
Doximity Ask + UpToDate: Clinical AI at the Point of Care
Doximity and Wolters Kluwer’s UpToDate are partnering to integrate UpToDate’s peer-reviewed, and, most importantly—trusted—medical content into Doximity Ask. These are two of my favorite AI companies whose platforms I use daily in my clinical workflow.
This was an unexpected partnership. When I woke up that morning, I had to take another shot of espresso and rub my eyes to make sure I was reading things correctly. Unexpected, sure. But this is such a strategic, powerful partnership. I had a lot of time to think about it on my flight home from the CHEST conference in Phoenix.
In this article, I’ll highlight the uniqueness of Doximity and UpToDate, discuss the partnership, and dissect what this means for physicians.
Doximity Ask vs UpToDate Expert AI
I previously compared these tools in Pressure Support, where I described how I use each in my clinical workflow. There’s overlap in the questions they can answer, but their strengths come from different places.
Doximity is a broader physician platform, with tools for calling patients, conducting telehealth visits, and documenting encounters through Scribe. Ask adds clinical questions to that workflow, which I explored in my deep dive on Doximity’s AI toolkit. Its acquisition of Pathway Medical brought structured guidelines, drug information, and landmark trials into the platform, supporting concise, citation-backed answers. I’ve used it for quick lookups, like checking a medication dose or reviewing a guideline recommendation.
You can learn all about Doximity’s clinical tools in my latest course to prepare residents to use these tools.
UpToDate’s strength is its editorial process. Expert clinicians synthesize the literature into peer-reviewed guidance, including recommendations for situations where research doesn’t give us a clear answer. That combination of evidence and expert judgment is what makes it useful for working through management decisions.
UpToDate Expert AI adds a conversational interface to that knowledge base. It identifies the topics informing its responses and makes its assumptions and rationale available for review. I’ve found that especially useful for more nuanced clinical questions.
The way I think about these tools: Doximity brings a range of tools into the physician’s daily workflow, while UpToDate brings depth in curated clinical guidance.
Doximity Ask + UpToDate Expert AI
Doximity and Wolters Kluwer announced their partnership on October 20. The agreement brings UpToDate’s evidence-based clinical content directly into Doximity Ask, so we can access its guidance without leaving Ask.
How it will work: Clinicians at health systems using both Doximity and UpToDate will see an answer based on UpToDate’s 13,000 clinical topics and best practice guides in addition to Ask’s own response. That gives us another source of guidance to review within the same workflow. Wolters Kluwer explicitly describes this as bringing UpToDate Expert AI into Ask, though it hasn’t detailed which features of the standalone Expert AI experience will be included.
When it’s coming: The companies said “in the coming weeks,” starting with those health systems using both products (like my institution). They haven’t announced an exact launch date or detailed how individual subscribers outside those systems will gain access.
The takeaway is that bringing Doximity Ask and UpToDate Expert AI together could reduce the back-and-forth between asking a question and checking the expert guidance behind an answer.
Dashevsky Dissection
This partnership makes sense because it brings expert-reviewed clinical guidance into a workflow many physicians already use. For clinicians, the potential benefit is fewer steps between asking a question, reviewing the guidance, and deciding whether it applies to our patient.
With AI medical search platforms, the workflow generally looks like this:
You ask a clinical question → The platform retrieves source material and generates an answer → You verify the answer and assess whether it applies to your patient.
That last step takes work. A citation can support a statement without supporting its application to the patient in front of us. We still need to check the population, the exceptions, and what the answer might have left out.
UpToDate adds expert judgment before we even ask the question. As I discussed in my previous article on how Expert AI builds trust, Expert AI is designed to ground its responses in content that clinicians have already written, synthesized, and reviewed:
Clinical experts write and review UpToDate content → You ask a clinical question → Expert AI generates an answer grounded in that content → You verify the answer and its relevance to your patient.
That editorial work is valuable when studies leave gaps or several reasonable management options exist. It gives the AI a curated foundation, though the generated answer can still omit something important or apply guidance incorrectly.
Doximity’s PeerCheck program adds physician oversight through evaluation and improvement of its outputs. That’s a different role from reviewing the underlying clinical reference content, and it shouldn’t be read as an expert signing off on every answer we receive.
The announced integration puts an UpToDate-based answer alongside Ask’s own response. I’m interested in whether that makes it easier to compare recommendations, follow the sources, and understand any disagreement. Two answers could also add work if they differ without making the reasons clear.
The distribution piece is important, too. At my institution, I have an UpToDate button in Epic. Access is easy… for me. Other clinicians may have to pay for their own subscription or practice somewhere without institutional access. Bringing UpToDate into Ask could put that guidance in front of more clinicians, but the announced rollout still focuses on health systems using both products. I’m interested in whether this partnership eventually expands who can access UpToDate, alongside making it more convenient for existing users.
My take is that this is a promising combination of clinical content and workflow access. I’ll judge its value by how easily I can check the sources, spot uncertainty, and apply the guidance responsibly, not simply by how quickly an answer appears.





