Skip to main content
NexITC
SERVICES/AI/RUN TIER/C12
C12 · AI · 12-MONTH MIN · RUN · OUTCOME · SECURE · CROSS-TAG · CYBERSECURITY

Govern what AI can see.
Control what AI can do.

C12 · AI SecurityOps™ is NexITC's managed AI-security operations subscription for organisations adopting enterprise GenAI, copilots and autonomous agents. Not AI portfolio governance. Not agent reliability operations. Not a recurring red-team engagement. A 12-month subscription operating enterprise AI discovery, AI-security policy enforcement, sensitive-data controls, agent/tool/MCP security posture, AI-security exception governance and release-level security assurance — with Practice Lead — AI as named account owner and Cybersecurity specialist support.

COMMITMENT
12 mo min
COMMERCIAL
Managed subscription
PRIMARY OUTCOME
Secure
C12 · AI-ADOPTED → AI-SECURED
C12
BASELINE
AI-adopted
STATE · AI-ADOPTED
C12
TARGET
AI-secured
STATE · AI-SECURED
BASELINE
NORMALISE
STEADY
REVIEW
AI SECURITY VISIBILITY
TRENDING ↑
SHADOW-AI RISK
TRENDING ↓
AGENT/ACTION COVERAGE
MEASURED
C12 · AI SECURITYOPS™
ILLUSTRATIVE
§ 00 · THESIS
01
AI GOVERNANCE DEFINES THE RULES.
AI SECURITY ENFORCES THE BOUNDARIES.

Employees can adopt AI applications without procurement. Developers can connect coding assistants. Business teams can expose enterprise data to copilots.

Agents can call tools, APIs and MCP servers and move from generating recommendations to taking actions. C12 operates the security boundary around those interactions.

STATE · AI-ADOPTED
Approved and unapproved AI coexist. Sensitive-data interaction is incompletely visible. Agent permissions expand project-by-project. Security policies exist but enforcement evidence is fragmented.
STATE · AI-SECURED
AI applications and agents inventoried. Sensitive interactions governed. Agent/tool permissions reviewed against approved boundaries. Security exceptions owned. Release-level security regression continuously evidenced.
§ 01 · OPERATING STREAMS

Six operating streams,
around the AI security boundary.

STREAM 01
CONTINUOUS

Enterprise AI discovery

Identify and maintain visibility over in-scope enterprise AI applications, workforce AI usage, production agents and material agent/tool/MCP connections. C12 discovers AI security surface. [[C2|C2]]'s agent registry governs portfolio status and lifecycle.

STREAM 02
CONTINUOUS

AI data-security policy enforcement

Operate controls governing sensitive data entered into, retrieved through or processed by in-scope AI applications according to approved policy.

OUTCOME
AI-SECURED
ENTERPRISE AI SECURITY
BOUNDARIES ENFORCED
STREAM 03
CONTINUOUS

Agent, tool & MCP security posture

Review material permissions, action boundaries, tool access and security controls around autonomous agents. C12 asks: “Can this agent access or perform something it should not?” [[C9|C9]] asks: “Is this agent behaving reliably while doing what it is authorised to do?”

STREAM 04
MONTHLY

AI security exception governance

Security exceptions around AI application use, sensitive data, user groups, agent permissions and tools reviewed with owners, expiry and escalation.

STREAM 05
RELEASE / CHANGE-TRIGGERED

Security regression assurance

Operate approved automated regression/security checks after material AI releases or control changes. Manual adversarial testing is explicitly excluded and remains [[A9|A9]].

STREAM 06
MONTHLY

AI-security executive scorecard

AI-security coverage, shadow-AI exposure, sensitive-data policy events, agent permission exceptions and regression-control status reviewed monthly.

EXPLICITLY NOT COVERED
Manual adversarial AI red teaming
That's A9 AI Safety & Red-Teaming Sprint™. A9 independently attacks the deployed system. C12 operates security controls between those assessments.
AI portfolio governance
That's C2 CoE-as-a-Service™. C2 owns use-case portfolio, charter governance, agent registry and pilot velocity.
Agent behaviour/reliability operations
That's C9 Managed Agent Operations. C9 owns output quality, task completion, functional tool-use anomalies and prompt drift.
SOC incident response
That's C7 SecOpsCommand™. AI-originated security events route into C7/customer SOC playbooks for investigation and containment.
Agent/application build
That's B14/B15/B20/B21 as appropriate.
§ 02 · ANNUAL CADENCE

Twelve-month subscription.
Baseline, normalise, operate.

