Secure Web Filtering & Shadow AI Governance

Shadow AI refers to the use of AI-powered tools and platforms without formal IT approval or governance. Unlike traditional shadow IT, Shadow AI often involves permanent, irreversible data exposure with limited auditability.

Employees routinely paste customer personally identifiable information (PII), financial records, source code, strategic plans, and proprietary data into public AI tools such as ChatGPT, often unaware this data may train public models or be retained indefinitely by third-party providers.

Blocking generative AI tools entirely stifles innovation, reduces workforce productivity, and drives employees to workarounds that create even greater risk (e.g., personal devices, unapproved accounts).

The right approach is secure enablement: allow AI usage within controlled guardrails that enforce corporate identity, prevent data training on public models, and mask sensitive information in real time.

Our solution transforms the browser into a secure gateway, replacing the outdated choice between “block everything” or “ignore the risk” with a govern and enable model that protects your organisation while empowering your workforce.

1. Shadow AI Discovery and Visibility

Gain immediate insight into:

• Which AI tools your workforce is using (ChatGPT, Claude, Copilot, Gemini, Perplexity, and hundreds more)
• How frequently employees access these platforms
• Whether files or sensitive data are being uploaded
• Time-based usage patterns and high-risk behaviours

Visibility is delivered through an AI Insights Dashboard that turns dangerous assumptions into data-driven security decisions.

2. AI Guardrails and Safe Enablement

Rather than blocking AI entirely, enforce secure usage policies:

• Mandatory corporate login: Require users to authenticate with organisational accounts before accessing AI platforms
• Disable public model training: Automatically lock settings such as “Improve the Model” to OFF, preventing user prompts from training public AI models
• Block desktop AI applications: Force AI tool usage into the browser where security policies can be applied consistently

These controls enable your teams to leverage the productivity benefits of generative AI while ensuring intellectual property and customer data remain protected.

3. Real-Time Data Loss Prevention (DLP)

Protect sensitive information at the browser level before it leaves your control:

• PII masking: Automatically detect and remove personally identifiable information (names, addresses, tax file numbers, Medicare numbers, financial data) before submission to AI platforms or external websites
• Content redaction: Sensitive data is masked in real time. AI tools still process the formatting request, but the actual data never reaches the cloud
• Obfuscation controls: Blur sensitive on-screen data unless the user actively hovers over it, protecting against visual shoulder-surfing

Custom data patterns can be defined using regex or keyword lists tailored to your organisation’s unique risk profile (e.g., client names, project code names, internal classification labels).

4. SaaS Application Control

Govern access to cloud platforms beyond AI tools:

• Allow download, block upload: Permit employees to retrieve files from platforms such as Dropbox or Google Drive while preventing data exfiltration
• Enforce corporate login: Ensure users access only organisation-approved SaaS accounts, blocking personal account usage on company devices
• Application-level policies: Apply different access levels based on user roles, departments, or risk classifications

5. Advanced Insider Risk Controls

Prevent data leakage through visual and physical channels:

• Screenshot blocking: Prevent users from capturing screen images of sensitive pages or applications
• Print protection: Block print-to-PDF or physical printing of controlled web content
• Forensic watermarking: Apply invisible, user-specific watermarks to on-screen content. If a photograph of sensitive data is taken and later uploaded or shared, the watermark identifies the individual responsible

6. Microsoft 365 and Copilot Security

Enhance Microsoft Copilot deployment with tenant-level protections:

• Commercial data protections applied at the organisational (tenant) level, not per-user configuration
• Security extends beyond the web browser to cover Copilot usage in Outlook, Word, Excel, and other Microsoft 365 applications
• Prevent data sharing and public model training across the entire Microsoft ecosystem

7. Compliance and Audit Transparency

For highly regulated industries, advanced logging delivers:

• Full audit trails of user activity, including AI prompts submitted (without capturing AI responses)
• PII-blocked event logs showing when policies prevented risky data exposure
• Executive-ready dashboards and reports aligned to ISO 27001, Essential Eight Maturity Model, and board governance requirements

Understanding the difference is critical for risk owners and decision‑makers.

Modern web filtering focuses not just on restriction, but on risk‑informed enablement.

No hardware appliances. No complex network reconfiguration. Protection follows the user across:

• Corporate office networks
• Home and remote work locations
• Public Wi-Fi and mobile connectivity
• BYOD and contractor devices

Deployment occurs via lightweight browser extension with centralised policy management through a cloud-based console.

• Fully managed: KMTech handles deployment, policy configuration, monitoring, reporting, and ongoing optimisation
• Co-managed: Your IT team retains day-to-day control while KMTech provides architecture, training, policy development, and escalation support
• Advisory and implementation: We design and deploy the solution, then hand over operational control to your internal team

Licensing is structured per user per month with tiered feature sets:

• Core: Web filtering, application control, basic reporting
• Core Plus: Adds Shadow AI protections, real-time DLP, screenshot/print blocking, watermarking, SaaS controls
• Premium: Includes AI prompt logging for highly regulated environments requiring full audit trails

Volume pricing is available for organisations with 100+ users. Australian-dollar pricing and flexible contract terms (annual or multi-year) align to your budget and procurement cycles.

Optional add-on:

• AI-Powered Analytics and Reporting (Agentic AI): Self-service executive dashboards with natural language querying (“Who is using AI most frequently?” “Show me PII-blocked events this month”). Single-license model provides organisation-wide visibility.