TrustWorks On Call: Three Graphics on AI

*|MC:SUBJECT|*

p{
margin:10px 0;
padding:0;
}
table{
border-collapse:collapse;
}
h1,h2,h3,h4,h5,h6{
display:block;
margin:0;
padding:0;
}
img,a img{
border:0;
height:auto;
outline:none;
text-decoration:none;
}
body,#bodyTable,#bodyCell{
height:100%;
margin:0;
padding:0;
width:100%;
}
.mcnPreviewText{
display:none !important;
}
#outlook a{
padding:0;
}
img{
-ms-interpolation-mode:bicubic;
}
table{
mso-table-lspace:0pt;
mso-table-rspace:0pt;
}
.ReadMsgBody{
width:100%;
}
.ExternalClass{
width:100%;
}
p,a,li,td,blockquote{
mso-line-height-rule:exactly;
}
a[href^=tel],a[href^=sms]{
color:inherit;
cursor:default;
text-decoration:none;
}
p,a,li,td,body,table,blockquote{
-ms-text-size-adjust:100%;
-webkit-text-size-adjust:100%;
}
.ExternalClass,.ExternalClass p,.ExternalClass td,.ExternalClass div,.ExternalClass span,.ExternalClass font{
line-height:100%;
}
a[x-apple-data-detectors]{
color:inherit !important;
text-decoration:none !important;
font-size:inherit !important;
font-family:inherit !important;
font-weight:inherit !important;
line-height:inherit !important;
}
table[align=left]{
float:left;
}
table[align=right]{
float:right;
}
#bodyCell{
padding:10px;
}
.templateContainer{
max-width:600px !important;
}
a.mcnButton{
display:block;
}
.mcnImage,.mcnRetinaImage{
vertical-align:bottom;
}
.mcnTextContent{
word-break:break-word;
}
.mcnTextContent img{
height:auto !important;
}
.mcnDividerBlock{
table-layout:fixed !important;
}
body,#bodyTable{
background-color:#FAFAFA;
}
#bodyCell{
border-top:0;
}
.templateContainer{
border:0;
}
h1{
color:#202020;
font-family:Helvetica;
font-size:26px;
font-style:normal;
font-weight:bold;
line-height:125%;
letter-spacing:normal;
text-align:left;
}
h2{
color:#202020;
font-family:Helvetica;
font-size:22px;
font-style:normal;
font-weight:bold;
line-height:125%;
letter-spacing:normal;
text-align:left;
}
h3{
color:#202020;
font-family:Helvetica;
font-size:20px;
font-style:normal;
font-weight:bold;
line-height:125%;
letter-spacing:normal;
text-align:left;
}
h4{
color:#202020;
font-family:Helvetica;
font-size:18px;
font-style:normal;
font-weight:bold;
line-height:125%;
letter-spacing:normal;
text-align:left;
}
#templatePreheader{
background-color:#fafafa;
background-image:none;
background-repeat:no-repeat;
background-position:center;
background-size:cover;
border-top:0;
border-bottom:0;
padding-top:9px;
padding-bottom:9px;
}
#templatePreheader .mcnTextContent,#templatePreheader .mcnTextContent p{
color:#656565;
font-family:Helvetica;
font-size:12px;
line-height:150%;
text-align:left;
}
#templatePreheader .mcnTextContent a,#templatePreheader .mcnTextContent p a{
color:#656565;
font-weight:normal;
text-decoration:underline;
}
#templateHeader{
background-color:#ffffff;
background-image:url(“https://mcusercontent.com/c877545641c837fafbce6c814/images/e269b796-ba8b-4a73-05d2-6e7403bf534f.png”);
background-repeat:no-repeat;
background-position:top;
background-size:auto;
border-top:0;
border-bottom:0;
padding-top:100px;
padding-bottom:100px;
}
#templateHeader .mcnTextContent,#templateHeader .mcnTextContent p{
color:#202020;
font-family:Helvetica;
font-size:16px;
line-height:150%;
text-align:left;
}
#templateHeader .mcnTextContent a,#templateHeader .mcnTextContent p a{
color:#007C89;
font-weight:normal;
text-decoration:underline;
}
#templateBody{
background-color:#FFFFFF;
background-image:none;
background-repeat:no-repeat;
background-position:center;
background-size:cover;
border-top:0;
border-bottom:2px solid #EAEAEA;
padding-top:0;
padding-bottom:9px;
}
#templateBody .mcnTextContent,#templateBody .mcnTextContent p{
color:#202020;
font-family:Helvetica;
font-size:16px;
line-height:150%;
text-align:left;
}
#templateBody .mcnTextContent a,#templateBody .mcnTextContent p a{
color:#007C89;
font-weight:normal;
text-decoration:underline;
}
#templateFooter{
background-color:#fafafa;
background-image:none;
background-repeat:no-repeat;
background-position:center;
background-size:cover;
border-top:0;
border-bottom:0;
padding-top:9px;
padding-bottom:9px;
}
#templateFooter .mcnTextContent,#templateFooter .mcnTextContent p{
color:#656565;
font-family:Helvetica;
font-size:12px;
line-height:150%;
text-align:center;
}
#templateFooter .mcnTextContent a,#templateFooter .mcnTextContent p a{
color:#656565;
font-weight:normal;
text-decoration:underline;
}
@media only screen and (min-width:768px){
.templateContainer{
width:600px !important;
}