M01 — AI security estate and policy baseline. M02–03 — security-control coverage normalisation. M04–12 — steady-state AI-security operations. Change-triggered — automated security regression. A9 — independent adversarial validation where required. M11 — annual review.

Q 01Q 02Q 03Q 04M01 · Estate & policy baselineM02–12 · Normalisation → steady-state operationsM11 · Annual reviewAI security estate & policy baselinedEND M 01Annual reviewM 11RenewalEND M 12OPERATING RHYTHMContinuous discovery & enforcement · Monthly exceptions & scorecard · Change-triggeredregressionNAMED ACCOUNTABILITYPractice Lead — AI (CEO escalation within 24 hours)
§ 03 · OPERATING MODEL

Discover, enforce, govern,
assure and report.

OPERATING MODEL · SIX STEPS
DISCOVER → REPORT
The operating model applied on every C12 subscription.
01
AI security-surface inventory
Identify what enterprise AI is security-relevant. Not every experiment requires continuous managed security. Production or materially sensitive AI does.
02
Risk classification
AI applications and agents are classified according to data sensitivity, action authority, business impact and exposure. The classification determines control depth.
03
Data-security enforcement
Approved policy translates into enforceable controls. Security teams gain evidence that rules are not merely written.
04
Agent/action governance
Agent identities, tool permissions and action boundaries remain visible as workflows evolve. High-risk write/privileged actions receive greater control.
05
Release security assurance
Material AI changes trigger security regression rather than waiting for the annual security review. Independent red-team assessments remain separately scoped.
06
Monthly executive review
AI and security leadership review one shared posture. This prevents AI governance from becoming a policy function disconnected from technical enforcement.
!
DISCLOSURE · INDEPENDENCE
C12 is not an AI-security software resale. The service may integrate AI security, DLP, endpoint/browser controls, agent-security technologies, application controls and customer-native platforms according to architecture and risk.
§ 04 · BASELINE VS MANAGED

From AI policy
to enforceable AI security.

BASELINE
TYPICAL STATE
STATE_01
AI usage partially known
STATE_02
Sensitive AI interaction difficult to evidence
STATE_03
Agent permissions project-specific
STATE_04
Security exceptions informal
STATE_05
Security validation episodic
C12 · CADENCE
MANAGED
STEADY-STATE
MANAGED_01
AI-security estate visible
MANAGED_02
Sensitive interactions governed
MANAGED_03
Agent/action permissions measurable
MANAGED_04
Exceptions time-bound
MANAGED_05
Release security assurance operating
MANAGED_06
Executive AI-security position measurable
↓ DISCOVER · CLASSIFY · ENFORCE · GOVERN · ASSURE · REPORT ↓

§ 05 · REPRESENTATIVE SCENARIO

A UAE enterprise,
from AI policy to enforceable AI security.

Illustrative composite — not a specific client.

SCENARIO / C12 / UAE ENTERPRISE · ILLUSTRATIVE COMPOSITE
COMMITMENT · 12 MO
SITUATION

A 2,000-user organisation has approved enterprise copilots while additional AI tools are being used by employees. Business units are also piloting agents connected to internal systems.

ENGAGEMENT

C12 establishes AI-security discovery, sensitive-data control, agent permission governance, exception management and release-level security assurance.

BOUNDARIES

C2 may simultaneously govern which use cases should exist. C9 may operate production-agent reliability. C7 SecOpsCommand™ (or the customer's existing SOC/IR process) owns security-incident response. The services remain distinct.

§ 06 · SERVICE ELEMENTS

Five service elements,
one AI security boundary.

E_01

AI Security Estate Discovery

Visibility of in-scope AI applications, agents and security surfaces.

E_02

AI Data-Security Controls

Sensitive-data policy enforcement.

E_03 · CORE

Agent / Tool / MCP Security

Security boundaries for autonomous actions and connected tools.

E_04

AI Security Exception & Regression Operations

Exception lifecycle and automated release security assurance.

E_05 · MONTHLY SCORECARD

AI Security Scorecard

Monthly executive reporting.

CADENCE
MONTHLY
§ 07 · OUTCOMES

Six outcome measures,
reviewed monthly.

Representative targets — not guaranteed results for a specific client.

01 · AI SECURITY VISIBILITY
TRENDING ↑
02 · SHADOW-AI RISK
TRENDING ↓
03 · SENSITIVE-DATA POLICY EXCEPTIONS
GOVERNED
04 · AGENT/ACTION SECURITY COVERAGE
MEASURED
05 · SECURITY-REGRESSION CURRENCY
MEASURED
06 · EXECUTIVE REVIEW
MONTHLY
§ 08 · FIT

Honest scoping.

