UC San Diego Health × Eureka Media

One intelligent network.
Every specialist. One front door.

A working concept for Dr. Khalessi.

Press → to advance · press N for the talk track

The problem, and the leak

When a patient can't be seen fast, they leak out to a competitor.

California says a non-urgent specialist visit should be offered within 15 business days. Nationally, the average new-patient wait is now 31 days and rising, up 19% since 2022. A directory of 2,000 names does not fix that. Patients wait, give up, or leave, and every one who walks takes years of downstream revenue with them.

Sources: California DMHC Timely Access to Care, 28 CCR 1300.67.2.2 · AMN Healthcare, 2025 Survey of Physician Appointment Wait Times (San Diego is one of 15 markets surveyed).

Why now

Patients now find doctors through AI.

36%

chose a doctor influenced by AI like ChatGPT, double the year before.

37%

trust Google's AI answer most, ahead of links, the map, or ads.

75%+

of AI health citations point to established institutions. That is you.

An academic medical center is built to be the answer AI gives, if the front door is intelligent.

How patients discover you now

Search split into three. You have to win all three.

SEO

Search Engine Optimization

Rank in Google's classic results. Still where most patient searches begin.

"spine surgeon near me" → you are on page one.

AEO

Answer Engine Optimization

Be the answer Google gives directly: AI Overviews, featured snippets, People Also Ask, and voice.

"who treats scoliosis in San Diego?" → you are the answer, not a link.

GEO

Generative Engine Optimization

Be the name ChatGPT, Gemini, and Perplexity give when a patient asks who to see.

"best spine surgeon in La Jolla?" → the AI recommends you.

All three together are your share of voice: how often the answer a patient gets is UCSD, not a competitor. Patients increasingly get one answer, not a list, so for the first time every specialist has an agency in their corner watching and growing it.

The concept, on one slide

An intelligent front door, on top of Epic.

Patient

A problem

"My back hurts. Who do I see?"

→

Eureka layer

The intelligent front door

Find the right specialist · get seen soonest · keep them at UCSD

→

UC San Diego Health

Epic & MyChart

The record, scheduling, the care. The source of truth.

We never replace the record or hold it. When a visit is scheduled, it runs through Epic's own scheduling APIs, so Epic stays the source of truth.

Live · the network

We already built it: all 2,084 providers, one network.

Directory of 2,084 UC San Diego Health providers

Filter the full network by specialty, then see who can be seen soonest. Every provider is real, with a photo and verified rating, and a page built to be found by Google and cited by AI.

Open the live directory →

Live · a specialist page

Every specialist, a page that earns trust.

Credentials, conditions, and their real verified patient reviews (Dr. Osorio: 4.84 across 227 ratings). This is what builds patient trust, and what makes AI recommend them.

Open a live profile →

The star feature

We keep patients at UCSD, even when the wait is long.

Step 1 · Coverage check

Dr. Osorio

Spine & Neurosurgery

Insurance accepted · referral on file

Next opening: about 5 weeks

→

Step 2 · Sooner, keep Dr. Osorio

Cancellation waitlistkeeps your doctor

Sooner virtual visitkeeps your doctor

Care-team bridgekeeps your doctor

→

Step 3 · Route to a peer? Set per doctor

Dr. Osorio · subspecialistOff

Fields any specialist can coverOn

Complex cases stay with their surgeon. Where any qualified specialist can help, we show who is available soonest.

A long wait is the moment a patient leaks to a competitor. So we check coverage and offer sooner options that keep them with their own doctor. Whether we then offer another specialist is set per doctor: on where any qualified specialist can help, off for subspecialists so complex cases stay with their surgeon. Either way, they stay inside UCSD. Wayfinding, never a diagnosis.

Illustrative flow. Live availability and booking run through Epic's scheduling APIs under a BAA; Epic stays the source of truth.

The agent layer

Agents that answer, remind, and win patients back.

Inbound: the voice navigator

Answers every call and chat, 24/7. It speaks and listens, identifies itself as an AI, points patients to the right specialty, and hands the booking to Epic. This is the voice assistant patients meet on a profile.

