Zocdoc Opens Booking Rails to AI Platforms and Health Insurers
Julie Yoo, Oliver Kharraz
Transactional scheduling cuts patient wait times from a month to under 72 hours by capturing perishable cancellation slots that static provider directories miss.
AI-assisted summary of Why AI Could Finally Fix Healthcare Access | Zocdoc CEO Oliver Kharraz — original source. Check the original for full context.
Source published:
Key ideas
- Zocdoc is opening its scheduling infrastructure to third parties, allowing payers, search engines, and AI platforms like Google Gemini to embed real-time doctor booking directly into their own interfaces. (0:00)
- While national doctor wait times have grown from 20 to approximately 31 days, Zocdoc averages under 72 hours by filling last-minute cancellations and backfilling unused schedule capacity. (26:45)
- Zocdoc maintains over 99% accuracy in provider directory data by using active booking transactions as a continuous, Wikipedia-style verification loop across more than 175 EHR integrations. (2:40)
- Under its discovery-based business model, cleared by healthcare regulators, Zocdoc charges providers solely for net-new discovery bookings, while direct name searches and existing patient bookings remain free. (23:08)
Detailed notes
## Background Published on October 17, 2026, by a16z General Partner Julie Yoo and Zocdoc CEO Oliver Kharraz via RSS and web podcast feeds, this episode examines how AI and open infrastructure can dismantle long-standing barriers to healthcare access. The show description highlights Zocdoc's two-decade effort to build scheduling plumbing, its partnership with Google Gemini, and why the company is opening its transactional booking rails to third-party AI agents, insurers, and external consumer platforms.
## TL;DR Oliver Kharraz outlines Zocdoc's strategic transition from a standalone consumer marketplace to a foundational access layer powering third-party AI agents, health plan directories, and web portals. He explains that provider directory accuracy depends on continuous transactional verification rather than static rosters, solving messy edge cases across more than 175 disparate EHRs. The discussion reveals how capturing hidden, perishable capacity—such as last-minute cancellations—can compress average national appointment wait times from over 30 days down to under 72 hours. Ultimately, Kharraz and Yoo make the case that healthcare should function like an open city rather than a walled garden, with Zocdoc providing the universal pipes that allow any patient search to terminate in a confirmed booking.
## Atmosphere & style The conversation carries the warm, easy banter of two seasoned health-tech veterans who have spent decades tackling the industry's thorniest structural problems. Pragmatic engineering war stories—like discovering clinic staff using Post-it notes on monitors instead of updating EHR calendars—ground the high-level strategy in relatable operational realities. The dialogue balances visionary optimism about conversational AI with disciplined skepticism toward ungrounded tech hype, leaving listeners with an insider's view of how persistent, incremental technical plumbing actually drives systemic change.
## Bulletin - `00:00` Zocdoc is opening its scheduling infrastructure to third parties, allowing payers, search engines, and AI platforms like Google Gemini to embed real-time doctor booking directly into their own interfaces. - `26:45` While national doctor wait times have grown from 20 to approximately 31 days, Zocdoc averages under 72 hours by filling last-minute cancellations and backfilling unused schedule capacity. - `02:40` Zocdoc maintains over 99% accuracy in provider directory data by using active booking transactions as a continuous, Wikipedia-style verification loop across more than 175 EHR integrations. - `23:08` Under its discovery-based business model, cleared by healthcare regulators, Zocdoc charges providers solely for net-new discovery bookings, while direct name searches and existing patient bookings remain free.
