Podcasts, articles, and executive intelligence for healthcare leaders who need useful context without wasting time.
Read the Weekly BriefWhat health IT leaders are talking about this week
New episode · October 15
The 229 PodcastOctober 15, 2026: Four September roundtables, one AI panel. A Piedmont physician built a mock ICU tracker on his own time, and Piedmont cannot support it today. At Scripps Health, autonomous coding cut the cost per encounter about 50%.
October 14, 2026: Jim Jacobs, President and CEO at MediQuant, sits with Bill Russell. Cold storage looks cheap until a team has to open the file. The last 3 years of MediQuant's own archive are greater than the previous…
Go to episode
New this week
October 5–11, 2026
Published Wednesday, October 14
Ambient and evidence-based AI vendors are both repositioning as a clinical intelligence layer, while the measurable capacity wins sit in credentialing, scheduling and bed turnover, not in per-clinician time savings.
Dominating the week
Drawn from conversations with Michal Nedoszytko, Venkat Kavarthapu and Yaw Fellin.
Read the full Weekly BriefWhat they’re talking about
Abridge is building pre-visit (agent conversations that prep the patient), in-visit (ambient capture), and post-visit (follow-up) components, and has already shipped real-time voice agents that let physicians verbally query charts and close operative notes instead of typing. The stated ambition is flagging missed medications, guideline changes and care gaps — with 'Care Gap' prototypes in development with health system partners including University of Chicago, aimed at the behavioral rather than demographic reasons patients skip prophylactic visits.
Symplr CEO Venkat Kavarthapu, citing a hospital CFO, put 10-15% of national bed capacity in the 'ghost bed' category, staffed beds locked by upstream credentialing and scheduling bottlenecks rather than by staffing shortfalls. Credentialing runs 120-180 days industry-wide and nurse managers reportedly spend half their time on scheduling, making both concrete automation targets; a Becker's survey found 85% of nurses, NPs and physicians want AI applied to administrative burden.
The panel agreed that per-user time savings from AI do not automatically convert into financial or capacity outcomes, the translation math remains unsolved for most systems. Meanwhile 2025 health IT reductions continued across Stanford (80 jobs), Trinity Health (557 positions outsourced), Wellstar (761 jobs), PeaceHealth (support shifting to Tech Mahindra/HCI Group in November), UPMC, MaineHealth and others.
Wolters Kluwer is embedding UpToDate natively into Epic workflows to end the link-out experience, positioning 7,000+ expert contributors and 30 years of editorial grading as 'earned trust' against generative AI error rates; its expert AI system surfaces its own assumptions (that a patient isn't pregnant, for example) rather than just citing sources. Separately, Michal Nedoszytko argues patients in most jurisdictions still have no legal entitlement to full visit transcripts even though the capture technology is already deployed — a gap he says protects neither patients nor physicians.
229 Signal
See what your peers are actually prioritizing, building, and worrying about. No vendors. No analyst spin.
Join 229 Signal
The organizations backing our podcasts, newsroom, and community.
Looking for help with something specific? Filter all 54 partners by the problem you’re solving.