42%

of calls to practices go unanswered, most during business hours.

67%

of unanswered patients call a competitor instead.

24/7

always on, it catches the after-hours and peak-time calls a front desk cannot.

Outbound: reminders, texts, win-backs

Calls and texts patients with appointment reminders and confirmations, and reactivates lapsed patients, all on its own.

−15pts

text reminders cut no-shows from 38% to 24% in a randomized trial (~$196 each).

98%

SMS open rate, ~45% reply, vs ~20% open for email.

15,000+

visits self-rescheduled by text, no staff involved, at one health system.

For UCSD, this replaces or shrinks an expensive, high-turnover call center. AI handles first contact for every specialty consistently, escalates to a senior human agent only when it should, and never misroutes, cutting cost and the errors a revolving-door front desk creates.

Every practice sees both: labor saved and more patients reached. Navigation and reminders only, never a diagnosis. The agent identifies as AI, runs under a BAA with minimum-necessary content and one-tap opt-out, and win-backs use a separate marketing consent.

Legal & compliance advisory

Built with health-law counsel.

Hal Katz

Partner, Husch Blackwell · Austin, TX

Board Certified in Health Law, Texas Board of Legal Specialization
Past Chair, American Bar Association Health Law Section
Ranked in Chambers USA and Best Lawyers for Health Care Law
Digital health, telemedicine, and healthcare regulatory counsel

He helped design the boundary in: non-diagnostic navigation, we never store the medical record, and scheduling and eligibility run only through Epic under a BAA.

Clinical decisions, the record, and the liability stay exactly where they belong, with your clinicians, in Epic. We grow patient volume on the safe side of that line, always reviewed by counsel.

The money

$200M+

leaks out every year. We keep it.

Only about a third of referrals ever reach a booked visit today. Recovering just 1% of leakage is roughly $2M a year, and every referral we keep seeds about $2.38M per physician in downstream care. Plus your physicians get time back, because reputation and patient inquiries get handled for them.

See it, and control it

Two dashboards: every doctor sees their practice, UCSD sees it all.

Each physician · operations first

Open a doctor dashboard →

UCSD leadership · money first

Open the network dashboard →

Every physician sees requests, reviews, and no-shows prevented. Leadership sees revenue recovered, capture, and where to act. This is what makes it measurable and yours to steer, and it stays non-clinical, never the record.

How we start

Start where there is no integration gate.

Phase 1

Own the front door

Now · nothing from your IT

The network & directory
Profiles & verified reviews
Navigation & AI visibility

Phase 2

Grow & capture

Next · via Epic, under a BAA

Reputation engine
Referring-physician program
"See sooner" scheduling

Phase one needs nothing from your IT team and stays entirely outside your systems of record. Phase two connects to scheduling only through Epic's own APIs under a BAA, with Epic still the source of truth.

Why you, why now

UCSD is built to win the AI front door. The window is open now.

You already made AI a core operating model, so you are ahead of your peers. Your academic authority means AI already wants to recommend you. The one piece missing is the patient-facing front door. That is the piece we bring, and it is a short window before every system races for it.

The ask

Let us prove, then scale.

Start focused: 3 to 4 specialties across different profiles, a high-volume referral engine (Orthopedics), a long-wait bottleneck (GI, Dermatology), a high-acuity surgical line (Spine), so the results generalize.

Launch

Front door live for the pilot specialties. Set the leakage baseline.

Optimize

Conversion, reputation engine, and AI + search visibility start compounding.

Prove

Referral-kept signal matures. Expand "see sooner" and referring-physician.

Scale

Full ROI incl. downstream revenue. Rollout plan to all 2,084.

A checkpoint at every step, so you can steer or stop at any gate. This is how we compound the value, not just test it.

Can we get a working session to pick the specialties?

Thank you

Thank you, Dr. Khalessi.

One network. Every specialist. One front door.

Eureka Media · let's pick the specialties and begin.