Practice Stories
Hospitalist · physician advisor · certified AI consultant
A hospitalist who has spent twelve years inside overloaded fifteen-minute appointments is now building the AI agents she wishes had existed for her — and asking why "that's so amazing" so rarely turns into "let's use it."
Dr. Augusta Uwah traces her path into medicine back to something unanswered. Growing up in Nigeria, she had a sibling who fell sick when she was thirteen, "and they were — nobody could literally give us answers. And that's kind of where my curiosity got sparked." In the British-modeled system, students commit to medicine straight out of high school. She did, at the University of Nigeria, then came to the United States for a master's in public health, then residency at Howard University Hospital in Washington, D.C.
Twelve years of internal medicine hospitalist work later, she still describes her second degree as something that changed how she sees a patient, not just what she treats. "I wouldn't exactly call this a superpower, but a different lens... maybe like an extra special lens," she says. "The patient is so much more than their disease." She rattles off what that means in practice — how poverty, education, and trust in the system shape not just whether someone gets care, but whether their outcomes improve once they do. "Sometimes, especially for those with the lower socioeconomic status, it's been highlighted that they tend to have worse outcomes."
Ask Augusta what technology can't solve in healthcare, and she doesn't hesitate. "Technology is a tool, okay, and it cannot replace the human interaction. It's great — I mean, if you talk to chat, it will talk right back. But it's not the same as a human interaction. The depth of the emotion and the empathy is different."
What she has watched AI do, instead, is protect the minutes that actually belong to that human interaction. She describes the math of a single primary care visit: "There's a thousand things competing for that fifteen minutes of their time with you." Documentation, pharmacy coordination, follow-ups, insurance — all of it stacked against the conversation the patient came for. Ambient scribes, she says, are "the gateway drug of AI" precisely because of what they give back: "It actually helps the physician not to have his back turned to you on the computer or scribbling while talking to you. It allows them give you that face-to-face, that attention that made you want to come seek healthcare in the first place."
On her days off from hospitalist work and physician advisory consulting, Augusta builds. Not casually — she holds a Harvard certificate in agentic AI implementation in healthcare, and what she calls her favorite stack is a multi-agent clinical intelligence platform she built herself: a narrative-generation agent that argues a hospital's case to insurers, a policy agent that checks that argument against the actual coverage rules, and a compliance agent that keeps the whole system inside governance. "That's the lead guy, so to say," she says of the first agent — and she's candid about how far outside her comfort zone the build took her. "I built it, you know... I'm like, 'Oh my god, this is so complicated.'"
Her framing for physicians eyeing this shift warily isn't a sales pitch — it's a reframe. "I would say don't think of it as artificial intelligence, think of it as augmented intelligence." And she's specific about why healthcare lags: not hostility, but risk. "Healthcare is a very tightly regulated field. So sometimes the physicians are not... anti-technology, they just don't know that it is safe or that it wouldn't add friction to their workflow." Her answer is to lead with governance, not enthusiasm — BAAs in place, HIPAA compliance, minimum information necessary. "Those are the kind of things that get the physicians warmed up a little bit more to the idea of AI."
What preoccupies Augusta isn't whether AI works. It's why knowing that so rarely changes behavior. She describes a conversation with a colleague who called AI "so amazing... are you for real?" and then did nothing with it. She sets the real adoption rate against the hype: "There are people who will tell you that ninety percent of businesses are ready to adopt AI. But in reality, I think the number is probably closer to like twenty percent — maybe, if that, and this is being generous."
She reaches for Blockbuster to make the stakes concrete — a company that was "offered the opportunity to buy Netflix... they laughed the guy out of their office, and here we are now." Her diagnosis of the hesitation is unsentimental: badly deployed AI doesn't create new problems, it just exposes old ones. "If you put an AI workflow into an already broken system, all you're going to do is amplify the brokenness of the system... they'll be like, 'Oh my god, AI made everything worse.' No, it just showed you how bad everything was."
Augusta doesn't pretend to have solved the adoption gap — only to have named it clearly. She knows the tools work. She's built several of them herself, from scratch, learning as she went. What she hasn't found is the lever that turns "that's so amazing" into a changed workflow, at scale, across a field she describes as risk-averse for good reason. Her closing advice to a physician who feels threatened by the headlines is less a solution than a bet on herself and everyone still deciding: "AI is not going to replace you, but the people that have learned it and learned how to use it — those are the people that are going to be the last man standing."
Watch the full conversation to hear how she thinks through that gap, agent by agent.