Cognita by Protecht

Cognita is the AI built into the Protecht platform to help risk teams capture, assess and understand risk in plain language. Unlike general-purpose AI tools, Cognita is grounded in Protecht’s trusted methodologies and authoritative GRC content, ensuring responses are aligned with real-world compliance practices and your own data.

Cognita works directly within the Protecht platform and respects your configuration, securely interpreting information in context, rather than generating generic responses.

No. Cognita recommends. People decide. It is designed to support your teams, not replace them, with human review and approval built into every interaction.

It automates repetitive tasks, improves the quality and consistency of records, and explains complex data in clear language, allowing your teams to focus on oversight and decision-making. Cognita never saves a change or overwrites what you’ve typed without your approval, and people stay accountable for decisions.

Version history keeps a record of what changed, so reviewers can see when a change was made and challenge it.

Cognita acts as an intelligent assistant, guiding users with clear, contextual explanations as they complete tasks, so users learn as they go. People with less risk experience can work with confidence, and experienced users spend less time checking others’ work.

• Answers questions about your risk data with AI Analytics. Ask in plain language and get answers with charts, so people get insight without waiting for a report. AI also helps users create Liveboards easily: describe the chart you want and AI helps build it, and Verified Liveboards show everyone which numbers are agreed. Liveboards include AI Highlights that point out what stands out, and Deep Analysis lets you ask deeper questions of the data.
• Adds an AI summary to scheduled reports, covering key themes, trends, priority takeaways and links to key items. Leaders get the main points without reading every page.
• Supports risk and control self-assessments with Risk Assessment Flow. Cognita reviews a risk owner’s risks, considering linked controls, related risks and what the owner tells it about their role. It sorts the risks into those ready to approve, quick updates and those needing a closer look, then suggests changes for each and explains each suggestion. The owner decides what to accept.
• Helps users log new incidents with AI Drafting. Users describe what happened in their own words and add photos or documents. Cognita turns this into a complete draft record for them to check, so they don’t need to know which fields to fill in or what each one needs. Incident logging is quicker, easier and more consistent.
• Helps users complete and improve records with AI-Assisted Editing. Users click Suggest on a field to get a recommended response when working on incidents, risks and controls. Cognita considers the record’s details and linked records, and explains its reasoning. Users can accept a suggestion, ask for more detail or choose their own response.
• Sums up what is linked to a record, such as new incidents, controls or audit findings. Users can see what has changed and what it means without opening every record.
• Answers questions about how to use Protecht and good risk practice. It draws on Protecht best practice Academy courses and help guides, with citations, so users get help when they need it.
• Assists with platform navigation by taking users to the page they need.

Yes. When Cognita answers questions about risk practice or how to use Protecht, it first searches Protecht’s own curated content, such as Academy courses, help guides and blogs, then builds its answer from what it finds. This approach is known as retrieval-augmented generation (RAG). It keeps answers grounded in trusted information instead of general internet knowledge, reduces the risk of bias, hallucinations and misleading answers, and lets Cognita reference where its answer came from.

When Cognita makes suggestions on a record, it works from that record and the linked records you have access to.

Whether it’s helping log an incident, explaining risk information, drafting a report summary or answering questions about your data, Cognita shows what the data says. AI-drafted content is clearly marked, and nothing is saved until you accept it.

Safety checks block harmful content and attempts to misuse Cognita. Subject-matter experts test outputs before release, and we keep monitoring them after launch to find and fix issues.

Cognita adheres to ethical AI principles, including human-in-the-loop oversight. People stay accountable for material decisions: Cognita prepares, surfaces and recommends, and people decide. All AI-generated outputs are subject to user review, so accountability and control remain with your teams.

Cognita works within your configuration and existing permissions, so it only uses records you already have access to. Records show when Cognita helped create or edit them. Customers control permissions for both human users and Cognita. Version history and audit metadata keep a record of what changed.

Yes. Users can always accept, edit or reject suggestions before they are saved. Cognita never enforces decisions or makes changes without user approval.

Version history keeps a record of what changed, so reviewers can see when a change was made and challenge it.

Yes. Your data is encrypted, hosted in your region, and never used to train AI models. Each customer runs in a private, single-tenant AWS environment, and no data is shared between customers. Cognita only accesses data required to deliver the requested functionality. It is designed for risk and compliance data, and sensitive details such as credit card numbers and government IDs are automatically redacted.

Cognita supports auditability through a combination of platform audit logs, version history and clear marking of AI-generated content. Protecht records user actions taken on records and workflows, and where AI-generated text is used, it is visible to users for review prior to saving. In Risk Assessment Flow, version history and audit metadata record what changed.

The level of AI interaction logging available can vary by feature and configuration, and Protecht can confirm the audit options for your environment.

Cognita is designed to help customers meet requirements such as ISO 31000, GDPR, the EU AI Act, the NIST AI Risk Management Framework (AI RMF), and, in Australia, APRA CPS 230, the Privacy Act, the AI Ethics Principles and the AI6 Framework. It enables compliance activities but does not override regulatory obligations or governance processes.

No. Cognita is optional and can be enabled or disabled at any time, so teams can adopt AI at their own pace.

Book a free demo, contact your Protecht account representative or customer success manager, or register your interest in upcoming Cognita webinars and demonstrations.