} @media only screen and (max-width: 480px){
body,table,td,p,a,li,blockquote{
-webkit-text-size-adjust:none !important;
}

} @media only screen and (max-width: 480px){
body{
width:100% !important;
min-width:100% !important;
}

} @media only screen and (max-width: 480px){
.mcnRetinaImage{
max-width:100% !important;
}

} @media only screen and (max-width: 480px){
.mcnImage{
width:100% !important;
}

} @media only screen and (max-width: 480px){
.mcnCartContainer,.mcnCaptionTopContent,.mcnRecContentContainer,.mcnCaptionBottomContent,.mcnTextContentContainer,.mcnBoxedTextContentContainer,.mcnImageGroupContentContainer,.mcnCaptionLeftTextContentContainer,.mcnCaptionRightTextContentContainer,.mcnCaptionLeftImageContentContainer,.mcnCaptionRightImageContentContainer,.mcnImageCardLeftTextContentContainer,.mcnImageCardRightTextContentContainer,.mcnImageCardLeftImageContentContainer,.mcnImageCardRightImageContentContainer{
max-width:100% !important;
width:100% !important;
}

} @media only screen and (max-width: 480px){
.mcnBoxedTextContentContainer{
min-width:100% !important;
}

} @media only screen and (max-width: 480px){
.mcnImageGroupContent{
padding:9px !important;
}

} @media only screen and (max-width: 480px){
.mcnCaptionLeftContentOuter .mcnTextContent,.mcnCaptionRightContentOuter .mcnTextContent{
padding-top:9px !important;
}

} @media only screen and (max-width: 480px){
.mcnImageCardTopImageContent,.mcnCaptionBottomContent:last-child .mcnCaptionBottomImageContent,.mcnCaptionBlockInner .mcnCaptionTopContent:last-child .mcnTextContent{
padding-top:18px !important;
}

} @media only screen and (max-width: 480px){
.mcnImageCardBottomImageContent{
padding-bottom:9px !important;
}

} @media only screen and (max-width: 480px){
.mcnImageGroupBlockInner{
padding-top:0 !important;
padding-bottom:0 !important;
}

} @media only screen and (max-width: 480px){
.mcnImageGroupBlockOuter{
padding-top:9px !important;
padding-bottom:9px !important;
}

} @media only screen and (max-width: 480px){
.mcnTextContent,.mcnBoxedTextContentColumn{
padding-right:18px !important;
padding-left:18px !important;
}

} @media only screen and (max-width: 480px){
.mcnImageCardLeftImageContent,.mcnImageCardRightImageContent{
padding-right:18px !important;
padding-bottom:0 !important;
padding-left:18px !important;
}

} @media only screen and (max-width: 480px){
.mcpreview-image-uploader{
display:none !important;
width:100% !important;
}

} @media only screen and (max-width: 480px){
h1{
font-size:22px !important;
line-height:125% !important;
}

} @media only screen and (max-width: 480px){
h2{
font-size:20px !important;
line-height:125% !important;
}

