Rachell Dumas woke up blind.

Six months after finally delivering a healthy baby - the miracle at the end of nine pregnancy losses, five surgeries, and years of being dismissed by doctors - the registered nurse opened her eyes to darkness and stroke symptoms. She did what any medical professional would do: she went to the emergency room.

The ER doctor refused to order a CT scan. He told her she had a stomach bug and sent her home with a recommendation for Pepto-Bismol.

For acute vision loss.

“If I hadn’t been a nurse, if I hadn’t known to go to a different ER, if I hadn’t been able to speak the language of neurology and pull in favors from neurosurgeons I’d worked with - I would have died,” Dumas says. “And most people don’t have any of those things.”

The second emergency room ran a stroke workup immediately. Six months of advocacy, specialist consultations, and her own clinical research later, Dumas received a diagnosis of idiopathic intracranial hypertension - a rare neurological condition where cerebrospinal fluid accumulates in the brain without explanation. Two brain surgeries followed. At one point, so much fluid had compressed her cerebral vessels that she was at imminent risk of stroke.

It was between those two surgeries - lying in a hospital bed, recovering from the first, contemplating how close she’d come to becoming a statistic - that Dumas decided the healthcare system needed something it didn’t have: a tool that could give every patient the same fighting chance her medical expertise had given her.

She called it HEARD.

The numbers behind Dumas’s personal nightmare are staggering in their scope. More than 7 million patients are misdiagnosed in the United States every year. Diagnostic errors cost the healthcare system over $100 billion annually. And 250,000 people die from those errors - making medical misdiagnosis the third leading cause of death in America, behind heart disease and cancer.

Yet the infrastructure for patients to advocate for themselves in real time - while they’re still in the hospital, while the misdiagnosis is still happening - essentially doesn’t exist. Patients can file complaints after discharge. They can leave reviews. They can sue. But in the critical moments when intervention could actually change outcomes, the system offers almost nothing.

“People don’t survive their stay in the medical system,” Dumas says plainly. “They’re getting misdiagnosed more often than anyone realizes. I survived because I had three things: medical knowledge, professional connections, and the ability to find credible research. HEARD gives all three of those to everyone.”

HEARD - Human Enabled AI Reporting and Documentation - is a mobile application designed to function as a patient’s real-time advocate inside the healthcare system. The premise is deceptively simple: open the app, start a case, describe what’s happening. The execution is anything but.

When a patient enters their situation, HEARD’s AI engine processes the information and generates advocacy scripts - giving users the specific medical language and escalation pathways they need to push back against dismissal, question a diagnosis, or request appropriate testing. It’s the equivalent of having a seasoned nurse whispering in your ear, translating your symptoms into the clinical terminology that gets doctors to pay attention.

“I was pulling my neuro ICU expertise to talk about the left transverse intracranial whatever,” Dumas recalls of her own advocacy battles. “People don’t have that vocabulary. And in the medical system, vocabulary is survival.”

Beyond scripting, HEARD offers tiered escalation. If the situation warrants it, users can connect with a health advocate for a live consultation. In cases involving potential malpractice or negligence, the platform facilitates legal consultations. And for systemic failures, HEARD can automatically generate and route formal complaints to hospital administrators, the Department of Health, the Joint Commission, nursing boards, or medical composite boards - pulling contact information through an API that maps each hospital’s internal advocacy infrastructure.

The platform also includes a caregiver role, allowing family members to follow a patient’s healthcare journey in real time. It’s a feature born directly from Dumas’s experience: during her own medical crises, her family was often left in the dark, unable to advocate because they didn’t know what was happening behind the curtain.

If HEARD’s origin story is one of medical perseverance, its development story is a crash course in the volatility of today’s tech landscape.

Dumas secured a $50,000 grant to fund initial development and partnered with Amazon Web Services through its startup program to find a qualified development team. She found one - a firm with deep healthcare app experience, well-versed in HIPAA compliance and patient data security. Payments were made. Work began.

Then the team was laid off.

The wave of AI-driven layoffs sweeping the tech industry had reached healthcare development. Dumas found a second team through Amazon’s network. This one moved fast, building out HEARD’s backend architecture with multiple weekly check-ins. Then, over the holidays, the emails stopped. The messages went unanswered. Weeks passed. That team had been laid off too.

“It’s becoming more and more common,” Dumas says of the development instability. “Six months ago it wasn’t as bad in healthcare tech. But with these no-code platforms becoming more popular, the layoffs are hitting everywhere now.”

