Your firm depends on data supplied by customers, partners, analysts, or specialist software. You need a practical way to catch problems before that data reaches the next person or system.
McQuillen Interactive can help turn your requirements and expert knowledge into clear, repeatable validation workflows. We help you find data problems earlier, highlight them clearly, and give your team more confidence in the data it accepts.
Better yet, we help you help your users (and their AI agents) validate data before you even see it... thereby saving you time.
Validation that fits the way your data is used
Good validation is more than a pass or fail. It captures what your team knows, applies those rules consistently, and gives people enough feedback and context to decide what happens next.
Define what “valid” means
We can help turn specifications, business rules, reference data, and expert review habits into checks that people can understand and maintain.
Make checks repeatable
A reusable workflow can apply the same rules each time, whether data arrives through a web interface, an API, a scheduled process, or an internal tool. This makes review more consistent and lets specialists spend their time on exceptions that genuinely require judgment.
Return useful results
Findings should help someone fix a problem. We focus on results that identify where an issue occurred, explain it in appropriate language, and preserve enough context for your team to understand what was checked. Results are structured so they're helpful to both humans and AI agents.
Validibot: open-source foundations for custom client workflows
We created Validibot as an open-source platform for turning data requirements and expert knowledge into clear, repeatable validation workflows. It gives us a practical foundation for testing ideas and building excellent custom workflows around a client's real data.
A workflow can combine several kinds of validator, run each check in a sensible order, and return one coherent set of findings. That lets simple structural problems be caught early while preserving space for the domain rules, simulations, and specialist judgment that make the validation genuinely useful.
Validibot gives us a powerful set of validation building blocks we can adapt, extend, or replace according to a client's formats, risks, existing systems, and long-term needs.
Examples from the Validibot blog
Validating inputs with EnergyPlus™
Parameterized IDF templates check inputs, run a controlled simulation, and compare selected outputs with engineering rules.
Provenance is not validation
A biodiversity CSV example combines tabular schemas, row-level CEL rules, and portable evidence of what passed.
XML Schematron support
A calibration-document workflow pairs XML structure checks with Schematron rules for cross-field and domain requirements.
From everyday files to specialist models
The right validation approach depends on the format, the risk, and what the data needs to prove. A single workflow may combine several kinds of checks while presenting one clear result to the person who supplied the data.
CSV and tabular data
CSV files and spreadsheets often look simple while carrying important assumptions. Validation can check expected columns, required values, data types, ranges, permitted values, uniqueness, cross-field relationships, and rules specific to your domain. Clear row- and field-level findings can make correction much less frustrating.
JSON and XML
Structured data can be checked against an agreed schema and then examined for business rules that a schema alone cannot express. This is useful for data exchanged between organizations, received through an API, or prepared for another system.
PDF documents
Where PDFs are part of a formal submission or reporting process, we can help design checks around document structure, required content, metadata, and consistency with accompanying data. Document validation can sit alongside other file checks rather than becoming a separate manual process.
EnergyPlus™ and IDF files
Building-energy workflows may need to do more than confirm that an IDF file can be read. Validation can examine important model inputs, run a controlled EnergyPlus simulation where appropriate, and assess selected outputs against project or engineering criteria. The exact checks are shaped with the specialists who understand what a credible model means in your context.
FMU/FMI models
Functional Mock-up Units can be exercised in a controlled simulation so that chosen inputs, outputs, and behaviours can be compared with expected limits or reference cases. This can help teams receive and review model-based deliverables more consistently without pretending that every question can be reduced to an automated rule.
Bespoke rules and existing tools
Some of the most valuable checks are unique to your organization. A workflow can incorporate calculations, cross-file relationships, trusted reference data, an internal API, or validation code your team already relies on. The aim is to connect useful pieces into a process that is easier to run and understand.
Validation that becomes part of the workflow
Validation can be made available through a straightforward web interface, integrated through an API, or placed behind an internal process your team already uses. Reusable workflows can bring together several validators, keep their rules visible, and return one coherent set of findings.
This can give data suppliers a chance to correct common issues before submission, while giving reviewers a more consistent starting point. It can also create a clearer shared understanding of the requirements as those requirements evolve.
How we can help
We can start with one important data flow and build outward from evidence. A practical engagement might include:
- Mapping who supplies the data, who relies on it, and where errors or uncertainty create the most cost and friction.
- Capturing written requirements and expert knowledge as a clear, prioritized validation approach.
- Building a proof of concept around representative files so the most important assumptions can be tested early.
- Delivering a maintainable workflow that fits your existing systems and can be refined as your requirements change.
Start with the data you need to trust
You do not need to begin with a finished specification or a plan for a large platform. A representative set of files, the people who work with them, and a few examples of costly errors are often enough to identify a useful first step.