} @media only screen and (max-width: 480px){
h3{
font-size:18px !important;
line-height:125% !important;
}

} @media only screen and (max-width: 480px){
h4{
font-size:16px !important;
line-height:150% !important;
}

} @media only screen and (max-width: 480px){
.mcnBoxedTextContentContainer .mcnTextContent,.mcnBoxedTextContentContainer .mcnTextContent p{
font-size:14px !important;
line-height:150% !important;
}

} @media only screen and (max-width: 480px){
#templatePreheader{
display:block !important;
}

} @media only screen and (max-width: 480px){
#templatePreheader .mcnTextContent,#templatePreheader .mcnTextContent p{
font-size:14px !important;
line-height:150% !important;
}

} @media only screen and (max-width: 480px){
#templateHeader .mcnTextContent,#templateHeader .mcnTextContent p{
font-size:16px !important;
line-height:150% !important;
}

} @media only screen and (max-width: 480px){
#templateBody .mcnTextContent,#templateBody .mcnTextContent p{
font-size:16px !important;
line-height:150% !important;
}

} @media only screen and (max-width: 480px){
#templateFooter .mcnTextContent,#templateFooter .mcnTextContent p{
font-size:14px !important;
line-height:150% !important;
}

}

*|MC_PREVIEW_TEXT|*<!–

View this email in your browser

Three Graphics on AI
July 30, 2026

Welcome to TrustWorks On Call, here with your healthcare business and strategy 411 for the week. If you enjoy our work, please consider forwarding it along to a friend and encouraging them to subscribe.

This week, we present to you a special edition of TrustWorks On Call. Rather than covering the week’s headlines, we go Beyond the Whiteboard with three graphics on the present and future of healthcare AI for consumers, physicians, and payment. Then, we are Dialing In on the challenges of implementing AI on top of faulty workflows.

<!–

–>

Beyond the Whiteboard
Visualizing key trends from the healthcare industry

Healthcare AI is Happening, With or Without Provider Involvement
Last week, OpenAI followed through on its January promise to launch ChatGPT Health, allowing users to upload securely their supported medical records. OpenAI claims that 300M of its 900M global weekly users are asking ChatGPT health-related questions. ChatGPT Health is a response to that trend, as OpenAI wants to help you “take a more active role in your health while supporting, not replacing, the care you receive from medical professionals.” According to companies like OpenAI and Anthropic, consumers naturally turn to their products with intimate health questions, and it is their responsibility to create secure environments that allow these conversations to happen responsibly. In other words, the train has already left the station.

A West Health-Gallup survey conducted in late 2025 found that 25 percent of US adults have asked AI chatbots health-related questions. Among those using AI chatbots for health, users particularly enjoy getting immediate, personalized answers to health questions and being able to do research before and after seeing a provider. Although the emphasis from all parties is on care supplementation, about 14 percent of AI-for-health users, or about 14M US adults, report skipping a recent provider visit because of the AI-generated health information they received.

Very few Americans strongly trust the accuracy or security of AI, but they nonetheless expect it to take on new and increasingly autonomous capabilities in support of their health. A Boston Consulting Group survey conducted in November 2025 found that over half of adults think AI will book appointments and manage referrals within two years, and over a third predict it will flag drug interactions, recommend personal treatments, and diagnose conditions with high accuracy. Just under one in five expect it to replace humans for basic care. (More on that later.)

Physicians Embracing Non-Clinical AI Tools First
A growing number of physicians are using AI tools professionally in more ways than ever before, but concerns over how to validate their safety, efficacy, and privacy hold back further adoption, especially as it relates to direct patient care.

From an American Medical Association survey, which polled almost 1,700 physicians in early 2026: 81 percent of physicians reported that their practice uses at least one AI tool, most commonly to summarize medical research (39 percent), create discharge instructions (30 percent), and provide documentation assistance (30 percent). In 2023, half as many physicians reported any usage of AI tools. In comparison, the share of office-based physicians adopting electronic health records (EHRs) went from 40 percent in 2012 to 82 percent in 2016, a similarly rapid spike which was preceded by a far longer tail for early adopters to work out the kinks.

