properly supervised and co-signed by the attending when required. The tool used to draft it is secondary. Residents who produce better notes faster are generally viewed favorably as long as the quality and accuracy hold up.

Is there AI specifically for medical education?

There are AI tools built specifically for medical education and board preparation. The tools in this guide are for clinical practice and residency workflows rather than structured test prep. If board preparation is the primary need, look at platforms built specifically for that purpose alongside what's listed here.

What about AI tools for clinical decision support?

Clinical decision support is a separate category from the documentation and research tools in this guide. Some institutions are piloting AI-assisted clinical decision support built into their EHR platforms. Those are distinct from general AI tools and are subject to institutional implementation decisions rather than individual choice. The tools in this guide assist with documentation and research, not with clinical decision-making itself.