ZAP AI

Your cloud estate.
Governed by AI.

ZAP AI is not a generic chatbot. It knows your infrastructure — every resource, every finding, every compliance score. Ask it anything. When it finds something wrong, it fixes it.

  • Runs in your tenancy — your data never sent externally
  • Vendor-agnostic — Azure AI Foundry · AWS Bedrock · GCP Vertex AI
  • End-to-end — detects → fixes → approves → closes
ZAP AI
Connected · Live data
80%
Findings auto-remediated
vs manual engineer time
Seconds
Evidence pack generated
vs 3–4 weeks manually
0
Data sent externally
Runs in your tenancy
3
Clouds queried
One answer, live data
What ZAP AI does

Two things. Both critical.

ZAP AI Co-pilot

Ask your estate anything. Get specific answers from live data.

Traditional cloud governance means opening three consoles, running three reports and spending an hour assembling an answer. ZAP AI gives you that answer in seconds — from live data across Azure, AWS and GCP simultaneously.

ZAP AI Remediation

Finds a problem. Writes the fix. Closes the ticket.

Most cloud security tools tell you what's wrong. ZAP AI fixes it — writes the IaC code, raises the PR, guides the approval, applies the change and closes the ITSM ticket. End-to-end. Full audit trail.

ZAP AI Co-pilot

One question.
Your entire estate answers.

ZAP AI queries live infrastructure data across Azure, AWS and GCP simultaneously — your CMDB, scan results, compliance scores, FinOps data — all within your tenancy.

Compliance
What is my current DORA compliance score?
Which CIS controls am I failing across AWS?
Generate an FCA PS21/3 evidence pack summary
What findings are blocking my CE+ assessment?
Security
What are my top critical findings right now?
Which resources have public access enabled?
Show me all findings by owning team
Any shadow AI detected in my clusters?
FinOps
How much could I save on cloud spend?
Show me all unattached disks and their owners
Which business units are over their cloud budget?
What are my right-sizing opportunities?
Operations
Which backup jobs failed overnight?
Show me drift events from the last week
Which environments have patch compliance issues?
What is the health status of my landing zones?
ZAP AI Remediation

Find it. Fix it. Close the ticket.

From detection to ITSM closure — every step automated, governed and documented.

01

Detection

ZAP Compliant Scan identifies a non-compliant resource. Finding raised with resource ID, cloud account, region, CIS control reference, severity and owning team from governance tags.

02

Root cause analysis

ZAP AI analyses the finding. Identifies the non-compliant configuration. Determines root cause — misconfiguration, drift, missing policy. Checks for related findings on same resource.

03

IaC fix generated

ZAP AI generates the Terraform fix automatically. CIS-compliant. Minimal change — only what is needed to remediate the finding. Validated by ZAP Compliant Scan before PR is raised.

04

PR raised

PR raised in GitHub or Azure DevOps. Fix description in plain language. CIS control reference. ITSM ticket number linked. Scan results showing current vs desired state. Rollback instructions.

05

Platform Lead notified

Platform Lead receives notification. Finding summary. Proposed fix. Risk assessment. One-click approval link in GitHub or Azure DevOps.

06

Approval and apply

Platform Lead approves. Terraform apply executes automatically. Change applied to live infrastructure. Full change record created.

07

Re-scan and verify

ZAP Compliant Scan runs on the remediated resource. Control passes. Finding resolved. If scan fails — PR flagged, Platform Lead notified.

08

ITSM auto-closed

ServiceNow incident auto-closed. Resolution timestamp. Remediation method. Approver identity. Scan verification result. Full Git commit reference. Immutable audit trail entry.

Shadow AI Detection

Unapproved AI in your clusters.
Detected in seconds.

As AI workloads proliferate across engineering teams, unapproved models in production Kubernetes clusters create regulatory risk and data sovereignty exposure. ZAP detects them automatically.

What ZAP detects
Local LLM inference
Ollama LocalAI LM Studio llama.cpp server
Model serving frameworks
Hugging Face TGI vLLM NVIDIA Triton BentoML · Ray Serve
Development environments
Jupyter Notebook JupyterHub JupyterLab
Vector databases
Chroma Weaviate Qdrant Milvus
Behavioural detection
GPU-accelerated workloads without approval AI inference endpoints on internal networks Custom model containers by GPU pattern
Detection to closure
1
ZAP scans all Kubernetes clusters every scan cycle (default: every 6 hours)
2
Unapproved AI container identified — image, namespace, cluster, first detected
3
HIGH severity finding raised immediately in ZAP Dashboard
4
ServiceNow incident auto-raised — resource ID, image hash, owning team
5
Platform Lead and Security Lead notified with containment options
6
ZAP AI generates: Remove · Quarantine · Whitelist options
7
PR raised → approved → applied → re-scan confirms → ticket auto-closed

Governed AI — not prohibited AI

ZAP enables governed AI use. Approved workloads are whitelisted and monitored. Unapproved ones are detected, contained and documented. Engineering teams can use AI — with CISO visibility and regulatory compliance maintained.