## Key takeaways - `00:00` Zocdoc has shifted strategically from being solely a destination consumer marketplace to functioning as open scheduling infrastructure for the broader healthcare ecosystem. - `02:40` Provider directory accuracy remains broken in static publishing models, but active transaction loops allow Zocdoc to maintain over 99% data accuracy. - `04:47` Electronic Health Record (EHR) integrations alone do not guarantee scheduling accuracy because practice staff often rely on physical workarounds like sticky notes. - `07:05` Healthcare functions like an open city rather than a closed building, making efforts by health systems to build proprietary walled gardens counterproductive. - `10:00` Data shows patients exhibit 75% to 80% organic repeat booking loyalty with the same specialist, proving that artificial system lock-in is unnecessary to retain patients. - `12:02` Conversational AI agents serve as personalized front doors that synthesize medical information, requiring foundational transaction infrastructure like Zocdoc to convert intent into booked care. - `14:45` Zocdoc treats autonomous AI agents as direct extensions of individual patients, delivering unfiltered available scheduling options rather than applying algorithmic gatekeeping. - `16:18` High-acuity specialties have widely adopted open digital booking, with tens of thousands of oncology visits booked in New York City alone. - `20:00` Embedding direct booking into payer directories, such as Blue Shield of California, cuts costs by diverting non-emergent visits away from emergency departments that are 10 to 12 times more expensive. - `23:08` Zocdoc transitioned from a flat subscription to a regulatory-compliant model where providers pay only for discovery bookings, with 97% of new clinicians joining via word-of-mouth. - `26:45` Doctors operate at load factors under 70%, allowing dynamic scheduling to capture perishable capacity and cut patient wait times to under 3 days. - `30:00` Top-tier clinicians frequently experience the lowest schedule utilization because they refuse to double-book, making automated backfilling of last-minute cancellations essential. - `30:00` Zocdoc enforces quality and credentialing standards that are more selective than state medical boards, actively de-listing non-compliant providers. - `31:45` Zocdoc's ultimate objective is to provide public scheduling APIs so that every emerging digital front door can seamlessly connect patients to care.
## Detailed walkthrough ### `00:00` Evolving Zocdoc from a Destination Site to Open Scheduling Infrastructure
Booking a flight or buying an item online takes seconds, yet booking a simple doctor's appointment in healthcare can still stretch across weeks. Julie Yoo, General Partner at Andreessen Horowitz (a16z), welcomes Zocdoc founder and CEO Oliver Kharraz to reflect on their decades-long shared history in health tech. The two joke about being "kindred spirits" who crossed paths and occasionally competed when they probably should have partnered. Yoo emphasizes that Zocdoc is one of the rare health tech businesses to establish itself as a true household consumer name.
Kharraz provides an update on where Zocdoc stands today. While most consumers still know Zocdoc as a destination website and marketplace where patients discover in-network doctors and book appointments immediately, the company has recently made a major strategic shift: unlocking the underlying infrastructure for third parties. Rather than restricting access to the Zocdoc consumer app, the underlying booking pipe is being unbundled so that health insurers, AI platforms, and consumer ecosystems can let patients book doctor visits directly from wherever they initiate their search. As Kharraz explains:
> “So whether you are a payer, you're an AI platform, you have a consumer audience, you can now give them the ability to schedule their doctors seamlessly, no matter where they start the journey.”
`02:40` Why the Provider Directory Problem Persisted—and How Transactability Solved It
The dialogue was sparked by a text exchange after Yoo posted a tweet noting that nothing in life is certain except "death, taxes, and the provider directory problem." Yoo notes that both early-stage startups and legacy incumbents still fail to maintain accurate information on which providers are bookable, where they practice, and what insurance plans they take.
Kharraz responds with a bold perspective: the directory problem is already largely solved if you approach it through the lens of transactionality. Zocdoc maintains scheduling directory data that is over 99% accurate, which contrasts sharply with standard, error-ridden payer directories. Kharraz compares legacy directory maintenance to compiling a static reference work:
> “When you put out a provider directory, frequently you're in the business of publishing the Encyclopedia Britannica, right? By the time you're able to really put it out there and present it to patients, it's already out of date.”
By contrast, an active transactional model works like Wikipedia: every completed booking acts as an active verification loop. If a mistake occurs, it surfaces and gets corrected on that single transaction, continuously updating the underlying record.
```mermaid flowchart TD accTitle: Static vs. Transactional Provider Directories accDescr: Comparing out-of-date static directory publishers with Zocdoc's real-time transaction verification feedback loop.
subgraph Static["Static Directory Model"] A["Collect Roster Data"] --> B["Publish Directory"] B --> C["Data Rapidly Becomes Stale"] C -.->|"Lacks Verification"| A end
subgraph Transactional["Transactional Model (Zocdoc)"] D["Active Booking Attempt"] --> E["Sync with EHR Schedule"] E --> F{"Appointment Booked?"} F -->|"Success"| G["Data Validated (>99% Accurate)"] F -->|"Failure / Mismatch"| H["Fix Edge Case Immediately"] H --> G G --> D end ```
Achieving true transactability required solving an extraordinarily complex technical "Tetris game" across the ecosystem: - Managing native integrations across hundreds of disparate Electronic Health Record (EHR) systems. - Accurately cross-referencing and mapping tens of thousands of complex insurance plans. - Adapting to unique practice workflows, office variations, and specialty-specific scheduling rules over a 20-year development effort.
