Memora Health
Winter 2018AcquiredDigitize and Automate Complex Care Workflows
Pressure-test this opportunity
Explore the risks and possibilities with a prompt for ChatGPT, Claude, or your agent.
Digitize and Automate Complex Care Workflows
Explore the risks and possibilities with a prompt for ChatGPT, Claude, or your agent.
Memora Health built a text-message layer between patients and care teams. Founded as a Harvard student project and launched as a company in 2017, it turned clinical protocols into automated journeys for medication adherence, symptom checks, education, appointments, and escalation. It raised almost $80 million and signed prominent health systems.
The product addressed a real problem, but the business appears to have confused contracted possibility with realized revenue. Axios later reported that leadership discussed roughly $20 million of annual recurring revenue in 2022 while a merger document showed only $1.2 million of revenue in 2023. Long pilots, custom integrations, and uneven product behavior made the gap hard to close. Commure acquired Memora in December 2024; Axios reported an estimated $30 million stock value, below total funding.[1]
Memora existed before its official founding date. In March 2016, Harvard's dental school described a project by Manav Sevak and Nisarg Patel that had reached the President's Challenge finals. The early product let clinicians manage medication lists and send automated adherence texts.[2] Kunaal Naik joined the founding team as the technical leader; the company later dated its launch to 2017 and entered YC's Winter 2018 batch.
The founders combined medicine, public health, and computing. Patel studied molecular bioscience, biotechnology, political science, dentistry, medicine, and biomedical informatics before surgical residency. Sevak had worked in public-health and computational-biology research. Naik's computing background supplied the technical complement.
Sevak tied the idea to a friend, John, who had colitis. John tracked medications, diet restrictions, and symptoms with paper, pill labels, searches, and phone calls. After observing rounds, Sevak concluded that hospital care was coordinated inside the facility but broke apart once patients went home. He wrote that the founders then spent months “doing cold outreach or walk-ins at every doctor's office” they encountered while driving around the country.[3]
The first narrow intervention sent medication reminders and tracked adherence. Early work with Brigham and Women's Hospital, Grady Memorial Hospital, and Emory Healthcare pushed the team toward a broader platform. Each health system used different pathways and records, so Memora made its workflow engine configurable. Patients would receive ordinary text messages; the system collected answers, sent approved education, and escalated concerns to clinicians.
The ambition grew from a reminder tool into an operating layer for outpatient care. In 2021, Sevak wrote: “Memora is rewiring healthcare systems with modern technology.”[3] That ambition won enterprise logos and capital. It also created the central execution risk: every new specialty, hospital, EHR configuration, and clinical pathway widened the distance between a demo and a dependable deployment.
Memora turned a health system's clinical workflow into a two-way SMS journey. A hospital supplied its preferred education, follow-up schedule, questionnaires, and escalation rules. Memora configured those steps, enrolled patients, and communicated without requiring an app or login. Patients could report symptoms, complete surveys, receive medication reminders, manage appointments, and ask common questions.
On the clinician side, Memora summarized responses, routed urgent concerns, and wrote routine follow-up work back into an EHR or its own dashboard. It supported care-team assignments and in-thread collaboration. The system was retrieval-based: approved clinical content and pathways constrained answers, reducing the risk that a generative model would invent medical guidance.
The product evolved through three scopes. The 2016 prototype centered on medication adherence. By 2020, Memora combined messaging with remote-monitoring devices and marketed a system for blood pressure, glucose, oxygen, weight, and activity. A California Health Care Foundation showcase said the company had processed 93 million patient interactions and claimed a 16% reduction in readmissions, though those were company-provided figures.[7]
By 2023, the company called the product “intelligent care enablement.” In an Axios interview, Sevak explained the configuration process: “What would your gold standard workflow look like?” He said Memora digitized about 60% of that workflow.[8] The stated customer set included Penn Medicine, Moffitt Cancer Center, Northwell Health, and Boston Medical Center.
The same breadth created delivery problems. Axios later reported allegations that EHR integration and language support sometimes fell short of what hospitals expected. Anonymous sources described a Stanford oncology pilot that did not expand, incomplete chatbot responses, and at least one cross-hospital message-routing error.[1] Memora and its investors did not respond to Axios's detailed reporting beyond Sevak acknowledging one ARR statement, so those claims remain attributed allegations rather than adjudicated facts.
Memora sold to health systems, health plans, life-sciences companies, and digital-health providers. Within a hospital, clinical operations, digital-health, and service-line leaders could sponsor deployments. Patients used the product, but institutions bought it through monthly subscriptions and implementation work.
The sales story paired staff relief with patient experience. Memora claimed it could automate routine questions, reduce inbox traffic, and let nurses focus on exceptions. General Catalyst wrote in 2023 that Memora had more than 70 healthcare partners and reported inbox reductions as high as 40%, equal to as much as 2.5 hours saved per employee per day.[9] These metrics came through company and investor material, not an audited cross-customer study.
The addressable need was large: every health system manages discharges, chronic disease, procedures, medication adherence, and patient questions. The commercial market was narrower than the clinical need. Each enterprise buyer required security review, EHR work, clinical governance, procurement, and proof that automation reduced labor or improved reimbursable outcomes.
