Step 01
Derive
Input your constraints. We bring in the regulatory, trial, financial and medical data for your indication and trial.
Design the trial your runway can fund.
Early-stage teams need to navigate hard constraints, including budget, burn and patient cap. We use these constraints to find the right design for you.
The platform for deliberate trial design.
The Trial Design Copilot makes every trade-off explicit. See how each constraint changes the design. Reach design lock faster, with fewer amendments downstream.
Step 02
Decide
Every perspective of the trial comes together in one place. Every decision is tracked, so the rationale behind the final design is clear.
Step 03
Deliver
CTD-style submissions and geographic expansion, plus protocol drafting, consent, site materials and notifications, with human review and sign-off throughout.
Six teams. One design. Every conflict named.
A trial design is never decided by a single team. Our platform surfaces the trade-offs across teams.
01_ statistical
What does it actually detect?
Operating characteristics, power, type I error — and the assumptions underneath them.
-
Pulls against
- ! vs financialPower 0.80 needs N=96. Funding covers 60.
- ! vs operationalEnrolment rate caps the N you can reach inside the readout window.
- ! vs regulatoryAn adaptive design needs agreement on type I error control.
02_ regulatory
Will an agency accept it?
Precedent designs, guidance, and the questions a reviewer asks first.
-
Pulls against
- ! vs statisticalAn adaptive design needs agreement on type I error control.
- ! vs commercialThe label claim needs evidence beyond the registrational endpoint.
03_ commercial
Does the readout support your claim?
Differentiation, and the evidence a partner or payer will ask for.
-
Pulls against
- ! vs regulatoryThe label claim needs evidence beyond the registrational endpoint.
- ! vs medicalThe claim needs a comparator this population cannot support.
04_ medical
Is this the right question for this population?
Endpoints, eligibility, comparator, standard of care.
-
Pulls against
- ! vs operationalThe assessment schedule raises site workload and patient burden.
- ! vs commercialThe claim needs a comparator this population cannot support.
05_ operational
Can it actually be run?
Sites, enrolment rate, schedule of assessments, and where the timeline slips first.
-
Pulls against
- ! vs statisticalEnrolment rate caps the N you can reach inside the readout window.
- ! vs medicalThe assessment schedule raises site workload and patient burden.
- ! vs financialMore sites accelerate enrolment and raise cost.
06_ financial
Does it fit the runway?
Cost per arm, per patient, per month — against burn and the next raise.
-
Pulls against
- ! vs statisticalPower 0.80 needs N=96. Funding covers 60.
- ! vs operationalMore sites accelerate enrolment and raise cost.
Surfacing disagreement early in the process.
Statistics wants 96 patients. Funding covers 60. That trade-off needs a decision. We don’t make it for you. We lay out the options, show how much each one costs, and give your CMO and statistician a clear basis for sign-off.
Built at Cambridge, for clinical-grade work.
Published trial-design methods and machine-learning research, applied to the decisions that have the greatest influence on clinical programme cost.
Start designing with Beaconsfield Labs.
We're working with a small group of biotech teams ahead of their next design lock. Leave your email — we'll keep you posted and reach out if there's a fit.
or email us directly — lukas@bf-labs.ai