`04:47` Cracking the EHR: Post-It Notes and the "1,000 0.1% Solutions"
Yoo breaks down physician data into three distinct layers based on her experience building Kyruus: 1. **Demographic / Administrative:** Medical school, practice location, board certifications, and payer networks. 2. **Clinical "Power" Data:** Real clinical scope, procedural volume, and referral relationships. 3. **Real-Time Availability:** Direct pipes into scheduling systems of record to route patients based on true capacity.
Kharraz confirms that while demographic data is fairly static and straightforward to collect, clinical power data and live calendar capacity must both be pulled from the EHR—and they resist easy standardization. While Zocdoc provides out-of-the-box integrations for over 175 different EHR platforms, the software is deployed and manipulated in drastically inconsistent ways across clinical staff.
To illustrate, Kharraz shares an anecdote about an evasive scheduling glitch. Zocdoc's system reported that a doctor was unavailable when a patient attempted to book, yet the practice's digital calendar and Zocdoc's digital replica appeared identical. After gathering engineering leads in a room, ordering pizza, and offering to walk their dogs to crack the mystery, Kharraz finally went into the practice and sat next to the office manager:
> “And what I realized is that sometimes the doctor would walk out and say, ‘Hey, I'm out next Friday,’ and the office manager would actually take a Post-it note and attach that to their monitor instead of going into the EHR.”
Because clinical workflows often bypass the EHR in favor of physical shortcuts, Zocdoc had to build predictive, mathematical models around real-world behavioral quirks. Kharraz emphasizes that making scheduling truly reliable did not stem from a single breakthrough, but from discovering and engineering "1,000 0.1% solutions." Yoo quips that in the early days of healthcare software, 3M felt like the industry's biggest competitor because of sticky notes.
`07:05` The "City" vs. the "Building": Abandoning Healthcare's Walled Gardens
Yoo highlights Zocdoc's identity as a consumer-first brand, contrasting it with B2B enterprise plays that prioritize hospital systems and health plans. Kharraz argues that beginning with the patient aligns incentives across the board, even if legacy enterprise software pushes the opposite direction.
Traditional health system strategies often rely on closed ecosystems designed to trap patients within their owned networks. Kharraz firmly rejects the premise that an organization should attempt to "own" the patient or monopolize the entryway to medicine:
> “Because healthcare isn't like a building where you lure the patient in and you slam the door shut behind them. It's a lot more like a city, right? And the patient walks the city and depending on what their need is at this current moment, they might knock on different doors.”
In Kharraz's view, patients naturally move across distinct points of access: - Visiting an urgent care clinic when a child is abruptly sick. - Seeking out a world-class subspecialist for a rare, complex diagnosis. - Finding an empathetic, accessible primary care doctor for longitudinal chronic disease management.
Attempting to restrict care through a single proprietary front door creates friction and blinds systems to patient needs. Instead, opening all the doors to healthcare allows technology providers to extract scheduling insights that no single EHR or health system could ever capture on its own.
`10:00` Consumer Loyalty and the Myth of Health System Lock-In
Julie Yoo asks Oliver Kharraz to reconcile two competing perspectives: health systems wanting to keep patients locked inside their own closed networks versus patients seeking the best care across a holistic set of options. Given that Zocdoc sits on millions, if not billions, of patient interaction data points, she asks what the empirical evidence shows about patient loyalty and whether patients actually achieve better outcomes when they are not artificially tethered to a single institution.
Oliver resets the conversation by comparing healthcare to everyday consumer expectations. In other industries, transacting is frictionless: “In e-commerce, you buy. In travel, you book. In healthcare, you're blocked.” He emphasizes that healthcare uniquely erects barriers that make it exceedingly difficult for patients to do business with providers, running counter to how people naturally make decisions. When patients search for care, they exhibit high-consideration behavior—carefully evaluating multiple providers, reading reviews, and checking credentials. Forcing them into an artificially narrow aperture or imposing friction simply causes a health system to fall off their radar entirely.
Counterintuitively for hospital administrators obsessed with retention, Zocdoc’s data demonstrates that once a patient finally selects a provider, they show immense organic loyalty without needing artificial software barriers. Specifically, Zocdoc sees a 75% to 80% rebooking rate with the same provider whenever a patient needs follow-up care in that same specialty. As Oliver puts it: “So it speaks to the fact that it's not really necessary to lock them in through the systems you build. You will gain their loyalty just by being a good doctor to them.”
