Oracle's clinical AI agent now writes the charge code — AI moved into the revenue workflow, not just the note
August 20, 2026 · 3 items
Oracle's clinical AI agent now writes the charge code — AI moved into the revenue workflow, not just the note
Healthcare IT News · Andrea Fox · August 20, 2026Practice operations
Oracle said Wednesday that its Clinical AI Agent, embedded in the Oracle Health EHR, can now analyze the conversation during a patient visit and suggest professional-fee charge codes inside the orders workflow. The clinician reviews and confirms before submitting — the human stays in the loop, but the first draft of the bill is now written by the agent.
Two other new capabilities shipped with it: clinician-controlled dictation into any text field (real-time transcription for review/edit/approve), and automated chart-review summaries that pull medical history, labs and meds into one pre-visit context view using semantic reasoning rather than keyword search.
Oracle's own number: over nearly two years the Clinical AI Agent has saved physicians across US health organizations more than 400,000 hours. That's a vendor figure, not an independent study — treat it as a claim, not evidence.
Seema Verma, GM of Oracle Health and Life Sciences: "Care teams can't afford to spend hours on documentation and administrative tasks."
Why it matters for an independent practice: The pattern to watch isn't the dictation — it's AI crossing from documentation into charge capture. Once the EHR drafts the code, coding accuracy becomes a software question, and the practices that lose money to undercoding get a lever they never had. For an independent regen or specialty practice this is the argument for asking your EHR vendor "what does your agent do on the revenue side?" — that's where the ROI conversation actually lands. The stewardship caveat is real though: a suggested code the physician rubber-stamps is still the physician's signature on the claim.
Deloitte surveyed 64 US healthcare finance leaders — 32 from health systems, 32 from health plans. It classified 44% as "AI scalers" (they'd deployed generative AI broadly). But CFOs at only 18% of those organizations said they consistently measure AI's contribution to revenue growth or cost savings.
The broader readiness gap: across seven areas of enterprise decision-making, 73% of respondents said they're expected to be regularly or heavily involved, but only 49% felt well-equipped — an average 24-percentage-point gap. The widest single gap was patient experience/access/affordability: 74% expected to weigh in, 41% prepared (33 points; 38 points among health-system CFOs).
Margin pressure is what's driving it: nearly 60% are targeting operating-margin improvement of at least 2 points over the next two years, and 27% want more than 5 points. Only 47% said their organization was prepared to manage the pressures affecting those margins.
Pharmacy-adjacent signal: 85% of health-plan finance leaders expect GLP-1 and specialty-drug costs to have a moderate-to-major margin effect — but just 38% said they're well-prepared for it.
Why it matters for an independent practice: This is the measurement gap sitting underneath every AI pitch a practice hears right now. Four in ten organizations are running AI at scale and fewer than one in five can attribute a dollar to it — which means the buyer, not the vendor, has to define the metric before signing. For MMR the sales angle writes itself: walk in with the measurement framework (baseline, attribution, dated before/after), not another tool. And the GLP-1 line is worth holding onto — plans expecting margin pain from GLP-1s is upstream of coverage decisions that hit compounding and cash-pay practices directly.
Google started rolling out the August 2026 spam update on Monday — check every client's Search Console from Aug 18 forward
Search Engine Journal · Matt G. Southern · August 18, 2026Patient acquisition
Google's Search Status Dashboard logged the update as a ranking incident beginning 9:27 a.m. Pacific on Aug 18, with the release note posted a minute later. It applies globally and to all languages, and Google says it may take a few days to complete.
This is the third spam update of 2026 (after March and June). For scale: the March rollout finished in 19 hours 30 minutes, June took 2 days 1 hour — so movement you see this week is likely still mid-rollout, not a settled result.
Google published no companion blog post and no new spam policy types. The spam-updates documentation page hasn't changed since December. The existing spam policies remain the rules.
Context worth remembering: the June update followed a May 15 clarification that Google's spam policies also apply to attempts to manipulate generative AI responses — AI Overviews and AI Mode included. Google has not said whether August targets that behavior. Don't claim it does.
Recovery, per Google's own documentation, can take months — its automated systems need time to see a site is compliant again.
Why it matters for an independent practice: Practice sites live and die on local search, and a spam update is exactly when thin AI-generated service pages and scaled content get repriced — the kind of shortcut a cheap agency may have left on a client's site. Concrete action this week: pull Search Console for every client from Aug 18 onward, note movement, and don't diagnose off a single day's data. If a site did drop, the fix is a compliance review against the existing spam policies — and the honest client conversation is that recovery is measured in months, not days. This is also the moment physician-authored, genuinely-expert content pays for itself.
