SynapsedBuilding the engineering science of Trustworthy AI.
Since 2025, Synapsed has been transforming Trustworthy AI from high-level principles into practical frameworks, lifecycle methodologies, assessment models, open technologies and evidence-based assurance.
- Applied research
- Open community
- Engineering methods
- Industrial adoption
Trustworthy AI is not achieved through governance, security or testing alone. It emerges from their continuous integration through evidence.
AI capabilities have advanced faster than our engineering disciplines.
Organisations now have access to principles, regulations and standards. What they often lack are practical methods for building, assessing and continuously verifying AI systems.
Synapsed conducts applied research at this exact intersection. We connect governance with engineering, translate requirements into lifecycle activities, define the evidence needed to demonstrate trustworthiness, and build tools that make assurance repeatable.
Integrating governance, engineering and assurance.
Trustworthy AI emerges from their integration, supported by measurable evidence throughout the AI lifecycle.
Governance
Set direction, define policies and ensure responsible AI.
Principles, standards and accountability.
Engineering
Build trustworthy AI systems by design and by default.
Research, methods and lifecycle practices.
Assurance
Verify, validate and continuously monitor AI behaviour.
Tools, testing and measurable evidence.
Six connected areas of research.
From organisational governance to the behaviour of AI systems in operation.
AI Governance
Governance models, responsibilities, decision-making and regulatory alignment.
AI Engineering
Lifecycle engineering methodologies and implementation practices across the AI lifecycle.
AI Assessment
Methodologies for evaluating organisational readiness and governance maturity.
AI Testing
Practical testing for LLMs, RAG applications and agentic AI systems.
AI Assurance
Evidence collection, verification and continuous assurance throughout the lifecycle.
Continuous Trust
Runtime monitoring, behavioural validation and continuous trustworthiness.
Building the foundations. Connecting the ecosystem.
2025
Establishing the foundations
- Trustworthy AI education programmes
Translating principles into practical skills.
- OWASP AI Maturity Assessment
Matteo Meucci — Creator and Contributor. Organisational maturity methodologies developed with the OWASP community.
- OWASP AI Testing Guide
Matteo Meucci — Creator and Contributor. Practical AI testing guidance developed with the OWASP community.
- Patent activities
Supporting long-term research and innovation.
2026
Building the research ecosystem
- Joint Lab Fiable with FBK
Connecting academic research and industrial application.
- AI-T-SDLC research
Engineering trustworthiness across the AI lifecycle.
- AI Governance Framework & 5D Assessment
Turning research into governance methods and assessment tools.
- TrustInspect
Repeatable testing and an open-source initiative.
- OWASP Global AppSec
Knowledge transfer at AppSec EU in Vienna and AppSec USA training, scheduled for 4 November.
- Wiley/IEEE book
Preparation of Trustworthy AI Systems.
Research milestones and ongoing initiatives · 2025–2026
From research to engineering practice.
Synapsed AI Governance Framework
A comprehensive framework connecting standards, lifecycle activities, technical controls and organisational responsibilities.
Research focus
The framework translates trustworthiness objectives into practical requirements and identifies the evidence needed to assess how those requirements are implemented.
5D AI Governance Assessment
Assess your organisation across five dimensions and build a practical roadmap to AI governance maturity.
Research focus
The five dimensions are AI Team, Processes, Guidelines, Standards and Guardrails. The assessment helps identify gaps, establish priorities and guide continuous improvement.
AI-T-SDLC & Trustworthy AI Baselines
Engineering methodologies and baselines that integrate trustworthiness by design throughout the AI lifecycle.
Research focus
AI-T-SDLC connects analysis, design, development, deployment and maintenance. Trustworthy Baselines, developed within the book, connect lifecycle activities with responsibilities, expected outcomes and evidence.
TrustInspect
Evidence-based testing for LLM, RAG and agentic applications, translating research into repeatable tests.
Research focus
TrustInspect records prompts, responses and findings to support assessment, comparison after changes and continuous testing. The research programme includes an open-source initiative alongside contributions to community testing methodologies.
Publications & Knowledge Transfer
Research papers, the upcoming Wiley/IEEE book and international training and conferences bring research into engineering practice.
Research focus
The upcoming book, Trustworthy AI Systems, brings these connected methods together.
Joint Lab Fiable with FBK
Collaborative research on Trustworthy AI, connecting academic research and industrial innovation.
Research focus
Joint Lab Fiable connects Synapsed with Fondazione Bruno Kessler. It brings together research and industrial experience to develop and apply methods for engineering, testing and governing Trustworthy AI systems.
Open knowledge.
Protected innovation.
We contribute openly to the community while protecting innovations that require long-term investment to create real-world impact.
Publications, education, community contributions and selected open technologies extend the reach of our research. Alongside this open knowledge, we protect intellectual property that supports continued investment and industrial application.
Collaborating to build a trustworthy AI future.
We work with academic institutions, communities and industry partners to connect research with practical adoption.
Our ecosystem includes Joint Lab Fiable with FBK, contributions to OWASP AI projects, academic education and collaboration with organisations applying AI in real operating environments.