- **E-commerce vs. healthcare friction:** Standard consumer interactions are immediate, whereas healthcare actively blocks patient engagement through administrative hurdles. - **High-consideration search:** Patients naturally conduct extensive diligence—browsing multiple profiles and reviews—before selecting a clinician. - **Empirical loyalty metric:** Zocdoc tracks a 75% to 80% repeat booking rate with the same specialist, demonstrating that retention stems from clinical quality rather than network lock-in.
---
`12:02` Becoming the Infrastructure for AI Front Doors
Julie shifts the discussion to the emergence of consumer copilot agents—pointing out that the industry is several weeks into the mainstream adoption of personal AI tools and assistants that consumers use to book travel, reserve tables, and navigate complex tasks. She asks Oliver how these emerging autonomous agents alter Zocdoc’s role in the ecosystem, and how Zocdoc plans to help providers manage the incoming wave of automated inquiries.
Oliver explains that conversational AI is driving patient engagement to unprecedented levels because it provides personalized, highly contextual health information that feels immediately relevant and spurs patients to take action. However, an informational response alone cannot deliver care. That bridge is where Zocdoc positions itself: “The AI can give you the information, but Zocdoc can make it actually actionable so that you can go see a provider.”
```mermaid flowchart LR accTitle: Zocdoc as Healthcare Access Infrastructure accDescr: Providers connect once to Zocdoc, which routes real-time availability out to various AI agents and platforms where patients search.
subgraph Supply["Provider Side"] P["Provider Schedules / EHR"] end
subgraph Hub["Infrastructure Layer"] Z["Zocdoc Integration Engine"] end
subgraph FrontDoors["Exploding Front Doors"] A1["Google Gemini"] A2["Consumer AI Agents"] A3["Zocdoc App / Web"] end
subgraph Demand["Patients"] C["End User / Patient"] end
P -->|"Single Integration"| Z Z -->|"Live Availability"| A1 Z -->|"Live Availability"| A2 Z -->|"Direct Booking"| A3 A1 --> C A2 --> C A3 --> C ```
Rather than trying to force every patient into Zocdoc's own direct-to-consumer app, Oliver views Zocdoc as foundational healthcare access infrastructure. With front doors multiplying across platforms—highlighted by Zocdoc’s active partnership with Google Gemini—individual medical practices cannot build custom integrations for every new bot or agent. Zocdoc promises providers that a single integration will distribute their live calendar availability across every emerging platform: “We don't need every booking to happen on Zocdoc.com or our app. We just want to make sure that whatever door the patient is knocking on, that door actually opens.”
- **The role of AI vs. Zocdoc:** Generative AI personalizes and synthesizes medical information, while Zocdoc provides the transactional layer that converts intent into a confirmed appointment. - **Partner ecosystem:** Zocdoc is partnering with major platforms like Google Gemini to embed real-time bookability directly within AI chat interfaces. - **The distribution promise:** Instead of managing fragmented third-party integrations, clinics plug into Zocdoc once to syndicate their schedules across all AI search interfaces.
---
`14:45` Routing, Patient Agency, and Treating AI as an Extension of the User
Julie explores the governance and fairness challenges of moving from an application layer down to an infrastructure utility. On Zocdoc’s consumer app, users retain full agency: they can sort through seven or eight search results, filter by their personal criteria, and make an autonomous choice. But when an AI agent acts as an intermediary, she asks whether Zocdoc will maintain an objective fairness layer or if it will have to apply judgment and preference logic when routing bookings back to third-party tools.
Oliver responds that Zocdoc has always anchored its design around giving full agency to the patient, presenting unfiltered available options and letting the consumer decide what constitutes the right match. In an agentic world, Zocdoc sees autonomous software not as an independent gatekeeper, but as a direct representative of the individual: “And in our view, you know, these AI agents are just going to be extensions of the patients, right?”
Because each personal agent will gather and apply that specific human's preferences, constraints, and decision criteria, Zocdoc draws no distinction between a human user and an AI agent querying its scheduling APIs. Zocdoc's responsibility remains surfacing the complete set of matched, bookable slots and letting the patient—or the agent executing the patient’s will—make the final selection.
