Design your next trial on the evidence of the trials you have already run.
It reads your own trial history and recommends who to enrol, what to measure, when and how many, each with the reason, the source and a match score behind it. Your experts accept, change or reject every choice.
See it on a study in your pipelineSee how it works
4
specialist agents reading your trial history in parallel
100
of recommendations carry a rationale, a source and a match score
4
outputs from one approved design: brief, protocol draft, synopsis and study definition
Study design is where time, cost and risk get locked in
It is also the stage still done most by hand, usually without the relevant precedent at hand.
Decisions are slow
Settling population, endpoints, schedule and statistics means weeks of digging through prior documents before the design holds still.
Precedent is trapped
What prevents a bad criterion or an unworkable visit schedule sits in old protocols, and with people who may have moved on.
Choices are hard to defend
Without a traceable link to a source study, teams re-argue decisions their own data already settled.
The intelligence is in the design decisions. The protocol is the output.
This is not a document assembly tool. It reasons about why a design choice is right, and your team stays the author of everything that follows.
Every recommendation shows its working
Four layers of the design, each with a reason, a source and a match score. Your experts accept, change or reject.
Phase II · Respiratory · Moderate asthma
Primary objective: lung function at week 12
Eligibility
Adults 18–75, FEV1 40–85% of predicted
Historical data
Used in 4 of 5 comparable prior studies
RSP-203, protocol §5.1
Match0.88
Endpoints
Change in trough FEV1 at week 12
Guideline
Matches asthma guidance and your last two Phase II trials
RSP-207, protocol §3
Match0.92
Assessment schedule
Spirometry at screening, baseline, weeks 2, 4, 8 and 12
Historical data
A week-6 visit added burden without changing the analysis
RSP-203, schedule of assessments
Match0.79
Statistics
186 patients, 90% power for a 100 mL difference
Statistics
Effect size and variability from your pooled prior trials
RSP-203 and RSP-207 study reports
Match0.74
Approved decisions become
Design brief
Protocol draft
Synopsis
Structured study definition (CDISC)
Built to hold up in a review
Not just in a demo.
Your data only
Grounded in your own historical protocols. Nothing is shared between clients and no design leaves your environment.
Your environment
On-premise, private cloud or public cloud. Excelra hosts no infrastructure, so there is no cross-client data.
Expert sign-off, always
An expert reviews every layer. It supports your writers and statisticians and replaces neither.
A review step
Accept, modify, reject or regenerate. The recommendation is a starting point, never a decision.
An expert reviews every layer. It supports your writers and statisticians and replaces neither.
Where it fits best: therapeutic areas where you have run trials before. A first-in-area programme with no prior trials is a weak fit today, because every recommendation depends on precedent.
See it on a study type in your pipeline
Tell us the therapeutic area and where your design process stalls today. We will walk you through the recommendations, the rationale and the review step on a comparable example.
"*" indicates required fields
Built in the Excelra AI Lab
Domain scientists and engineers turning a client conversation into a working prototype, on one governed architecture.
40
AI solutions across the R&D value chain
8
live in client production today
2–4
to a working prototype
