Designing a Clinical Trial in One Day with Poorya Amini of Risklick

What we talked about.

In this episode of the Sunny Ray Show, host Sunny Ray talks with Poorya Amini of Risklick about what it would take to design a clinical trial in one day. Amini, a veterinarian turned pharmacologist, explains how his time at a clinical research organization exposed the mess of protocol writing, where dozens of people edit a 500 to 1,000 page Word document with tangled versions and track changes. That frustration led to Protocol AI, a data-driven, digital native platform launched in 2024 with strategic investment. He discusses why strategic money helped test the product in real pharma environments, how a partnership with Citeline connects clinical trial data to the right step of the workflow, and why small and medium companies can move faster than large sponsors. Amini describes a database built over almost three years that maps trials end to end, and explains why time to market matters most to buyers. He is candid that human decision making and regulatory documentation, not technology, remain the main barriers to a one day design.

A veterinarian turned pharmacologist explains why clinical trial protocols take so long to write and how Protocol AI aims to cut it to a day.

The questions, and the answers.

What pulled you from animal medicine into the machinery of clinical trials?

After my PhD in pharmacology, I worked at a clinical research organization, where I was responsible for designing and developing clinical trial protocols for clients. There I faced the challenges and saw the inefficiency firsthand. That pushed me toward the idea of transforming the way protocols are developed and building a solution for it.

Can you give me the specific scene that made you call protocol development inefficient?

Imagine 25 people with different expertise working on one document of 500 to 1,000 pages, mostly in Microsoft Word, with different versions and track changes. Everything in it is connected, so a change on page one must be reflected on page 1,000. You easily get lost, and a deletion in one version can go unnoticed.

What did taking strategic money teach you?

It was a good decision, because business knowledge matters and you must test your solution in real practice. A solution can look amazing in an isolated world, but once you implement it in a compliant, regulated pharma environment, you may find it does not work. A strategic investor helped make sure our idea works in practice.

Where are you against the goal of designing a study in one day?

Honestly, it is still the human factor. Technology can provide information in minutes, analyze it, suggest options and challenge your ideas. But a person must make the decision, and in companies that process is slow because approvals are needed from many people. The current processes work against one day design more than technology or data do.

How did the partnership with Citeline start?

We felt a synergy immediately. Citeline holds global data on clinical trials, sites and investigators. We have the workflow and process for how to use that data at the right point to make things efficient. Together we serve it as one single platform that all users can access and follow.

What does a medical writer do differently with Protocol AI?

Instead of hunting for information here and there, everything is already prepared. The writer presses a button to generate the document. Their role shifts from drafting text to reviewing, checking that the AI output is correct, directing the AI to fix it or correcting it manually, and validating the system's output.

Which number do buyers care about, half the document time or fewer amendments?

For sponsors, time to market matters most. From compound to market can take about 10 years, and selling starts only at launch. With Protocol AI we try to reduce time not just in document development but also in implementation and operation, cutting overall time to market.

What breaks first once protocols can be designed in a day?

The process has to align with the technology, and regulatory requirements slow things down. Quality checks, for example, can run with AI in minutes, but documenting them in a way regulators accept can take days. That documentation needs to change too, so studies can be designed and implemented very fast.