---
`16:18` Expanding into High-Acuity Subspecialties and Clinical Qualification
Julie pivots to the supply side of the marketplace. In Zocdoc’s early years, adoption centered primarily on lower-acuity, transactional specialties such as primary care, dental, and urgent care. Back then, highly specialized practices were culturally hesitant to embrace open digital scheduling. Julie asks what shifted across the market and within provider psychology to persuade complex medical practices to make their schedules openly bookable.
Oliver counters that Zocdoc’s breadth across subspecialties has been established for years, even if primary care and dentistry naturally see higher everyday booking volumes than neurosurgery. Zocdoc currently supports over 250 distinct specialties. To illustrate the scale of high-acuity adoption, Oliver shares a striking metric from an internal meeting earlier that day: “I just came out of a meeting where I learned that we have literally tens of thousands of oncology bookings going through Zocdoc in New York City alone.” Patients turn to Zocdoc for everything from routine dental hygiene to acute cancer care, largely because it eliminates the painful alternative of calling clinic reception desks and playing what he terms "scheduling Tetris" with an office manager.
Addressing historical concerns that high-acuity specialties require rigid clinical triage before an appointment slot can be assigned, Oliver notes that progress has occurred on two fronts: 1. **Consumer judgment and referrals:** Providers have recognized that patients exercise remarkably sound judgment regarding their clinical needs. Most complex specialty bookings are preceded by a formal referral from a primary care doctor, meaning the patient uses Zocdoc not to guess a diagnosis, but to research, vet, and "pre-meet" the specialist online. 2. **Technological filtering:** Intelligent matching and filtering tools on the platform have advanced substantially, guiding consumers to the right provider without relying on the brittle, antiquated scheduling rule sets traditionally hardcoded into legacy electronic health records (EHRs).
---
`19:24` The Next Generation of AI-Native Practice Management and EHRs
Julie turns to what many healthtech founders consider the industry's ultimate holy grail: building a truly AI-native electronic health record (EHR) and practice management system. Recognizing that Zocdoc has steadily expanded upstream and downstream from its core search marketplace—vertically integrating deeper into the administrative and workflow layers of medical practices—she asks Oliver to explain his vision for Zocdoc’s role in modernizing practice operations and whether the company aims to redefine core clinical software.
`20:00` Expanding Beyond Zocdoc.com and Partnering with Payers
Julie Yoo asks whether Zocdoc ultimately envisions itself becoming a system of record for scheduling and practice administrative tasks, or how it navigates operating alongside deeply entrenched incumbent software platforms in healthcare. Oliver Kharraz explains that Zocdoc is guided primarily by consumer needs in an industry notorious for neglecting the patient experience. Historically, the company operated under the assumption that every interaction had to take place directly on Zocdoc.com. However, Kharraz notes that the company underwent a major philosophical shift and now builds infrastructure that embeds directly into existing provider workflows—including software that powers online scheduling on a provider’s own homepage, tools deployed inside provider call centers, and internal routing products designed to eliminate operational friction across various modalities.
This omnichannel approach extends into the payer landscape, where health plans historically struggled with inaccurate, static member directories. Julie points out that insurance is often the "tail that wags the dog," since securing an appointment is meaningless if a patient discovers the visit will not be reimbursed. Kharraz points to Zocdoc’s partnership with Blue Shield of California as a prime example of their healthcare access infrastructure at work: members browsing Blue Shield's directory can directly book confirmed appointments on the spot. For health plans, instant access provides an immediate financial and clinical return by directing members to outpatient clinics and keeping them out of emergency rooms, which are ten to twelve times more expensive than ambulatory care and frequently yield worse health outcomes. Kharraz reflects on how quickly payer adoption escalates once integrated:
> “What was interesting for us on this journey is actually that the amount of change on the payer side is relatively minimal, right? They can now plug into our healthcare access infrastructure. So once they realize that that's happening, you start with one area in a payer and then suddenly other people come and say, can I use this also? And I have this use case and this other use case. And pretty quickly, you're working with the entire organization, which is very, very exciting to us.”