Memora's funding history shows investors priced the category for broad adoption. The company announced $10.5 million in 2021, $40 million in 2022, and $30 million in 2023.[3][4][5] Fierce Healthcare later cited $79.8 million raised.[10]
Memora competed with patient-engagement vendors such as Artera and Luma Health, care-management systems, point products for remote monitoring, and the communication features inside Epic and other EHRs. Its claim was broader than a reminder tool: reusable clinical pathways, conversational patient support, triage, and EHR actions across many specialties.
That breadth weakened differentiation when deployments remained custom. EHR vendors owned the clinical record and could add messaging. Point products could go deeper in oncology, maternity, or chronic care and tie value to a defined outcome. A horizontal startup had to integrate everywhere while proving expertise in each care pathway. Commure offered another answer: assemble patient navigation, ambient documentation, revenue cycle, and operational data under one vendor. It acquired Memora to add care navigation to that suite.[6]
Healthcare organizations paid monthly subscriptions, often beginning with a pilot or service-line deployment. Forbes reported more than 50 paying organizations in 2021.[11] The company could expand an account by adding specialties, patient populations, or workflows.
The missing distinction was between signed potential and recognized revenue. Axios reported that former employees saw projected multi-year deal value entered into annual contract fields. The same report said Sevak acknowledged telling investors Memora had about $20 million in ARR in 2022, while a transaction document listed $1.2 million of 2023 revenue and about $1.5 million through October 2024.[1]
If those figures are comparable, the gap is the business model's central failure. A five-year expansion possibility is not recurring revenue until the customer commits, deploys, and pays. The public record does not provide enough contract detail to reconcile bookings, contracted ARR, recognized revenue, pilots, or implementation fees. That ambiguity should have increased diligence, not supported a $430 million valuation discussion.
Memora had genuine use. Its customer materials named Mayo Clinic, Penn Medicine, Moffitt Cancer Center, Intermountain Health, Aetna CVS, and others. In one 2023 postpartum example, the company said more than 32,000 messages produced fewer than 150 manual clinical responses.[5]
Yet customer-logo counts blurred pilots, historical customers, and live revenue. Fierce reported more than 50 organizations but noted after the acquisition that Edward Elmhurst Health was no longer a customer.[10] Axios cited former employees who estimated only six to 12 hospital clients at a time and 15 to 25 total by the end of 2024.[1] These definitions can coexist, but only one measures an active commercial base.
Memora's primary failure mechanism was measurement, not lack of need. Enterprise-health contracts expand slowly: a pilot may cover one pathway, then require clinical proof, integration work, security approval, and new budget before it grows. Treating the possible expansion as current ARR removes the feedback that tells a company whether deployments repeat.
Axios reported that employees sometimes entered projected multi-year value into annual contract-value fields and were encouraged to track potential revenue as actual revenue. Sevak acknowledged telling investors the company had roughly $20 million of ARR in late 2022. The merger document's later revenue figures were far smaller.[1]
The company attempted to grow into the projection. It raised $30 million in 2023, added health systems as investors, and planned to sell deeper into existing hospitals, payers, and digital-health companies.[8] Axios reported at least four layoff rounds in 2023 and 2024. Cutting costs could extend runway, but it could not turn pilots into renewals.
Memora sold one configurable platform across oncology, maternity, surgery, chronic care, and other pathways. The pitch sounded efficient: build the engine once, then encode each institution's gold-standard workflow. In practice, every specialty carried new content, clinical thresholds, languages, EHR data, escalation rules, and safety risks.
The company tried to control risk through retrieval-based responses and clinician-authored programs. That was a sound choice. Axios nevertheless reported that some promised bidirectional EHR and language capabilities were limited, and that incomplete knowledge produced bad or repetitive answers in some deployments.[1] The remedy was more configuration and implementation work, which made the horizontal platform less repeatable.
Commure officially disclosed no price. It said Memora would strengthen patient engagement and care navigation inside CommureOS, and Kunaal Naik said the combination could take the product to millions of patients.[6] Memora's site now identifies the product as part of Commure.[12]
Axios reported an estimated $30 million in Commure shares before adjustments and said liquidation preferences of up to $87 million left common shares with no value.[1] That report relied on a shareholder document and anonymous sources; no public filing confirms the exact consideration. Even with that caveat, the reported value against nearly $80 million raised explains why the transaction was described as a fire sale.
The strongest counterargument is that Memora built clinically useful software and sold into famously slow institutions. Its patient engagement and inbox-reduction examples show real value, and Commure kept the product. That does not answer the financial gap. A useful product can still be a poor venture investment when capital, valuation, and revenue recognition get far ahead of repeatable deployment.
Define revenue before celebrating it. Memora's pilots, contracted possibilities, and live revenue were not interchangeable. A board should reconcile bookings, contracted ARR, deployed ARR, recognized revenue, and cash collection every quarter.
Customer logos need status labels. “Worked with” can include a pilot that ended, a strategic investor, or a live enterprise account. Memora's public roster sounded large while Axios's sources described a much smaller concurrent base.
Horizontal clinical software accumulates exceptions. Each new specialty and hospital added workflow, language, EHR, and safety variation. The common engine only saved work when the remaining configuration stayed small and measurable.
Safe model design does not guarantee safe operations. Memora constrained responses with approved content, yet Axios still reported routing and knowledge failures. Testing must cover tenant isolation, institution-specific content, escalation, and production monitoring.
Strategic capital cannot substitute for renewal. Health-system investors gave Memora access and credibility. The venture outcome still depended on deployments expanding into durable recognized revenue.