How AI agents help restore sight for patients with inherited retinal diseases.
Maybe We Don't Need a CRO After All
There is a lot of marketing B/S around AI these days, but there are some success stories, and not just Harvey for law firms.
This week I decided to write a “once-upon-a-time” story to illustrate how agentic AI is changing clinical trials.
I’ll show you AI company and drug company perspectives.
One races to win in a white-hot market. One races to give sight back to the blind.
Let’s get it on.
Intro
This week, Ram Yalamanchili, founder and CEO of Tilda Research was my guest on Life Sciences today.
Several weeks ago, I hosted George Magrath, who leads Opus Genetics. Opus Genetics develops gene therapy for ultra-rare inherited retinal diseases and restores sight to people who went blind at a very early age. Opus has taken a platform approach, developing 7 assets to maintain continuous revenue flow for groups of very small patient populations (hundreds in the US).
When George Magrath's team at Opus Genetics connected with Ram, they discovered something powerful: AI agents could handle the operational grunt work that typically requires armies of CRO staff.
The results for Opus are impressive:
Protocols adapted automatically across their 7-asset pipeline
Informed consent forms generated instantly
Real-time edit checks eliminating data queries
Administrative overhead cut by 70%
It was a glimpse of what happens when clinical operations shift from armies of people to a coordinated team of AI agents.
AI-powered autonomous agents that perform clinical trial tasks
100% human supervised with 90-92% acceptance rate
Can scale instantly from 1 to 100+ sites without additional resources
Utility-based pricing model: customers only pay for work completed
No-fixed costs for customers
Enable 100% resource utilization
See this episode of Life Sciences Today here - AI Teammates for clinical trials
Once upon a time, there was a princess.
The princess went to Stanford and did a PhD in theoretical computer science and a post at Carnegie Mellon in AI.
When she was in 5th grade, the princess announced to her parents.
“I will change the world for drug development. I will build AI agents as modular systems driven and enabled by LLMs and LIMs for task-specific automation. I will solve problems in each clinical paradigm including hallucination, brittleness, emergent behavior, and coordination failure, and implement ReAct loops, and causal modeling. I will develop robust, scalable, and explainable AI-driven systems for drug development”.
The queen sighed and said, “I don’t understand a word of what you said dear, but if it makes you happy — go for it”.
And so it was.
Imagination Lab School - Palo Alto
By 5th grade she was already programming and building things.
She graduated high school and built them at Stanford.
She built them at Carnegie Mellon at her post-doc.
Her need for building things was like her need to run (the princess was a marathon runner).
After Carnegie she worked at a big virtualization company doing security research. After 8 years at the virtualization company, she left and started a mobile security company with a friend from school. 5 years and 2 startup acquisitions later, she became an entrepreneur in residence at a Sandhill Road VC.
She learned at the VC that there were 2 kinds of entrepreneurs:
Builders and Players
Always a builder
One evening, over dinner, her husband told the princess,
“You could take some time off.
We have the money.
You don’t have to work.
Your parents are healthy. They don’t need you at the Royal Family Office.
You’ve earned it.”
She shook her head, “My friend Tim is working on an AI super model to develop biomarkers for rare disease trials. He’s been pitching us for his Series B how they can make trial recruitment more efficient and speed up results by bringing the right patients into the right interventions.
I’m not done. I want to see how this world really works. From the inside.”
She bought a clinical research site in Milpitas.
Not because it made sense on paper — it didn’t — but because she wanted to see first hand what her friend Tim was talking about.
She came to work as a site coordinator.
And she saw it all.
Binders fat with tabs and paper.
Data queries drifting in weeks late.
Invoices aging in drawers.
People working hard, and always behind.
Endless email and track-changes threads in Word documents.
She remembered her 5th grade dream and her lecture to her folks about AI.
The 5th grade dream comes true
The princess raised seed money from the royal family office, hired some developer friends from her mobile security company and a year later launched her agentic AI system to run trial clinical operations. They started with regulatory and site startup loops and continued with agents that write protocols and monitoring plans.
The General Partner at the VC where she had been entrepreneur in residence started sending her leads - Bay area biotechs who were short on time and frustrated with inflated CRO budgets and timelines.
One day he called her up, “Princess - I have a lead for you. A physician from Triangle Park working on rare genetic defects that cause blindness at a young age. A buddy of mine told me about them. I think they’d be a great early adopter for you guys”.
University of South Carolina
Meanwhile, across the country, there was a boy from South Carolina who loved healing people.
He studied medicine at the University of South Carolina with a specialty in ophthalmology.
He studied economics at Johns Hopkins.
He started a full-service ophthalmic-focused biopharmaceutical solutions company.
As CEO, he led the company through substantial growth and an acquisition by a top tier private equity firm.
After the acquisition, he got interested in ultra-rare inherited retinal diseases and founded a startup to develop gene therapy to restore sight.
Rare disease therapy - a blessing and a challenge
Gene therapy for rare genetic eye diseases is a human blessing and an economic challenge.
A blessing because if you get it right, you restore sight for life.
A challenge to sustain revenue and ROI from very small patient populations.
He realized that the key was a platform approach and super-efficient clinical development that would use adaptive trials to move rapidly from Phase 1 to commercialization.
The startup was acquired and they became a clinical-stage drug development company with a pipeline of 7 indications.
They used ML to predict treatment response and he published a paper “Use of a Convolutional Neural Network to Predict the Response of Diabetic Macular Edema to Intravitreal Anti-VEGF Treatment”.
But he was missing a critical piece for clinical trial operations. He called a friend at a West Coast VC and asked him if he knew companies replacing CROs with AI agents.
The friend put him in touch with the princess and they started a partnership to use their AI-powered clinical trial platform and enable efficient scaling across their multiple rare disease programs.
They developed AI agents for the gene therapy pipeline that:
Adapt protocols from existing products in his pipeline
Automatically generate informed consent forms
Automatically create eCRF forms
Generate monitoring plans
That do real-time edit checks
That automate administrative functions like user and site management
And they both had an AHAH moment:
We don't need a CRO to run a trial.
We can move fast and build things ourselves with a team of AI agents.
Afterword
Tech companies in emerging AI markets and biotech companies targeting rare diseases share the same paradox: they must move fast under severe resource constraints.
In agentic AI, speed is a race against competitors—markets shift weekly, and early movers capture outsized value. Success demands lean, agile teams that iterate rapidly while balancing engineering investment against customer acquisition and burn.
In rare-disease biotech, speed is a race against time for patients—every delay prolongs suffering for those with few options. Success demands focused research teams that advance through milestones efficiently while managing the enormous capital requirements of drug development.
The stakes differ—market share versus human lives—but the challenge is identical: move fast enough to matter, and make every resource count.
About Me
I’m a former pharma-tech founder who bootstrapped to exit.
Now I run a private community with 900+ life science leaders helping them maximize their revenue with the right partners.
I hear these insights first-hand every week from the founders building the future of TechBio.
If you want to get them before your competitors do, join my private network for techbio entrepreneurs.
If you are a techbio leader contact me here to be a guest on the Life Sciences Today podcast.
About Tilda Research
Tilda Research AI Teammates for clinical trials already do a lot of the grunt work in clinical trial operations with 100% utilization and utility-based pricing.
🔗 Visit Tilda Research