```mermaid flowchart TD accTitle: Zocdoc Integration Touchpoints accDescr: How Zocdoc infrastructure connects patients, providers, and payers across channels.
subgraph Channels["Patient Entry Points"] A["Zocdoc.com & App"] B["Provider Homepages & Call Centers"] C["Payer Member Directories<br/>(e.g., Blue Shield of CA)"] end
subgraph Core["Zocdoc Access Infrastructure"] D["Real-Time Scheduling & Verification Engine"] end
subgraph Outcomes["Downstream Impact"] E["Timely Outpatient Care"] F["Diversion from High-Cost Emergency Rooms<br/>(10x–12x Cost Savings)"] end
A --> D B --> D C --> D D --> E E --> F ```
- **Payer integration footprint**: Embeds directly into insurer directories (such as Blue Shield of California) with minimal IT overhead for the health plan. - **Cost avoidance**: Ambulatory care redirection diverts non-emergent visits away from emergency departments that cost 10 to 12 times as much.
---
`23:08` Regulatory Clearance and the Discovery-Based Business Model
Julie probes the mechanics of Zocdoc’s three-sided marketplace, asking how the company monetizes relationships across consumers, providers, and payers, particularly given stringent healthcare regulations like the federal Anti-Kickback Statute and the Stark Law. Kharraz recounts the company's early days, when Zocdoc operated strictly as a flat-rate subscription service charging every clinician the same monthly fee simply to be listed. That uniform pricing proved suboptimal—prohibitive for smaller or lower-volume practices while vastly undercharging high-volume clinics. To transition to a value-aligned pricing model without running afoul of federal kickback prohibitions on referral fees, Zocdoc engaged directly with healthcare regulators to demonstrate that facilitating open appointment access actively served patients, doctors, and payers alike.
The resulting regulatory clearance enabled Zocdoc’s modern business model: charging providers strictly for a "discovery booking," where a patient searches generally and gets matched with a new clinician. If a patient is simply searching for their existing doctor by name and books an appointment, Zocdoc provides that booking software to the provider completely free of charge. By stripping out upfront financial barriers, practices can join the network without financial risk. Kharraz notes that the model has sparked organic momentum across the clinical community, explaining:
> “One stat that I'm particularly proud of is that 97% of the providers we signed up, you know, that's word of mouth. Their friends tell them this is a system that really works. You should join too.”
On the health plan side, monetization is structured differently depending on the specific enterprise contract. Payers pay to unlock turnkey access to Zocdoc’s entire scheduling network and directory tools, enabling their members to instantly book validated in-network providers with verified insurance participation, eliminating billing surprises and out-of-network leakage.
- **Discovery vs. direct booking**: Providers pay only when Zocdoc introduces a net-new patient via discovery searches; bookings for patients searching for a specific doctor by name are 100% free. - **Provider acquisition channel**: 97% of newly onboarded providers join via peer-to-peer word-of-mouth recommendations rather than aggressive outbound sales. - **Payer monetization**: Enterprise agreements grant health plans full network access, automated real-time insurance verification, and embedded directory scheduling.
---
`26:45` Cracking the Capacity Problem: Slashing Wait Times to Under 72 Hours
Julie raises the chronic issue of appointment backlogs, citing that when she previously worked at Kyruus, the average wait time for a primary care doctor in Boston hovered around 45 days. She asks whether national wait times have improved and what drives capacity bottlenecks. Kharraz shares a grim reality: across the broader United States healthcare landscape, appointment availability has actually deteriorated over time. While the national average wait time sat around 20 days when Zocdoc first launched, it has since ballooned to approximately 31 days.
In stark contrast, patients booking through Zocdoc regularly see a doctor within 72 hours—frequently within three days or less across a wide variety of medical specialties. Kharraz explains that this dramatic difference does not stem from creating new doctors out of thin air, but from systematically capturing hidden, perishable capacity. Medical practices experience a relentless churn of last-minute cancellations, no-shows, and reschedulings that conventional scheduling fails to backfill. Drawing a historical parallel to the travel industry, Kharraz points out:
> “If we compare that with airplanes, right? Like airplanes used to fly 60% full, but that was in the 70s. With the advent of the internet, we had the Kayaks and the Expedias and the Bookings.com, and now airplanes have a load factor of over 90%. In the wild, doctors have load factors around 70%, just under 70%. Now, with Zocdoc, we fill that last-minute capacity, and this is how we can get to the three-day average time.”
