Part 2 of this series traced how Mosai's models evolved from a manually printed PDF into a continuously updating signal that helped a BAYADA care team catch a health decline early enough to prevent a hospitalization. But a better model doesn't automatically mean a trusted one. Every signal in this series so far only mattered because a clinician chose to act on it — and that choice has to be earned, visit by visit, with real skeptics in the room.
This post covers two sides of the same coin from the panel: how BAYADA built that trust from the ground up, and how leadership is thinking about the flip side of efficiency gains — protecting clinicians from burnout rather than simply asking them to do more.
Amy Hirsch didn't romanticize the rollout. Asked directly how BAYADA got experienced clinicians to trust a signal over their own instincts, she was candid that skepticism was the starting point, not an obstacle to route around:
"Nurses like to see things before they believe them. And, you know, that initial start off... this is going to tell me when my patient's probability is going to die in the next seven to ten days. I don't know if I believe that." — Amy Hirsch, BAYADA
Rather than arguing clinicians into compliance, Hirsch's team let the tool prove itself in the field. When a signal didn't match a clinician's own read on a patient, the response wasn't to override the clinician — it was to go verify:
"In the end, explaining to the clinician that this is a tool that we use — it doesn't replace your clinical judgment, but maybe we should go out a few more times to see what it looks like." — Amy Hirsch, BAYADA
Nine times out of ten, Hirsch said, the clinician came back and confirmed the signal was right — the patient really was declining faster than expected. That repeated pattern is what created what she called the "light bulb" moment:
"Once that switch happens, it's really fun to see the light bulb come up and say, 'oh my god, I get it now.' ...Then you've got a champion for the program." — Amy Hirsch, BAYADA
The mechanism behind that trust matters as much as the moment itself. When clinicians ask how the system knows what it knows, Hirsch's answer reframes the whole relationship: the model isn't a black box guessing at outcomes — it's reflecting the clinician's own documentation back to them, faster and with more context than they could piece together alone. "It's all in your documentation," as she put it — the system is surfacing what the clinician already put into the chart, just sooner and more legibly. Once that clicks, champions start recruiting the next skeptic themselves, and adoption compounds without anyone mandating it from the top down.
Justin Searle placed this in a broader context: adoption friction isn't unique to AI, and it isn't really about the technology at all.
"There's some broader trepidation that exists with our frontline clinicians who are very naturally asking, like, hold on — you're moving their cheese a little bit." — Justin Searle, BAYADA
He pointed out that introducing a tool like this often forces an organization to confront workflow and process questions it had been avoiding — standardizing operations that were previously ad hoc, for instance — and that discomfort can get misattributed to "the AI" when it's really a deeper operational reckoning. His conclusion, though, was pragmatic rather than resigned: technology is already reshaping every other part of daily life, and healthcare isn't going to be the exception. The partnership with Mosai, in his framing, is less about adopting a tool and more about forcing a useful audit of whether BAYADA's own processes are set up to deliver the best care and take the best care of its clinicians.
Late in the session, an audience member asked the sharper version of a question every provider eventually has to sit with: if AI is saving clinicians time by compiling and translating data, doesn't that just create pressure to raise the number of patients each clinician is expected to see? Does efficiency just become an excuse to ask for more?
Justin Searle didn't dodge it — he named the tension directly and pointed to where the real answer has to live:
"I think there's a way to harmonize both of those opportunities... if you can take a start of care down to two hours or an hour and a half, or you can optimize and drive better quality in a forty-five-minute visit because of one's ability to be totally present — [that matters]. But the objective is we want to be able to give them their nights back and have them at top of license and top of title." — Justin Searle, BAYADA
He was candid that this isn't a solved problem industry-wide — it's a choice every organization has to make deliberately, visit by visit:
"Each company is going to have to wrestle with, hey, what do you do with that efficiency? Do you give it back to the clinician, or are we going to try to drive more productivity? I think a lot of us will be trying to figure that out over the coming months." — Justin Searle, BAYADA
For BAYADA, the answer isn't ambiguous: reducing clinician burnout while improving patient outcomes is the whole point, not a side effect. Time saved through better-informed, better-targeted care is only meaningful if it's actually protected and handed back to the clinician, rather than quietly absorbed into a higher caseload. It's the same logic driving BAYADA's guiding principle from Part 1 of this series — frontline clinicians are the organization's lifeblood, and a strategy that erodes their bandwidth in the name of productivity works against its own stated purpose.
It's worth naming the connective tissue between these two topics, because they're often treated as separate workstreams. A clinician who doesn't trust a signal will ignore it, wasting the time it was supposed to save. A clinician who's overloaded won't have the bandwidth to build that trust in the first place — there's no time to "go out a few more times to see what it looks like," as Hirsch described, if the schedule is already maxed out. Getting adoption right and protecting against burnout aren't two separate initiatives. They reinforce each other, or they both fail together.
None of this happens by accident, though — in every story so far in this series, someone in a leadership role made a deliberate choice to frame why the technology mattered before asking anyone to use it. That's where this series turns next
This is Part 3 of a 4-part series recapping Mosai's live webinar with BAYADA. Next up: "Leadership, Culture, and the Home Health–Hospice Continuum" — the single factor Mosai's CEO says predicts success more than any feature of the technology itself, and how a unified view across service lines changes outcomes for patients and families.
→ Watch the full webinar on-demand
→ Continue reading with Part 4