AI Helps Healthcare Leaders Turn Intuition into Data-Driven Action
As pressure to personalize the patient journey grows, healthcare organizations that embrace AI are finding ways to move faster and smarter. Here’s how three leaders use AI to turn data into actual improvements.
// By Wendy Margolin //
In a field that’s constantly changing, one truth in healthcare marketing that’s here to stay is data-driven decision-making. An increasing number of consumers expect a personalized healthcare journey. Meanwhile, marketing departments are tasked with driving service-line volume and are inundated with data across EHR, CRM, digital, and call center systems.
These demands, coupled with the sheer volume of data available in healthcare today, make AI tools a necessity. When used correctly and in compliance, these tools contribute to brand trust, help marketers make smarter decisions, and improve the user experience.
In a session during AI Week in Healthcare Marketing in December, leaders from AdventHealth, UCI Health, and Mount Sinai Health System shared practical use cases on how data informs their marketing decisions and how they use AI to support marketing communications strategies.
Left to Right: John Davey, vice president marketing technology, Mount Sinai Health System; Tara Nooteboom, director of consumer digital strategy, UCI Health; Josefer Montes, executive director, marketing & branding, AdventHealth; Micha Siegel, director, product innovation and strategy, BPD
Click here to watch the session: “AI and Data-Driven Transformation in Practice: Stories and Advice from Three Major Health Systems.”
Use Case 1: AdventHealth Uses AI to Grow Medicaid Patients
When Josefer Montes, PhD, executive director of marketing & branding at AdventHealth, was tasked with increasing Medicaid patient volume, he turned to AI to support his team throughout their entire marketing campaign.
Challenge
In August 2025, AdventHealth Gordon in Calhoun, Georgia faced a critical challenge to increase Medicaid patient volumes to maintain its 340B status. The Federal 340B Drug Pricing Program allows eligible hospitals and health systems to purchase certain outpatient drugs at significantly discounted prices, provided they meet a specific threshold of Medicaid patients.
Solution
Montes used AI to research the Georgia Medicaid landscape. He turned to expert colleagues to confirm the results and learn more. His team began by promoting AdventHealth to expectant mothers because every delivery brings in at least two patients: the mother and the baby.
They grew volume but didn’t solve the specific business problem. “While we did grow our Medicaid volume in the OB, we also grew all the other volume, so it didn’t help the threshold issue as well as we would have liked,” says Montes.
That’s when the marketing team’s effort tightened into a focused Medicaid campaign.
They identified potential Medicaid patients by leveraging AI to interpret a massive amount of data including local demographic data, patient experience data, HealthGrades, Yelp, Facebook, and more. AI allowed them to pinpoint high-value areas down to the Zip Code level for precise outreach.
AdventHealth used AI to interpret demographic data in its market, enabling the marketing team to target advertising across primary and secondary service areas.
The most common question among the target audience about Medicaid was simply whether the hospital accepted it. “It became a clear moment that this is mostly an awareness campaign to let people know we accept it,” says Montes.
Next, they used economic indicator data in every zip code to determine where to target prospective patients.
AdventHealth used AI to identify families in the area that are most likely to be on Medicaid.
Audience insights guided budgeting and media activation strategy, helping them identify channels, geographies, and tactics most likely to drive awareness and increase Medicaid patient volume.
The team focused on areas where their audience naturally spends time in the priority geographic areas surfaced through the data. That led to some unusual advertising placements, like hair salons, laundromats, convenience stores, and childcare centers.
Results
The campaign began in Q4 2025, and the initial results won’t be available until the end of Q1 2026. Montes is clear on the process, even if the results are not ready. “This campaign was the first one where we used AI from out of the gate at every step of the way. And the only bad news is I wish I had started using it sooner,” he says.
Use Case 2: AI Makes Data-Driven Transformation Doable
Digital marketing leaders at UCI Health faced a conundrum many healthcare marketers are dealing with: The infinite possibilities of data-driven decision-making with AI — combined with compliance, ethical, and significant industry barriers to its use in healthcare — can lead to decision paralysis.
Tara Nooteboom, director of consumer digital strategy, offers a pragmatic counterpoint. Her use case focuses on how to pick one area as a pilot program to show how using AI to interpret data can improve processes and strategy.
Goal
The UCI Health team is determined to build more direct patient feedback and first-party consumer input into their marketing strategy and brand experience prioritization. “We want patient feedback to drive the projects we moved to the top of the queue, whether that’s in our digital experience or in operations,” says Nooteboom.
Method
The team turned to NRC survey comments to gain insights about positive and negative sentiment. They first used traditional data analysis, without entering the survey results into AI.
