senior product designer
consumer growth ux lead
consumer growth ux lead
.svg){kind=logo}
josh cheriyan
BASED IN SEATTLE
7+ YEARS OF EXPERIENCE
.png)
I was a visual designer, newspaper editor, and photographer as I learned my way into UX. Being a "creative" is in my bones. Modern design approaches, high-quality mockups, and the marriage of form+function guides my every next frame.
Whether it's sitting with PMs, leadership, and stakeholders to roadmap at a high level or diving into the weeds of edge cases with a developer during handoff, I thrive on collaborating, making friends, and doing more than just be a "design resource".
I'm not my user even if I dogfood what I'm building. Customer interviews, usability testing, session replays, analytics dashboards, A/B tests, surveys, reviews, competitor analyses—all give me yet another vantage point to test against my assumptions and hypotheses.
Some screens, concepts of a case study or two, and a fun variety of direct-to-consumer and enterprise projects.
With every burgeoning telehealth company chomping at the bit to sell compounded medication, it's paramount that patients are more valuable than their wallets. PlushCare—a telehealth company owned by Accolade—wanted to jump into the market by extending services we already in place for urgent/primary care to those needing chronic care support, particularly in weight management. We opened the floodgates and then soon realized we were ingesting far too many patients that didn't qualify for the medications they were seeking.
Customer pain point
Leaving disgruntled within minutes of the doctor visit because they didn't realize they didn't hit the parameters to qualify for weight management prescriptions.
Business pain point
Doctor utilization is pertinent indicator of overall customer satisfaction and the mental health of our clinical staff. As weight management patients poured in, appointments were getting booked well in advance all to be wasted upon discovery of patient disqualifications. We needed to get utilization rates down and have more qualified bookings.
My contribution
I conducted several interviews with existing patients, brainstorms across teams, worked through concepts, usability testing, and arrived at an MVP that added friction to what was already a longer booking flow. The intent was that even if we lost a bit on conversation rate, we'd save the business on utilization rate and an increase on billable appointments. The MVP was a clinically-designed set of questions that would triage patients into a few flows based on their health history. We'd be able to recommend the right medication from the start, set better expectations for the patient, and let them choose to continue on to booking the appointment only after they'd been told whether they can be treated.
The result
Not only did we lower doctor utilization rates to a manageable level and increase billable averages in appointments, but we also managed to improve conversion rates by 10% in A/B tests. How did added friction to an onboarding flow produce healthier conversion? Our hypothesis is that our triage added a trust factor for patients who wanted to stay away pill factories and sketchy businesses.
Next steps
The images you see on this post are part of that next step. We have a cut-and-dry triage right now, but we started exploring what we can do to flesh out the MVP into something more robust. With that research, we found that we need to further educate patients on our process, encourage them with testimonials (in an organic way), and use product placement to speak to the real reason many of them had arrived at the top-of-funnel in the first place (most searches included "ozempic online", "weight loss prescription", "compounded GLPs", etc.). We're hoping that this next round of A/B testing yields an even healthier conversion and also drive down the amount of customer reviews that complain about surprises in our weight management treatment process.
.png)
This was initially a quick visiontype of a future in which some automation, conversational AI, and generative AI could be compiled to onboard patients into a telehealth app.
Instead of tedious forms, repetitive data entry, and winding paths to end up with a doctor's visit, this approach serves to converse with the patient, triage them into certain flows and follow-up questions that gives the doctors more context around the patient's health, use available technology to parse image uploads, and use all inputs (in this conversation and also based on previous patient experiences) to recommend doctors that are particularly able to help them.
I served as PM and product designer on a hackathon team that set a goal of reducing the 7+ minutes it takes to book an appointment, reducing the scrolling users are forced to deal with to find an appointment time that works for them, and mitigate form field exhaustion—particularly on registration and insurance pages.
Our team introduced a set of LLM agents that were engaged on a particular set of tasks and would interchangeably hand off context to other agents in the chain as the user-AI conversation progressed. The conversation used natural language processing, but also embedded microservice widgets into the chat itself that would render booking flow components like the appointment list UI. We then ported the artifacts of the conversation into the existing booking flow with the intent to pre-populate fields and interaction choices that had already been covered within the AI conversation.
Our team was one of four finalists, with us winning the Innovation and Creativity category.
Your social graph is an exquisite and yet haunting mirror of your tendencies and desires. Some companies harvest that data; others ingest. For the sake of convenience, folks have mostly resigned to what's out there, but then a hacker hits and suddenly the dark web is gorging on stuff you thought really was going to be securely held.
That's where Integris Software came in. As the sole product designer for what eventually was bought by industry leader OneTrust (and repackaged as their current DataDiscovery product), I had the opportunity to shape the future of protecting user data using machine learning (ML).
In this case study, I dive into the importance of the dashboards I built.
In the fall of 2020—before a single Rivian vehicle hit the streets—I was asked to come up with their post-purchase confirmation and insurance design.
I worked with a PM who spec'd out most of the flow in Lucidchart, a research team that conducted customer interviews and a creative director that kept my work aligned with other parts of the Rivian digital experience.
From there, I created the flows, wireframes, and near-final mocks for what dictated the desktop view for this feature.
You go to an urgent care, sign forms you don't read, see a doctor for 5 minutes, and then get charged $1000. Patient billing must be more transparent.
I was able to work with a great Amazon Web Services team to develop a hub of resources for the hundreds of cloud partners with now almost 1,000 videos.
I was tasked with presenting the information in a way that was
1. Featured in an all-up page
2. Searchable by series
3. Searchable by a deeper set of filters
I worked with other agency designers to come up with a brand theme that said it was AWS (dark blue, grid-heavy), but also something really fun and almost playful in what's normal a very sterile environment.
I heavily leaned on the restraints set up by the existing CMS especially since the search functionality had to work like it does on the overall platform. I then handed this off to a developer and worked alongside them to implement and test.
I was tasked with a take-home design test during an interview that I thoroughly enjoyed. Although exercises like this lack the essentials of thorough research, cross-functional collaboration, testing, and implementation, I still get to work some of my muscles to think my way through a creative but functional output.
One of the first projects in which I was purely a UX designer for was truly a monster to behold. Nearly 10,000 pages of an older website were going to be brought into a totally new system designed by our team. I worked on a few pieces here and there with a microsite being one of the more invested projects.
The overall company had several product microsites. These few pages needed to sell the product with a great story and give potential customers an easy way to gather the information they needed to contact sales for a purchase.
Our UI team worked on fonts, colors, and basic components found throughout the microsite. I chose photography throughout the overall site experience as well as this microsite, set up the general user experience here, and created custom components like the product comparison chart and mobile in-page navigation.