Software for evidence-intensive decisions

Turn messy evidence
into decisions
people can defend.

Orelian builds decision-support software for teams working with complex, unstructured data. Our flagship product, PFD Toolkit, turns thousands of coroners' reports into transparent, research-ready evidence.

Who we help

For organisations making consequential decisions across civil society and public life.

Government departments, public bodies, funders, research organisations and civil society institutions make consequential decisions using evidence spread across administrative systems, accounts, registers, grants, contracts and complex reports. Orelian connects and maintains that evidence so it can be trusted, interrogated and put to work.

Working withGovernment and public bodiesFunders and foundationsResearchersInfrastructure bodiesCharities

01 What do we know?

Evidence foundations

Bringing fragmented data and documents together into reliable, well-documented evidence infrastructure that can be stewarded or transferred.

02 What does it mean?

Analysis for important decisions

Answering focused questions about need, reach, outcomes, resources and risk, with assumptions and uncertainty made clear.

03 What should we do?

Decision tools

Turning evidence and analysis into useful workflows, monitoring systems and responsible AI tools that work in practice.

  1. Structuring messy evidence at scale

    Turning inconsistent documents, scans, and free text into reliable data that people can search and analyse.

  2. Automating evidence workflows responsibly

    Replacing repetitive review without losing the source links, audit trail, or human oversight that serious research demands.

  3. Making AI outputs inspectable

    Showing the evidence behind an extraction or classification so users can verify results instead of trusting a black box.

  4. Finding patterns across thousands of records

    Surfacing recurring concerns, organisations, locations, and themes that are difficult to see through manual review alone.

  5. Moving from prototype to dependable product

    Packaging models as maintained web applications and reusable Python tools that work with real users and changing data.

  6. Supporting reproducible research

    Giving researchers consistent workflows, structured exports, and transparent methods they can reuse and scrutinise.

03 Product development

Four steps. No black boxes.

Discover

Map the evidence, user workflow, and constraints before choosing models or writing product code.

Build

Develop the smallest useful product slice, joining data pipelines, models, interfaces, and auditability.

Validate

Test against expert review and real operating conditions, then measure both model quality and user value.

Ship

Release through accessible web interfaces and reusable packages, with documentation and monitoring built in.

  1. Start with the decision

    We design around real users, workflows, and incentives. Products are shaped to work in practice from day one.

  2. Keep evidence traceable

    The best technology is the technology people can use. We do not pick the most advanced option for its own sake.

  3. Be honest about the limits

    We make uncertainty, missing information and competing explanations visible. Sometimes the most useful result is knowing what the available evidence cannot support.

  4. Test before scaling

    We evaluate methods against real examples, measure errors and include proportionate human review. Promising ideas earn further investment through evidence, not enthusiasm.

  5. Build for continuity

    Everything we ship is designed for the people who will use it on Monday morning, not just the people who see the demo on Friday.

  6. Build for independence

    Open interfaces, clear documentation, and reproducible methods let users take their work further without depending on a services team.

Sam Osian, Managing Director of Orelian
Sam Osian Managing Director

05 Founder-led by design

“The best analytical work does more than produce an answer. It helps people see a decision differently.”

Sam founded Orelian to bring together two ways of seeing. His background in social science and doctoral research in data science mean he can go deep into technically demanding work without losing sight of the systems, people and consequences behind it.

Orelian grew from a belief that consequential decisions deserve evidence equal to their complexity. Sam helps organisations turn ambitious questions into practical, trustworthy evidence systems without losing sight of the people and public purpose behind the data.

Orelian is founder-led. Sam remains involved from framing the question through to delivery, keeping responsibility for the thinking and the outcome in the same hands.

LiverpoolWorking across the UKRemote and on site

06 Start with the decision

Tell us what you need to understand, predict, prioritise — or defend.

[email protected]