Independent Innovation Lab - Est. 1998

How do you implement
responsible AI ?
Sedasoft built the answer
before the question was asked.

Responsible AI - auditable, evidence-based, provenance-tracked, compliance-ready - is now the central challenge for every serious AI deployment. The EU AI Act's high-risk provisions come into force in August 2026. Organisations are scrambling to retrofit accountability onto systems that were never designed for it.

Sedasoft built the answer before the question was asked. Not because they predicted the regulation, but because they have always held to the principle that compliance isn't a layer you can add. It is a foundation you build on.

SedaSoft takes its name from its founders: Seamus and David. That is not a branding decision. It is a statement about what this has always been - a collaboration between two minds, from different intellectual traditions, who kept arriving at the same conviction.

The Founding Belief
“Structure before scale.Provenance before generation.Accountability before deployment.”

These are not responses to the current AI moment. They are the principles Sedasoft built from since 1998. Every architectural decision in SiteEngine AI, from the citation system to the knowledge graph to the compliance dashboard to the token-level audit trail, is their direct expression.

Why a Lab, Not a Company

Built to be right,
not built to ship fast.

The great technology research labs - Bell Labs, Xerox PARC or the teams who built the foundational internet protocols - shared one characteristic: they were protected from the pressure to commercialise before the work was ready. The result, in each case, was that they built infrastructure that the world is still running on decades later.

SedaSoft has worked the same way. While the industry was scaling fast and shipping first, Seamus and David were asking whether what they were building was actually correct - whether the structure was right, whether the provenance was genuine, whether the accountability was real. That is a slower process. It produces a different kind of result, a body of work that is deeper, more coherent and more defensible than anything built under commercial pressure.

25+ years in production

The architecture that underpins SiteEngine has been running in live environments since 1998. One resource, Management-Issues.com, has published over 7,200 articles across 25 years using SiteEngine without interruption.

Comprehensive documentation

Every major component of SiteEngine AI is documented to research standard - literature reviews, comparative analyses, reproducible benchmark evaluations.

A consistent track record of firsts

First European military online recruitment system. First municipal searchable historical archive in the UK. First production Efficiency Engine in RAG. The pattern of arriving early and building properly spans three decades.

The People

The depth and the range.

Most technology is built by people who think the same way and learned from the same tradition. SedaSoft was not. David came to technology through History - through the discipline of treating information as evidence, of building structures that make knowledge trustworthy over time. Seamus came through engineering and systems thinking - through the discipline of building things that hold under pressure, that can be documented, defended, and improved. One thinks in meaning. The other thinks in mechanism. Both, as it turns out, are necessary for what they set out to build.

Lead Architect & Researcher

Seamus Waldron

The Range

Seamus Waldron As a problem-solving innovator, Seamus has built a career in technology which spans across data intelligence, enterprise architecture and edge computing. As Founding Partner of Sedasoft, his collaboration with David Bosdet has been defined by years of market-leading innovations, almost all of which stemmed the original architecture of SiteEngine, a declarative, object-based information system that turned out to be, in effect, a vector database before vector databases existed.

He now builds bespoke AI-powered knowledge systems that combine Retrieval-Augmented Generation, knowledge graphs and cognitive AI to support auditable, evidence-based decision-making, including applications within UK local government.

His most recent innovations include the Efficiency Engine architecture the first per-query carbon accounting framework for AI inference at the application layer and Hydrate, the first ever tool to give LLM coding assistants portable working memory across sessions, tools and models.

He has also led data warehouse transformations for financial institutions operating under SOX and FCA compliance, built ETL and migration systems for local government meeting statutory data retention obligations and advanced software and hardware solutions for central government agencies.

In their earlier days, Sedasoft created the first web-based online recruitment system in Europe for an international bank and the first European military on-line recruitment system for the UK Royal Air Force. They later built the first municipal searchable online historical archive in the UK, a subscription publishing platform for niche busines information and an online corporate learning platform whose concept of 'communities of practice' predated Google Circles by several years.

Earlier in his career, Seamus was a Senior Research Engineer at Corechange in Boston, where the company's portal software and signal sign-on technologies defined the corporate portal market. He was also a Systems Analyst at MAID Plc where he co-invented Profound for Internet - one of the first business intelligence alerting systems for knowledge workers.

Solutions Architect

David Bosdet

The Depth

David Bosdet David studied history at Cambridge University, then taught himself an entire technical discipline from scratch - starting from building websites by hand in the mid-1990s before moving onto information design, data architecture and infrastructure for online applications. The result, acquired methodically over three decades, is a level of information architecture expertise that many formally trained engineers do not reach.

That background is an important factor in SedaSoft has built. A historian doesn't treat information in the same way as a computer scientist - as data to be stored and retrieved. A historian treats information more as evidence, as something to be structured, cross-referenced, sourced and weighted by reliability.

His interest in foundational AI and LLM research has underpinned the development of SiteEngine AI as a local-first RAG pipeline - in the citation system, in the knowledge graph's entity relationships, in the confidence scoring, in the evidence provenance trails. These represent a particular way of thinking about how information works and how this can be expressed in code.

David was part of London's Shoreditch technology scene in the 1990s, among the first generation of practitioners who laid the foundations for the modern web we now take for granted. As a Solutions Architect, he has operated ever since at the intersection of business complexity and technical possibility: defining strategy, translating requirements into structure, identifying risk early and ensuring that what gets built is not just scalable and secure, but meets long-term business objectives.

David’s career has included several years as head of software development at a financial service e-learning company specialising in compliance. He also founded a company producing ‘walk through’ videomontage imagery of renewable energy projects based on the Cesium 3D Geospatial platform.

David's photography portfolio on Flickr has over eight million visits. The work of someone who sees structure and pattern where others see noise - expressed in a different medium, with the same precision.

“Structure before scale. Provenance before generation. Accountability before deployment.”

If you are an investor, an acquirer or an organisation that needs reliable AI it can actually trust, we would welcome a conversation.

Start a conversation