Kharraz emphasizes that Zocdoc dynamically manages its provider supply to maintain this standard: whenever appointment inventory drops within a specific specialty, geographic zip code, or insurance plan quadrant, network teams actively recruit local providers to replenish open slots. Julie then counters with a common consumer skepticism: if a physician has an open slot available within three days, doesn't that imply they are a low-quality doctor whom nobody else wants to see? Kharraz prepares to debunk that perception by detailing the statistical mechanics of provider scheduling and clinical quality.
| Metric | Traditional Healthcare System | Travel Industry (1970s vs. Modern) | Zocdoc Network | |:--- |:--- |:--- |:--- | | **Average Appointment Wait Time** | ~31 days nationally (up from ~20 days; up to 45 days in Boston) | N/A | **≤ 3 days (under 72 hours)** | | **Capacity Utilization / Load Factor** | Just under 70% (lost to cancellations, no-shows, and gaps) | 60% in the 1970s → **>90% today** via aggregators (Kayak, Expedia) | Near-optimal utilization via real-time last-minute backfilling | | **Network Optimization Strategy** | Passive waitlists and manual phone rescheduling | Dynamic revenue management algorithms | Proactive provider recruitment targeted by zip code, specialty, and insurance quadrant |
`30:00` Counterintuitive Utilization, Last-Minute Cancellations, and Provider Vetting
Julie Yoo explores the operational mechanics of platform accountability, pressing Oliver Kharraz on whether Zocdoc enforces strict service level agreements (SLAs) or de-lists providers who fail to demonstrate satisfactory patient outcomes or honor schedule commitments. Kharraz responds that quality control operates firmly on both sides of the marketplace, highlighting a paradoxical dynamic in physician scheduling. High-demand clinicians frequently struggle with schedule efficiency because their patient demographics lead to unpredictable churn. As Kharraz explains: “So one kind of counterintuitive bit is that actually the best doctors tend to have the worst utilization because they frequently treat patients that have very busy lives and cancel on them more often and they don't double book.”
To illustrate how the platform salvages this wasted capacity, Kharraz shares an anecdote involving the dean of a prominent academic medical center in New York. The physician initially hesitated to join the platform, telling Kharraz that because his calendar was already fully booked three months into the future, listed availability on Zocdoc seemed unnecessary. Kharraz persuaded him to test the system anyway, and the integration captured roughly 40 appointments that would have otherwise expired unfilled due to standard last-minute patient cancellations. Instead of empty clinic hours, patients desperate for expedited care secured prompt visits, and the dean achieved substantially higher operational utilization—keeping him active on Zocdoc to this day.
Regarding network standards, Kharraz stresses that Zocdoc does not passively accept anyone holding an active medical license. “In terms of who gets to be on ZocDoc, we are more selective than your neighborhood state board, we have our own qualification entities and rules,” he explains, noting that the company maintains proprietary credentialing standards and routinely expels doctors who fail to satisfy those requirements.
```mermaid flowchart TD accTitle: Salvaging Doctor Schedule Utilization accDescr: Flowchart showing how Zocdoc captures last-minute cancellations from fully booked doctors to serve urgent patient demand.
subgraph Clinic["Provider Schedule Dynamics"] Dean["High-Demand Specialist / Academic Dean<br>Booked 3 Months Out"] Cancel["Busy Patients Cancel Last-Minute<br>No Double-Booking Policy"] Dean --> Cancel end
subgraph Platform["Zocdoc Integration"] Sync["Automated Schedule Sync<br>& Qualification Vetting"] Match["Instant Availability Surfacing"] Cancel --> Sync Sync --> Match end
subgraph Demand["Patient Access"] UrgentPatients["Patients Needing Prompt Care<br>~40 Appointments Preserved"] Match --> UrgentPatients end ```
- High-demand providers often suffer from poor utilization because they refuse to double-book, leaving them exposed when patients with hectic schedules cancel on short notice. - A New York academic medical center dean who was booked three months out recovered approximately 40 appointments through last-minute Zocdoc bookings. - Zocdoc enforces internal qualification rules and vetting criteria that exceed basic state licensing board requirements, de-platforming providers who breach standards.
---
`31:45` Opening Every Front Door in Healthcare: Infrastructure, AI, and the Future
Looking toward the horizon, Yoo asks Kharraz to outline Zocdoc’s forward-looking product roadmap and share what anxieties keep him awake at night after decades of evolving the business model and expanding market reach. Kharraz deflects the existential worry with humor, quipping, “Oh, my superpower is sleeping.” Turning serious, he articulates the company’s core long-term objective: converting Zocdoc’s proprietary plumbing into an open, universal healthcare access infrastructure. Rather than forcing all digital interactions into the standalone Zocdoc app, the company aims to ensure that scheduling functionality seamlessly powers every point of patient origination.