The UCI Health team leveraged AI to quickly identify broad themes in a large dataset.
Next, they used their custom AI tool, UCI ZotGPT, in a data-safe, protected environment to ask questions about how to prioritize a long list of brand experience projects based on survey results.
They drew on intuitive ideas from within the organization to improve the brand experience and used AI to validate which ones resonated most with patients’ survey responses. “Going from intuition to intelligence, validating some of those intuitions early at a low threshold is definitely something that AI-assistant-driven analysis can help support early in the game,” Nooteboom says.
For example, the team partnered with the patient experience deployment team on scripting and language for delays and explored ways to trigger proactive communication via SMS. “Communication around wait times and scheduling delays is one of the biggest dissatisfiers that comes up in comments, regardless of the quality of care that’s being delivered,” she says.
AI didn’t solve the problem, but it helped the UCI Health team focus faster. “It indicated where we needed to dig deeper and lift the stone to look underneath,” says Nooteboom.
The team even got UCI business school students involved. Using anonymous data, they asked students to investigate trends and issues related to physician reputation.
Traditional Analysis vs. AI-Assisted Exploration
The combination of traditional and AI data analysis helped the UCI Health team determine where to invest time and money to improve brand experience. Nooteboom acknowledges that AI data-driven research isn’t always the answer to rigorous research in healthcare marketing. She explains the distinction between the two as follows:
Traditional Analysis
- More appropriate for high-rigor situations like policy implementation and safety
- Allows deeper and more trackable categorization
AI-Driven Analysis
- Allows early exploration and trend discovery
- Sets up good questions for more rigorous, deeper dives
- Exponentially speeds up the process
Speed vs. Safety
Finally, Nooteboom emphasizes how to responsibly handle patient data, such as anonymizing comments to remove identifiers before sharing insights with vendors or business school students.
She offers a reality check for healthcare leaders eager to scale AI quickly. “The tools available in this space today require a lot of judgment and responsibility. There’s nothing I want to click and turn on for everybody today.”
Use Case 3: New Strategy for a New Search Landscape
For John Davey, vice president of marketing technology at Mount Sinai Health System, AI proved to be both the problem and the solution to getting more patients in the door.
Problem
Mount Sinai delivers world-class care in New York City, but so do several other health systems in the area. Meanwhile, AI overviews mean fewer people are going to the Mount Sinai site for answers to basic health questions.
But Davey’s team saw that traffic continued to increase for content like doctor profile pages and appointment information. Ensuring physicians, locations, and service lines appear as trusted answers in Google search and AI overviews is more important than ever.
Solution
Davey outlined a pilot in partnership with Yext focused on cardiology provider profiles. Yext is a platform for managing structured location data in an AI-optimized format. Mount Sinai’s goal was to improve visibility for its cardiology team by about 10 percent in search results, measured across both Google and AI rankings. “We want to see what’s working for one service line’s provider profiles and then scale it up. If you put it all together correctly, it’s a grand slam home run,” he says.
They improved profile attributes and listings to increase review invites and responses and then measured the impact. Businesses with more frequent and high-quality reviews rank higher in Google Search.
Yext helped Mount Sinai focus on where to prioritize its time and resources, and measure results.
Results
To measure results, the team focused on two key indicators:
- Google Rank: The average position providers appear in Google search results
- AI Rank: How prominently providers are featured in AI overviews
By focusing on increasing the volume and quality of reviews, Mount Sinai hoped to rise above the competition in search results. Over a three-month period, the “review score” for this provider nearly doubled.
Finally, Davey summarizes what’s true for all three healthcare leaders. “Data is great, but it’s not just for running reports. You have to determine the drivers of the data and then decide what actions to take to move your organization above the others.”
5 Practical Tips for Healthcare Marketers
- Start with intuition and use AI to interpret data that will help you validate your strategy.
- Use data you already have, such as patient reviews, call center notes, site search logs, and MyChart messages. These can serve as digital focus group, if you use them safely and responsibly.
- Use AI tools with guardrails. Ensure you have HIPAA-compliant secure environments, define what data can be used where, and require human review for anything public-facing.
- Go hyper-local where it matters. Zip Code and neighborhood-level strategy can sharpen budgets and channel mix, especially in competitive markets.
- Move from insights to action fast. AI doesn’t create ROI, but execution does. Tie outputs to specific changes you can implement and measure and then iterate.
As owner of Sparkr Marketing, Wendy Margolin helps busy healthcare marketing communications teams create more content. She’s on a mission to build a better medical web, one article at a time. Her favorite form of content is hospital brand journalism, which ties together her 20-year career in journalism, marketing, and healthcare.