AI and eClinical systems consulting
AI for teams where mistakes are expensive.
BC Consulting trains clinical teams to build with AI, develops the applications they cannot buy off a shelf, and consults on the IRT, RTSM, and clinical supply systems those teams run on. We work where a wrong answer has real consequences: the clinical trial ecosystem, for sponsors and the service providers that support them. Everything we teach, we run our own business on first.
- AI training sessions and bootcamps delivered since 2025
- 11
- clinical supply professionals trained at one top-5 pharma
- 75
- AI-enabled control tower, replacing dozens of spreadsheets and SharePoint files at a clinical supply company
- 1
Who we work with
BC Consulting works in one field today, clinical research, and speaks the language of the people who run trials: their systems, their examples, and their buyers.
Sponsors and the service providers that support them
Clinical Research
We offer two services to sponsors and service providers: AI training and development for clinical teams, and eClinical systems consulting for IRT, RTSM, and clinical supply chain systems, backed by twelve years on the vendor side of those systems.
See the clinical research services →What we do
Three AI services and one systems consulting service. The clinical research page shows how they fit together.
AI Training
Programs that end with working tools, not a certificate of attendance. Every participant builds something they can use the next morning, in their own account, against their own work.
How AI Training works →AI Application Development
Agents, document pipelines, and decision tools built around your systems, your data boundaries, and your rules. You own what we build, and your team learns how it works while we build it.
How AI Application Development works →AI Implementation
The path from a few enthusiasts to an organization with a position: tool selection, governance, rollout, and the measurement that lets you report honest numbers to a board.
How AI Implementation works →eClinical Systems Consulting
IRT and RTSM vendor selection, implementation oversight, UAT, and optimization, plus the clinical supply chain systems around them, advised from the sponsor side by someone who spent twelve years on the vendor side.
How eClinical Systems Consulting works →Why a former teacher runs this firm
Bryan Clayton, M.S., taught high school music for nine years before twelve years in clinical trial software, most of it on IRT and RTSM systems where a bad configuration shows up as a dosing error. He has coached baseball at North Penn High School since 2016. That combination is the whole method: the patience to teach a room of skeptics, and the scar tissue to know which shortcuts get a team in trouble.
This site, our learning platform, and the agents behind our own operations are built and maintained with the same workflows we bring to clients. When we say a team can build its own tools in a day, it is because we have watched it happen.
Start today with a self-paced course
The same curriculum we teach live, on the BC Consulting Learning Platform. Enroll online and start the same day; the first course is free.
Intro to Prompt Frameworks
A practical, hands-on intro to structured prompting for clinical trial work. Learn the C.R.E.A.T.E. and I.D.E.A. frameworks (plus five more), practice on every exercise with a live AI coach built right into the page, and see how agentic AI applies to clinical operations. No prior AI experience needed, and free to take.
Start free →$129Claude for Clinical Trials: Foundations
Start using Claude confidently in your clinical trial work: safely, effectively, and without the hype. Learn how large language models actually behave, master the C.R.E.A.T.E. prompting framework, build verification habits that hold up in a regulated industry, and leave with a working Claude setup plus a starter prompt library built around real clinical operations tasks. No prior AI experience required.
View course →$179Claude for Clinical Trials: Building Your AI Workflows
Move from one-off prompting to repeatable AI systems. Build Claude Projects for your studies, write context files that teach Claude your therapeutic area, author reusable Skills that run your recurring work on demand, and add the verification layer that makes AI output trustworthy in a GxP world. Every module ends with a template, skill file, or checklist you keep. For clinical trial professionals already comfortable chatting with AI.
View course →$249Claude for Clinical Trials: Agents, Automation & Governance
Delegate whole workflows to Claude, with the guardrails a regulated industry demands. Learn the delegation ladder from chat to scheduled autonomous tasks, connect Claude to your files, inbox, and browser, design recurring briefs that run without you, and classify every AI workflow with a risk-based governance framework aligned to where FDA and EMA are heading. The capstone: a complete agent design for your own role, stress-tested before you build it.
View course →Latest writing
What we learn building AI tools, training clinical teams, and working on the systems trials run on, written for the people who do the job.
Clinical Research
All Clinical Research posts →
Headless IRT: Where Does Authority Live?
In a headless IRT, authorization, trial rules, blinding, and the audit record sit behind the API rather than inside a screen. A website, an integration, and an AI agent then reach the same engine, each held to what its own credential permits.
4 min read

Where the Safety Limit Lives When an AI Agent Runs Lab Hardware
Anthropic's Model Hardware Standard puts an AI agent's safety limits in the device driver, the layer closest to the equipment. Layered software design, GxP, and CSV have kept limits there for years, and that experience tells you exactly what to check before an agent reaches your instruments.
15 min read

What It Takes to Get Claude to Write in Your Voice
The fingerprints that make AI writing feel like AI, and a practical method for turning your own emails into checkable style rules that Claude applies everywhere you write.
5 min read