Kharraz summarizes this cross-platform ambition: “Our goal is to connect every single provider and make sure that every single front door in healthcare opens, right?” He envisions an interoperable ecosystem that equips external tech ecosystems to book appointments natively: “This is why we've opened up our healthcare access infrastructure so that every payer, every AI platform, every consumer platform, review platform, search site, they can connect their patients directly to a doctor and it will work and it will work well.” For Kharraz, solving this coordination problem represents a lasting, tangible contribution to American healthcare delivery and overall societal welfare.
In his closing reflections, Kharraz emphasizes that realizing this vision depends on talent, noting that Zocdoc is actively hiring. While generative AI will make workers significantly more productive, he maintains that high-caliber human engineers and operators remain indispensable to building reliable infrastructure. For emerging startups building in digital health or generative AI, Yoo and Kharraz confirm that access to Zocdoc's scheduling APIs is publicly available via direct inquiry on their website. Yoo congratulates Kharraz on the company's sustained longevity and impact across the evolving healthcare landscape.
```mermaid flowchart LR accTitle: Zocdoc Open Access Infrastructure accDescr: Architecture diagram showing diverse front-door channels connecting through Zocdoc APIs to underlying provider schedules.
subgraph FrontDoors["Universal Front Doors"] AI["AI Agents & Chat Platforms"] Payers["Health Plan Directories"] Search["Search Engines & Review Sites"] end
subgraph Core["Access Layer"] API["Zocdoc Open Infrastructure & APIs"] end
subgraph Delivery["Care Delivery"] EHR["EHR Integrations & Provider Calendars"] end
AI --> API Payers --> API Search --> API API --> EHR ```
- Zocdoc’s central strategic objective is opening its scheduling infrastructure to third-party search engines, consumer review sites, health plan directories, and conversational AI agents. - The company is actively recruiting, driven by the philosophy that human talent augmented by AI is critical to scaling healthcare access tools. - Early-stage startups and external developers can directly apply for and integrate Zocdoc’s scheduling APIs through the company's website.
## Notable quotes > `02:40` When you put out a provider directory, frequently you're in the business of publishing the Encyclopedia Britannica, right? By the time you're able to really put it out there and present it to patients, it's already out of date. — Oliver Kharraz
> `30:00` So one kind of counterintuitive bit is that actually the best doctors tend to have the worst utilization because they frequently treat patients that have very busy lives and cancel on them more often and they don't double book. — Oliver Kharraz
> `12:02` We don't need every booking to happen on Zocdoc.com or our app. We just want to make sure that whatever door the patient is knocking on, that door actually opens. — Oliver Kharraz
> `12:02` The AI can give you the information, but Zocdoc can make it actually actionable so that you can go see a provider. — Oliver Kharraz
## Mentioned Zocdoc, Oliver Kharraz, Julie Yoo, Andreessen Horowitz, Google Gemini, Blue Shield of California, Kyruus, 3M, Kayak, Expedia, Booking.com, Anti-Kickback Statute, Stark Law
## Video description Why is booking a doctor's appointment still so difficult, and can AI finally change that? Zocdoc CEO Oliver Kharraz argues that the real problem isn't finding doctors, but connecting patients to accurate availability and making appointments bookable wherever they start their search. Oliver joins a16z General Partner Julie Yoo to discuss how Zocdoc spent nearly 20 years building the infrastructure behind online scheduling, and why it's now opening that technology to AI agents, insurers, and other platforms. They explore Zocdoc's partnership with Gemini, why patients shouldn't be locked into a single healthcare platform, and how surfacing last-minute availability can reduce appointment wait times from weeks to days. They also discuss what happens when AI becomes a new front door to healthcare and why Oliver believes the future of Zocdoc extends far beyond its own app.
[Source](https://mgln.ai/e/1344/afp-848985-injected.calisto.simplecastaudio.com/3f86df7b-51c6-4101-88a2-550dba782de8/episodes/60df6c4d-a29a-4c66-994e-bbd659a2e577/audio/128/default.mp3?aid=rss_feed&awCollectionId=3f86df7b-51c6-4101-88a2-550dba782de8&awEpisodeId=60df6c4d-a29a-4c66-994e-bbd659a2e577&feed=JGE3yC0V)