Engagements

Three ways to
work together.

A structured engagement ladder — from boardroom diagnostic through technical architecture to executive rehearsal. Each step is mapped to specific domains of the Secucloud AI Security Framework, with fixed scope, defined deliverables, and clear ownership.

Grounded in two decades of enterprise cloud security architecture across AWS, Azure, Microsoft 365, Kubernetes, and modern identity platforms — because AI security advice is only as credible as the cloud foundation it sits on.

The questions that actually matter inside an enterprise AI rollout are not strategic. They are concrete, technical, and unforgiving.

Prompt injection in tool-calling agents. RAG poisoning through user-controllable context. Agent identity propagation across service boundaries. MCP server trust boundaries. Cross-tenant isolation in vector stores and embedding databases. Data lineage from prompt to response. LLM gateway controls. Shadow AI discovery across the SaaS estate. Plugin and tool-call authorisation. Model supply-chain risk.

Every Secucloud engagement is grounded in these specifics — and in the cloud-native architecture, identity propagation, segmentation, and assurance mapping that surround them. Strategic AI security advice without this technical floor is just opinion.

Each engagement opens the door
to the next.

01 FIXED SCOPE Assessment Diagnostic & roadmap 02 PER SYSTEM Architecture Per system review 03 EXECUTIVE Tabletop Executive exercise DIAGNOSE DESIGN REHEARSE

Built on the SAISF
framework.

Every Secucloud engagement is grounded in the same intellectual scaffolding: the Secucloud AI Security Framework. Six domains, five maturity levels, and an explicit mapping to NIST AI RMF, ISO/IEC 42001, the EU AI Act, and the OWASP LLM Top 10.

The framework is what turns subjective findings into a maturity score a board can read, an auditor can verify, and an engineering team can act on. It is also why every engagement produces an output that survives the audit cycle.

Explore the Framework in Full →
01
Govern
Authority & Accountability
02
Discover
Visibility & Inventory
03
Protect
Boundaries & Provenance
04
Secure
Architecture & Identity
05
Detect
Telemetry & Reaction
06
Assure
Testing & Evidence

Four ways to start.

01
SVC · 01 · Diagnostic

AI Readiness Assessment

A structured engagement covering all six SAISF domains, producing a board-ready map of where you are, where you need to be, and what stands between.

Most organisations have AI activity already underway — sanctioned tools, shadow Copilots, vendor-embedded features — without a coherent view of the risk it carries. The Readiness Assessment establishes that view. Discovery interviews, document review, technical sampling, and gap analysis produce a maturity score per domain, a prioritised remediation roadmap, and a one-page board summary.

  • Six-domain SAISF maturity scoring
  • NIST AI RMF & ISO 42001 gap analysis
  • EU AI Act exposure assessment
  • Shadow AI discovery sample
  • Prioritised remediation roadmap
  • Board-ready summary report
02
SVC · 02 · Technical Deep Dive

Secure AI Architecture Review

A focused one-to-three week review of a specific AI deployment — Copilot for M365, Bedrock, Azure OpenAI, custom RAG, or agent frameworks — assessed against modern threat models.

Where the Readiness Assessment surveys the landscape, the Architecture Review goes deep on a single system. Identity model, data boundaries, prompt injection resistance, tool-call security, network egress, monitoring strategy. Output is a written architecture critique, threat model, and remediation backlog ready for engineering teams to execute against.

  • Per-system threat modelling
  • RAG & vector store access review
  • Prompt injection & tool-call analysis
  • Identity & secrets management review
  • Logging & monitoring blueprint
  • Engineering-ready remediation backlog
03
SVC · 03 · Executive Exercise

Deepfake & AI Tabletop

A half-day executive simulation putting your board, audit committee, and finance leadership through three AI-era incident scenarios their current playbooks were never written for.

Voice-cloned CEOs authorising fraudulent payments. Deepfake video calls in the middle of a Teams meeting. AI-generated spear phishing that bypasses every awareness control you have. The tabletop is structured around three scenarios drawn from real recent incidents, run by a facilitator who has lived the cloud security side. Output is a debrief, a gap report, and three to five concrete playbook changes.

  • Three live scenarios from real incidents
  • Voice-clone, video deepfake, AI phishing
  • Designed for boards & finance
  • Facilitated debrief session
  • Gap report & playbook recommendations
  • Optional follow-up tabletop in 6 months

A consistent shape
across every engagement.

i.
Scope

A 30-minute discovery call to confirm fit, scope, and timeline. No obligation. No proposal until both sides agree the engagement is the right one.

ii.
Discover

Document review, stakeholder interviews, and technical sampling. We learn your environment before we offer opinions on it.

iii.
Synthesise

Findings are mapped to the SAISF framework and underlying standards. Recommendations are prioritised by risk reduction per pound spent.

iv.
Deliver

A written report, a board-ready summary, and a debrief session. The deliverables are designed to outlive the engagement and survive a regulator's eye.

Cloud Foundations

AI security sits on top of cloud security.

You cannot secure AI workloads without first securing the cloud they run on. Secucloud's AI specialism is built on years of cloud security architecture work across regulated industries — financial services, public sector, healthcare.

When AI engagements surface underlying cloud security gaps, we are equipped to address them as part of the same conversation rather than handing you to another supplier.

  • i.
    Cloud Security Architecture
    Landing zones, perimeter design, segmentation, and platform engineering across AWS, Azure, and Microsoft 365 — with a bias toward defensible patterns over feature theatre.
  • ii.
    Identity & Access
    Zero-trust design, conditional access, privileged identity, federation, and identity governance — extended naturally to AI agent and service principals.
  • iii.
    Data Protection
    Data classification, DLP, encryption strategy, key management, and data sovereignty — the foundational layer beneath every credible AI data boundary.
  • iv.
    Security Engineering
    Detection engineering, infrastructure as code, policy as code — translating security architecture into things that run, not slides that don't.
  • v.
    Compliance Architecture
    ISO 27001, SOC 2, NIS2, DORA — translated into cloud and AI controls that auditors can verify and engineering teams can implement.
Start a Conversation

Not sure which engagement fits?

A 30-minute scoping call. No pitch, no proposal until we both agree the engagement is right. The fastest way to decide is to talk.