Rather than wait for a third team to materialize and potentially disappear, Dumas took a different approach. Working with Simply Built, she used Lovable - a no-code development platform - to create a fully functional prototype of HEARD that could serve as both a beta-testing vehicle and a definitive blueprint for her development team.

The strategic pivot to Lovable did something Dumas hadn’t anticipated: it collapsed the distance between concept and market validation.

Instead of describing HEARD through slide decks and static wireframes, Dumas could now put a working application in the hands of beta testers, hospital administrators, and potential investors. Users could tap through the advocacy flow, experience the caregiver dashboard, and interact with the escalation pathways - all before the production app was fully built.

“Lovable allowed me to share a functional app,” Dumas says. “It’s a mock, of course, but it’s still functional. People can click around and see what it’s truly supposed to look and feel like. It brought my vision to life in a way that just talking about it never could.”

The prototype also transformed her conversations with development teams. Rather than describing features through lengthy specification documents, she could point engineers directly at a working interface and say: build this. Cybersecurity consultants could evaluate the architecture. HIPAA compliance experts could audit the data flows. The functional mock-up became the common language between Dumas’s clinical vision and her technical team’s execution.

More critically, beta testing through the prototype generated real user feedback. Patients confirmed the advocacy scripts were intuitive. Caregivers validated the family-tracking dashboard. Healthcare providers expressed interest in clinic integration. The market was speaking - and it was saying yes.

This approach - building a functional prototype to test market fit while the production app is under development - represents a growing pattern among health-tech founders navigating an industry where development timelines are unpredictable and the cost of waiting can be measured in lives. For Dumas, the prototype wasn’t a workaround. It was a strategic accelerator.

Dumas’s ambitions for HEARD extend well beyond the consumer app. The immediate B2C play - patients downloading HEARD to advocate for themselves during hospital stays - is only the entry point of a much larger platform strategy.

Phase two is a B2B hospital dashboard. In the current system, patient safety incidents are reported in one of two ways: a healthcare provider files an internal report, or a patient fills out a survey three days after discharge. Both mechanisms are retrospective. Both are slow. And both miss the critical window where intervention could prevent harm.

HEARD’s hospital dashboard would surface patient-reported issues in real time, giving administrators a live view of what’s happening on the floor. A medication error flagged at 2 a.m. could be investigated at 2:05 a.m. - not three days later, not after a lawsuit is filed. Misdiagnoses, near-misses, and escalating situations would appear on the dashboard as they’re happening, enabling hospitals to do what they’ve never been able to do: respond in the moment.

Beyond hospitals, Dumas envisions HEARD becoming an employer benefit - a standard offering alongside dental and vision coverage. She’s also targeting Medicaid and Medicare reimbursement for patient advocacy visits, which would make the platform accessible to the populations most vulnerable to diagnostic errors.

“Healthcare is facing billions in lawsuits, readmissions, and CMS penalties because of quality and safety gaps,” Dumas says. “HEARD closes that gap. It’s not just a patient tool - it’s a lifeline for the entire system.”

Dumas was recently selected as one of ten founders for a Google-connected investor showcase in New York, where venture firms including Legion VC, Bison Ventures, and Millennia Ventures will be writing checks. She’s also pursuing her doctorate in nursing informatics - the same discipline that gave her the technical foundation to conceptualize HEARD in the first place.

Her third development team is now building where the second left off. The $50,000 grant is being deployed with the urgency of someone who knows exactly what’s at stake. And the Lovable prototype continues to onboard beta testers through myheardapp.com, generating the user data and feedback loop that will shape the production launch.

When asked where she sees HEARD in ten years, Dumas laughs - “Do we talk about ten years in tech?” - before turning serious. The vision is total integration: patients, caregivers, providers, hospitals, insurers, and regulators connected through a single platform where advocacy isn’t an afterthought but the operating system of care.

It’s an audacious bet. But Dumas has already survived the kind of odds that would break most founders. Nine losses before a miracle. Two brain surgeries. A misdiagnosis that should have killed her. A development team that vanished. Then another.

She’s still here. And now, so is HEARD.

To become a beta tester for HEARD, visit myheardapp.com. Follow Rachell Dumas on LinkedIn and social media at @rdumasrn.

Sequoia Blodgett is the CEO of Lexore Spark and founder of Simply Built. She writes about the founders building at the intersection of human stories and emerging technology.