Threat Intelligence & Research
Understand which AI threats and developments matter to your decisions.
Example outputs
- Threat or incident intelligence brief
- Sector threat landscape
- Technical research report
- Executive decision brief
Services
Research, standards and assessment for decisions about AI.
Bring a threat, requirement, research question or assurance decision. Explore the service routes and the outputs that can be agreed for your context.
Our services
Choose the question you need answered.
Standards Development
Turn governance and security requirements into specifications people can examine.
Example outputs
- Requirements and control specifications
- Framework mappings and gap analyses
- Technical specifications and evidence schemas
- Standards research and submission support
Applied Research
Investigate a question that existing evidence cannot adequately answer.
Example outputs
- Feasibility & exploratory studies
- Experimental & comparative studies
- Evidence & evaluation method development
- Research prototypes & proofs of concept
Planned · not available
Compliance Advisory
Connect applicable obligations with control and evidence requirements.
What to expect
- Detailed scope and deliverables to be published when available
Academy programmes · in development
Training & Professional Development
Build the knowledge and judgment needed to apply AI security and evidence methods.
What to expect
- AI security
- Incident response
- Evidence practice
Certification in development
Assessment Services
Examine what defined evidence supports—and where the conclusion must stop.
Example outputs
- Evidence Readiness Assessment
- Bounded Control Assessment
- Bounded Assurance Engagement
SERVICE AVAILABILITY
Read availability with each offer.
Commissioned research, standards development, applied research and defined assessments are open for scoping. Compliance Advisory and Academy programmes remain in development. Certification is not operating.
Published frameworks and methods are separate from the availability of a service. Acceptance of an engagement depends on scope, competence, independence and delivery capacity.
Notably Absent: Published work and scoped assessment do not establish accreditation, regulatory endorsement or operating certification. A framework mapping does not demonstrate control effectiveness.
A first engagement
Start with the right scope.
01
Describe the decision
Tell us what you need to establish, the system or workflow involved and the records available.
02
Agree a paid scope
Confirm competence, independence, access, outputs, ownership, confidentiality, fees and timing.
03
Receive findings and next steps
Review the agreed output, what the evidence supports and the limitations that affect its use.
Research, standards work and assessment have different entry points. An Evidence Readiness Assessment is one route; it is not a prerequisite for every engagement.
Published work
Explore our research and methods.
Publications help you understand the questions and methods behind our work. They do not represent client delivery claims.
Explore Research → Explore Assurance →Other requirements
Your question may not fit a service label.
We welcome relevant enquiries beyond the examples listed. Describe the decision or problem; we will assess whether we can accept an appropriate scope.
Introductory discussions establish fit and commercial scope. Technical analysis, study design, validation and advisory work require an agreed paid scope, independence and data-handling checks, written terms and payment before work begins.
Discuss your requirement →SYSTEM AND EVIDENCE CONTEXT
Find the work in your system context.
PHYSICAL AI
Evidence for systems that act.
Explore applied research and defined assessments of authority, controls and records along a physical action path. Security assurance and functional safety remain distinct.
Explore Physical AI evidence →PLATFORM INTEROPERABILITY · IN DEVELOPMENT
Preserve meaning between systems.
Explore the developing evidence structures and tooling direction, and discuss a research or specification question. This is not a production integration offer.
Explore platform interoperability →