(It is possible that these 2026 AI adoption numbers are undercounts: OpenEvidence, a company that operates an AI-chatbot research service for medical providers, claims that 65 percent of US doctors used their service in the month of April.)

There is also an important discrepancy between what physicians currently feel comfortable using AI for, and what consumers want or expect AI to do for their health. Despite consumers’ expressed desire and expectations for personalized, AI-enabled care, assistive diagnosis has grown less rapidly than other, less clinical applications of AI, such as documentation assistance and patient communication. Physicians’ comfort levels with AI can be expected to increase with time, as both the technology and protocols around its governance improve. Giving physicians clear protocols on how to use new software, a fair distribution of AI oversight, and designated accountability for when AI-generated mistakes affect the patient will go a long way toward getting providers to embrace new AI tools.

Technology Innovation Demands Payment Innovation
Clinical AI, which is the direct (assistive or autonomous) application of AI tools to patient care and medical decision-making, will never be fully realized under our current payment system. This conclusion was the starting point for the Peterson Health Technology Institute’s roundtable, which last May convened senior leaders from care delivery, insurance, academia, and government to discuss current and future payment models in the age of AI.

The immediate threat AI poses to our current payment system is that AI scales human effort in unprecedented ways, and fee-for-service payments reward volume, creating a significant risk of AI-enabled services being inflationary. A payment system built around value-based care is of course the answer to this problem, but our industry’s decades-long collective efforts to reward value have (with a few exceptions) fallen woefully short. Even successful pay-for-performance and capitation models were not designed with AI in mind and will require upfront investments for the reworking of workflows, payer incentives, and cost attribution.

AI delivering care autonomously looms as a true paradigm shift, not only for its potential impact on patients’ access to care but also because of the fundamental rethinking of healthcare payment it would require. Physicians’ time is the most valuable resource in healthcare, and much of today’s payment system derives at least in part from its value and scarcity. Autonomous clinical AI would face no such constraints, scaling with previously unimaginable ease. Accounting for this in our payment system would transform the medical profession, as the services AI proves capable of handling would no longer be worth (in reimbursement) a doctor’s time.

One can argue we have been through this before, unsuccessfully. The Affordable Care Act, the adoption of EHRs, and the industry-wide embrace (at least rhetorically) of value-based care painted a bright future for our healthcare system back in 2012. Almost 15 years later, those beliefs never materialized and now seem naively optimistic. We now have another chance. This roundtable identified good principles to guide another take at payment reform, asserting that it should be deflationary, outcomes-based, and adaptive to evolving evidence. We must overcome our industry’s strategic intransigence and the momentum of the status quo in order for these ideas to make it beyond the whiteboard and into reality.

<!–

–>

Dialing In
Sharing insights from our work with clients

AI Does Not Fix Bad Workflows
A health system engaged us to co-design a blueprint to replace its fragmented, clinic-centered outpatient scheduling system with a centralized patient-navigation center. Their leadership wanted it to showcase as many best-in-class AI applications as were suitable to the scope, and we shared their enthusiasm that AI agents for scheduling support and patient communication were going to be central to the project. Our disagreements emerged on the human side of the ledger.

Taking resources and administrative control out of clinics and placing them in a central system office proved immediately contentious. Leadership assured us that doctors would revolt if their clinics’ autonomy were jeopardized. Instead of overhauling their scheduling process from the ground up, they wanted to layer an AI-enabled system overtop, one that solved their patients’ problems of navigating different access channels with minimal disruption to how the clinics already operate. But when the problem is specialists not accommodating referral slots for the system’s primary care patients, no AI tool can solve this on its own. Political problems, such as resource allocations between clinics, require political solutions, for which AI can only be a helpful technological expedient. Strategic planning about cutting-edge tech is a lot more fun than talking to your physicians about workflow redesign, but you cannot advance the former without addressing the latter.

<!–

–>

Thank you for tuning into this week’s TrustWorks On Call. We’ll see you next Thursday with another round of TrustWorks Takes. With your help in sharing TrustWorks On Call (subscribe here!), we’re living up to the “Collective” in TrustWorks Collective. And if you ever need help thinking through a healthcare problem, don’t hesitate to reach out to us.

Still written by human hands,
Anthony D’Eredita,
Lisa Bielamowicz, MD,
and TrustWorks Collective