Self-hosted AI

ZAP AI runs inside your tenancy.

ZAP AI is connected to your own AI model deployment — inside your tenancy, in your chosen region. Your infrastructure data and queries never leave your environment.

Azure AI Foundry
Recommended for: Azure-primary estates
  • GPT-4o · Phi-3 · Llama 3
  • UK/EU Azure regions
  • FCA · DORA · NHS data residency
  • ZAP connects to your Foundry instance
AWS Bedrock
Recommended for: AWS-primary estates
  • Claude · Llama 3 · Titan
  • AWS region of your choice
  • Data stays in your AWS account
  • ZAP connects to your Bedrock deployment
GCP Vertex AI
Recommended for: GCP-primary estates
  • Gemini Pro · Gemini Flash · Claude
  • GCP region of your choice
  • Data stays in your GCP project
  • ZAP connects to your Vertex endpoint
Who uses ZAP AI

One AI. Every team uses it differently.

CISO

Morning briefing in 30 seconds

What are my top critical findings across all three clouds?
Any shadow AI detected in the last 24 hours?
What is my DORA score vs last month?
Full picture before 9am — without opening a single cloud console.
Platform Lead

Remediation without the toil

Show me all HIGH findings with ZAP AI fixes ready
What changed in prod-lz-02 in the last week?
Which open PRs are waiting for my approval?
Platform Lead reviews and approves — no investigation, no manual triage.
Compliance Manager

Audit prep in seconds

Generate an FCA PS21/3 evidence pack summary
What findings were remediated in the last 30 days?
Show me compliance trend across all frameworks
3–4 weeks now takes seconds. Evidence always current.
CFO / FinOps

Cost visibility without the spreadsheet

How much cloud waste identified this month?
Which business units are over budget?
Show me right-sizing opportunities by cost centre
Cost attributed from day one. Every saving documented.
Frequently Asked Questions

Everything you need to know
about ZAP AI.

Eight questions our customers ask most. Tap to expand.

No. ZAP AI runs entirely within your own cloud tenancy. You choose the model provider (Azure AI Foundry, AWS Bedrock or GCP Vertex AI) and ZAP connects to your instance. Your infrastructure data, compliance findings and queries never leave your environment.
Azure AI Foundry, AWS Bedrock and GCP Vertex AI. You choose based on your cloud primary and data residency requirements. Model selection within each provider is configurable.
ZAP AI can auto-remediate the majority of misconfiguration findings — storage access, NSG rules, encryption settings, backup policy drift and more. Findings requiring architectural decisions, data classification or access reviews are flagged for human-led remediation with ZAP AI analysis provided.
Yes. ZAP AI generates the IaC fix and raises a PR — but no change is applied without Platform Lead approval. This is a governance requirement, not a limitation. Every change has a documented approver and audit trail.
Yes. In addition to known AI container signatures, ZAP detects unknown AI workloads by behaviour — GPU usage patterns, inference endpoint exposure and model serving signatures.
EU AI Act enforcement begins August 2026. High-risk AI systems must be registered, documented and governed. Unapproved AI containers running in production clusters will create compliance exposure. ZAP provides the detection, governance and audit trail required for EU AI Act compliance.
GitHub Copilot and Microsoft Copilot are code assistance tools — they answer from training data. ZAP AI answers from your live infrastructure data — your actual compliance scores, your actual findings, your actual resources. It also takes action — generating IaC fixes, raising PRs and closing ITSM tickets.
ZAP AI remediation is deterministic — it generates fixes based on CIS benchmark requirements and your approved configuration baseline. The fix is validated by ZAP Compliant Scan before the PR is raised. Post-apply re-scan confirms the finding is resolved.
Free · 30 minutes · No commitment

See ZAP AI answer questions
about your estate.

30 minutes with our team. We show you ZAP AI live — compliance queries, shadow AI detection, remediation pipeline. No slides.

Book a session Talk to sales
No commitment · 30 minutes · [email protected]