This site is not optimized for Internet Explorer 9 and lower. Please choose another browser or upgrade your existing browser in order get the best experience of this website.

Case Study

Developing and Launching a Knowledge Graph-Backed Telemedicine API In Just Eight Weeks

Cover Developing and Launching a Knowledge Graph-Backed Telemedicine API In Just Eight Weeks

Download the PDF


Medical care is a deeply personal business that begins impersonally: log into your insurance provider’s database of in-network physicians and find the doctors closest to you who can serve your need.

That’s an unacceptable option for those living in the public eye. For them, seeing a new provider for anything – but especially for medical consultation or procedure – entails reputational risks the rest of us don’t face. Discretion is worth paying for. So, to fill the gap, a startup telemedicine provider built a digital referral system on GraphGrid’s Connected Data Platform (CDP), allowing doctors to identify networks of trusted partners for their most visible patients.

Knowing users of the system would demand both privacy and an exceptional standard of care, the telemedicine startup had to build a scalable, searchable, context-rich knowledge graph. Doctors had to be able to give their visible clients trusted options for care in those moments when a virtual visit from 10,000 miles away simply wouldn’t provide relief.

In this case study, we’ll review how the telemedicine provider went from concept to live in just eight weeks, saving time and money that could have gone into infrastructure that GraphGrid CDP provides out of the box.



Scale a global referral network for in-person treatment while preserving privacy.


Use knowledge graphs to capture and encode the elements of trust.


Move from finished concept to finished product in eight weeks.



Scale a global referral network for in-person treatment while preserving privacy.

Health care is an inherently complicated business. But it can be even more so when catering to high-profile people who, because of their public visibility, demand the utmost privacy.

Telemedicine can be an excellent option for this cohort: it’s personal, safe, and outside the view of prying paparazzi. Trouble is, not all medical networks are built equally. A celebrity booking through a common-use platform might run into an administrator willing to feed the gossip media a tip.

It happens more often than you might think. Remember: it was the gossip site TMZ that first broke the story of pop legend Michael Jackson’s heart attack and subsequent death in 2009. The new telemedicine platform would be built to help prevent just this sort of leak, and pent-up demand meant the founders would need to move quickly in developing the components.

But they’d also have to be careful about using third-party tools when building the system. Every outsourced component could introduce a security risk, potentially invalidating the startup’s core value proposition of discreetly providing an exceptional level of care.

Even worse, many of the startup’s core customers would travel frequently to distant locations. Finding trustworthy on-site care – especially with the level of discretion these high-end patients would expect – would force the startup to find ways to vet providers sight unseen. The platform itself would have to prove trustworthy.


Use knowledge graphs to capture and encode the elements of trust.

In building out its offering, the startup tackled three big challenges in order to establish trust:

  1. Capture clinical and preferential data. Most doctor’s offices focus on family history, known conditions, and other clinical data. The startup decided to also capture preferences. Why? A client choosing homeopathic care when near home and their primary care physician will likely need that same expertise when getting care on the job in a different country. Understanding preferences could also help to foster trust.
  2. Make it easy to search using a wide range of criteria. A skilled physician whose office has no clear private access point might be less desirable than a comparable physician operating from a more discreet location. The system would be built to match patients and doctors according to their needs not only for care but also for personal privacy and safety.
  3. Build a system for scoring trust. Most importantly, the startup needed a way to reassure clients that doctors they’d never seen could be trusted to care for them. How to do that? Leverage other doctors, giving each clinician the option to build a referral network right on the platform. Patients seeking care would see who their doctors trusted, and why.

Patients needed choices they could feel confident were relevant, trustworthy, and reassuring. These patients aren’t like the average American, who sees their doctor four times per year. Instead, the startup’s clients travel the globe frequently and may get sick or suffer unusual injuries on the job. It’s entirely possible they would see their primary care physician half as often – or less – than a third-party clinician in a different part of the world.

The system had to find these trusted providers quickly, and securely, and provide a private means to arrange care at any time. To fill that need, the startup built a series of knowledge graphs capturing rich, contextual information about both patients and their doctors. GraphGrid CDP made it easier to leverage the data and build business processes around it, including payments and documentation.


Move from finished concept to finished product in eight weeks.

Ultimately, the startup and its partners developed over 200 APIs using “Showmes,” a low code feature in GraphGrid CDP for creating dynamic interfaces that would help to define the essential features of the telemedicine platform.

Specifically, GraphGrid CDP provided the tooling to enable security, search, natural language processing, and other services for generating value from the knowledge graphs at the heart of the system. As a result, the startup team was able to get from concept to testing and deployment in just eight weeks.

Today, the collection of APIs developed with GraphGrid CDP has allowed the provider to iterate rapidly and scale quickly. Every new patient-doctor interaction – especially referral visits in new locales – helps bring new context into the knowledge graphs, improving trust scores. CDP makes that data instantly searchable and provides tools for creating new workflows.

For example, a regular patient might upload their itinerary for an upcoming job. CDP could scan the details and alert the patient’s primary care physician to add any relevant referral partners in the region. The system could also provide a tailored list of the highest-scoring caregivers in the region. In each case, the startup can use context to provide value.

Most importantly, the telemedicine startup didn’t need to spend years identifying and building the underlying technology platform. Instead of committing precious time and capital to operating and maintaining the structural bits and bytes of a global system, the startup team was able to focus on sharpening the knowledge graphs that represent the working model of the business and the data that defines it.

Patients win. Doctors win. And so do the founders and investors who’ve seen their idea come to life.

Get started now for free

On Sept. 15, 2021, GraphGrid unveiled the Connected Data Platform downloadable package in two editions: Enterprise and Ecommerce. Both editions include a full feature set. Enterprise edition includes Natural Language Processing (NLP) capabilities so you can turn text-based information into connected data. Ecommerce edition includes tools to create a smart shopping experience for customers and manage things like payment processing, invoicing, and order tracking.

The current 1.4 edition will be followed by a bi-annual release schedule. Current customers get 8 CPU cores, 32 GiB memory, and 1 GPU for all editions and their features and can be used through production for FREE.

Need more capacity or interested in having us run it in our cloud? Contact us now!