MGMA was celebrating its centennial this year, and the turnout showed it. We spent three days in San Antonio with healthcare executives, CXOs and practice administrators from groups of every size and specialty, and the conversations were some of the most interesting ones we have had.
San Antonio seemed to put everyone in a bit of relaxed mode. And who could blame them? With so much Texas history to explore and such a laid-back vibe, the city kept spirits high.
And our pen plotter fit right into that atmosphere, drawing AI portraits onto postcards in under five minutes and giving people another reason to stop and chat. As their portraits took shape, we got into some great conversations about the documentation challenges they face every day. A little creativity turned out to be a great conversation starter.
Two years into ambient AI scribes, the conversation is evolving. As specialty groups gain experience with these tools, they’re looking beyond capturing the visit to supporting the complexity of specialty care - The clinical nuances, longitudinal patient history, and documentation needs that shape each encounter.
That evolution is bringing two areas into sharper focus:
Practice leaders told us they need help with pre-charting, and several are spending multiple FTEs on it inside their organizations.
The work ranges from reviewing external hospital discharge notes that run to a hundred pages, down to the note prepping that a lot of MAs handle before visits. It becomes extensive because they are constantly pulling that data from the EHR multiple times a day, every day.
Most AI companies skip this part altogether. And this takes far more time than ambient listening, which only summarizes the encounter that just happened.
Marvix AI handles it before the appointment with Patient Recap summaries. Prior notes, assessments, imaging, labs and external records are pulled directly from the EHR into one structured chronological summary, including scanned PDFs and handwritten intake forms.
Everybody we spoke with agreed on how technical insurance requirements have become, and what they need are notes incorporating these guidelines.
The guidelines are also local. Local Coverage Determinations vary by Medicare Administrative Contractor, so the same procedure can call for different documentation depending on where the practice sits.
Marvix AI generates ICD-10-CM, CPT and E/M coding with MDM rationale, and inserts procedure-specific medical necessity language into the note while the visit is happening.
This is where our articulation resonated most. The providers and administrators we met recognized their own situation in it straight away.
Several groups we met are still evaluating AI options, not because they aren’t convinced about AI, but more so because nothing has fit their requirements yet, and a number of them are still relying on humans.
The openness and curiosity about AI are real and nobody is denying that it is here to stay.
What they want is something more meaningful, with clear and tangible impact on outcomes for the group, whether that shows up as revenue or as clinical results they can point to.
Rashie Jain, our co-founder and CEO, on what MGMA made clear:
“Specialty groups have moved past evaluating AI on note quality alone. The work consuming their staff sits before and after the encounter, in pre-charting and in insurance requirements that change by locality. Marvix AI is built across that full span, which is what makes the impact measurable at group scale.”
A fantastic conference, and a room full of people who know their own operations in detail.
Two years of AI scribes have shown practice leaders exactly which parts of documentation got solved and which parts still sit with their staff. They can tell you where the hours go, down to the headcount.
That is where the next round of AI decisions in specialty care will be made.
If you missed us at MGMA 2026, book the same demo that wowed many healthcare leaders at the conference. Book your 30-day free trial, integrated with your EHR from day one, for your entire team.
.png)