PREREQUISITES
Move fast when these conditions are in place at onboarding.
01
Material enterprise AI adoption

02
AI/product/data and security sponsorship

03
Approved or identifiable AI policy boundaries

04
Relevant telemetry/control access

05
12-month commitment

NOT SUITABLE IF
These patterns indicate a different engagement is a better fit.
You only need AI use-case discovery

→ A11/A5

You need independent adversarial testing

→ A9 AI Safety & Red-Teaming Sprint™

You need agents built

→ B14/B15/B20/B21

You need agent reliability operations

→ C9 Managed Agent Operations

You need enterprise AI governance

→ C2 CoE-as-a-Service™

§ 09 · COMMERCIAL

Managed subscription.
Monthly cadence. No surprises.

COMMERCIAL MODEL
Managed AI-security subscription, 12-month minimum

Pricing based on protected workforce scope, AI application/agent estate, control complexity and agreed integration surface.

COMMITMENT & CADENCE

12-month minimum subscription with monthly delivery cadence. Renewal negotiated at annual review gate (M 11). Scope amendments negotiated through the Practice Lead.

INCLUDED IN SUBSCRIPTION
  • ✓Five named service elements
  • ✓Monthly executive AI-security scorecard
  • ✓Named Practice Lead — AI
  • ✓Change-triggered automated security regression
OUT OF SUBSCRIPTION
  • —Manual A9 red-team engagements — separately scoped to preserve validation independence
  • —Agent/application build (B14/B15/B20/B21)
  • —SOC incident response (C7)
COMMERCIAL PRINCIPLES
01

Subscription, not billable hours

No hourly billing. Subscription priced against service elements and SLA commitments agreed at kickoff.

02

12-month minimum commitment

The operating cadence needs time to establish. Shorter commitments produce onboarding costs without steady-state value.

03

Change orders authorised

Practice Lead has authority to negotiate scope amendments in the same conversation, not through a separate commercial cycle.

§ 10 · QUESTIONS

The questions AI and security leaders actually ask.

Q_01Is this the same as C9?

No.

C9 manages reliability.

C12 manages security boundaries.

Q_02Is this the same as C2?

No.

C2 governs which AI initiatives should exist and how the portfolio progresses.

C12 governs how enterprise AI interactions are secured.

Q_03Does C12 perform red teaming?

Not as part of the retainer.

A9 remains NexITC's independent manual adversarial-testing SKU.

Q_04What happens to an AI security incident?
The event routes into C7 or the customer's established SOC/IR process.
Q_05Can C12 cover workforce AI and autonomous agents together?
Yes. Both form part of the enterprise AI security surface, while their operating controls differ by risk.
§ 11 · NAMED ACCOUNTABILITY

One name.
Six accountabilities.

Specialist consulting means the person who onboards the subscription is the person who owns the cadence — with escalation to CEO on any material issue within 24 hours.

THE ROLE

Practice Lead — AI

Named account owner for the duration of the subscription, with Cybersecurity specialist support. Present at every monthly review. Available for escalation on material issues, with escalation to CEO within 24 hours.

SIX ACCOUNTABILITIES
01
AI-security operating cadence

02
Monthly executive review

03
Policy/exception governance

04
Security assurance coordination

05
Cybersecurity escalation

CEO within 24 hours.

06
SLA accountability

§ 12 · ADJACENT WORK

What runs before,
beside, and with C12.

A11 · ENTRY
PRIOR

Agentic AI Readiness

Entry prior.

Read A11 →
C9 · RUN
PEER

Managed Agent Operations

C9 owns output quality, task completion, functional tool-use anomalies and prompt drift.

Read C9 →
A9 · ENTRY
INDEPENDENT VALIDATION

AI Safety & Red-Teaming Sprint™

A9 independently attacks the deployed system. C12 operates security controls between those assessments.

Read A9 →
§ 13 · BOOK A CLINIC

30 minutes.
One AI-security question.

Bring the AI-security question blocking your board conversation. C12 is scoped in the clinic — AI estate, policy boundaries, telemetry/control access, sponsorship and commitment appetite. If C12 is not the fit, the clinic surfaces the honest alternative.

CLINIC · C12
  • —Enterprise AI and agent estate
  • —AI policy boundaries
  • —Telemetry and control access
  • —Fit assessment against A9, C2, C7, C9
Practice Lead — AI attends every clinic.