Our planned AI capabilities tailor phishing exercises, explain simulation outcomes, recommend lessons and help employees apply approved security policies at work.
These capabilities are planned and are not yet available. Scope, integrations and availability will be confirmed after implementation and validation.
The proposed experience connects relevant practice with learning support and clear reporting. Every capability below is part of the development roadmap.
Generate exercise drafts around an employee’s role, language and approved learning objectives.
After an exercise closes, explain the warning signs and the employee’s decision using the approved scenario.
Recommend a short lesson and the next exercise’s difficulty for the programme owner to review.
Example learning experienceAn employee misses a suspicious request to change payroll bank details. The debrief explains the verification step, then suggests a short refresher.
Use quiz results and completed lessons to recommend a relevant learning path.
Explain difficult topics with examples and approved policy sources, and refer unresolved questions to the security team.
Draft role-based practice questions, explain answers and adapt approved lessons to the learner’s language.
Example learning experienceA customer-service employee practises account recovery, receives an explanation of a missed check and reviews the relevant policy.
Help employees apply approved security policies to everyday work questions, with source references.
Suggest brief reminders and learning activities for topics that need reinforcement, within an approved programme.
Explain which information may be shared with AI tools using the organisation’s approved guidance.
Example learning experienceAn employee asks whether a customer ticket can be pasted into an AI assistant. The coach cites the policy and explains permitted options.
Draft phishing and social-engineering exercises around a role, language and learning objective. Approved policy material would guide context; reviewers would control every release.
Editable scenarios with learning objectives, warning signs and review history.
Recommend short learning activities from quiz results and simulation outcomes. Explain each recommendation and let programme owners adjust it before assignment. After an exercise closes, provide a personalised explanation grounded in the approved scenario and observed response.
A learning path with reasons for each suggested next step.
Help employees navigate approved security policies through questions and practical examples. Answers would cite accessible source passages and escalate questions without sufficient evidence.
Source-linked guidance, with a clear route to the security team.
Turn campaign results into readable summaries for security leaders and programme owners. Link observations to actual metrics, comparison periods and supporting records.
Reviewable summaries, evidence links and suggested follow-up actions.
Plan role-play exercises about voice impersonation and deepfake requests using licensed or authorised synthetic media. Focus on independent verification and escalation.
Controlled voice scenarios, facilitator notes and verification exercises.
Adapt approved scenarios and lessons for different languages, roles and local terminology. Keep versions linked and require language review before publication.
Localised drafts, a shared glossary and an approval record.
Help the security team review employee-reported suspicious emails. Group related reports, highlight supporting indicators and suggest review priorities, with the final decision kept with an analyst.
An explained review queue, linked evidence and draft feedback for the reporting employee.
Turn programme goals, team needs and learning gaps into an editable campaign plan. Suggest audiences, timing and follow-up activities while checking exclusions, overlaps and training fatigue before approval.
A reviewable campaign calendar with audience choices, reasons and approval history.
Interpret simulation outcomes alongside exercise difficulty and audience context. For email exercises, use the NIST Phish Scale as a reference for human-reviewed difficulty ratings and explain the limits of each comparison.
Difficulty-aware comparisons, traceable metrics and suggestions for the next exercise.
Planned coverage for evolving phishing, identity fraud and AI risks, with reviewed scenarios, short practice and measurable learning outcomes.
AI-assisted drafts for email, QR-code, SMS, collaboration-message and callback scenarios. Begin with safe in-platform practice; live delivery through each channel needs separate validation.
Related capability: Phishing across channelsPractise responding to unexpected MFA prompts, device-code requests, app permissions and help-desk impersonation. Teach verification of the requested action, even when a sign-in page looks familiar.
Related capability: Deceptive sign-ins and access requestsRehearse account-recovery, employee-data and supplier-payment decisions with synthetic records. Approved voice role-play teaches verification through a known channel, beyond judging how convincing a voice sounds.
Related capability: Business fraud and voice impersonationPractise protecting customer data, checking AI answers and recognising instructions hidden in documents. Learn to inspect the recipient, data and scope before approving an AI assistant's action.
Related capability: Safe AI and assistant useRecommend short lessons from observed learning gaps, then revisit key decisions through spaced practice. Reviewed question banks, clear explanations and programme-controlled frequency keep the experience useful.
Related capability: Adaptive practice and reinforcementTrack reporting, time to report and retained knowledge alongside exercise difficulty. Show comparable cohorts and data limitations, with supportive follow-up instead of unsupported predictions about individual employees.
Related capability: Evidence of safer decisionsPrioritise reviewed email and in-platform scenarios, everyday verification decisions, learning debriefs, source-based policy answers and a clear measurement baseline. Start personalisation with transparent rules.
Add authorised voice exercises, validated delivery channels and analyst integrations after the learner journey works. Review new threat sources and model changes before they affect a live programme.
Illustrative role-based scenarios for organisations across industries. Select a team to explore the proposed exercise.
A person impersonating a customer pressures an agent to bypass the account-recovery process.
Follow the approved verification process and escalate exceptions without disclosing customer information.
A message impersonating an employee asks HR to change payroll bank details without the usual verification.
Verify the request through a known contact channel and follow the approved payroll-change process while protecting personal information.
A supposed supplier asks for emergency maintenance access outside the normal ticket process.
Verify the contact and change request independently, then use the approved access workflow.
A voice message attributed to a senior executive requests an urgent supplier-bank-detail change.
Confirm through a known channel and follow the established payment-approval process.
These are proposed acceptance requirements. Hosting choices and integrations require agreement and technical validation for each deployment.
Separate draft, review and release permissions; retain campaign and content approval history.
Define tenant separation, permitted AI inputs, retention, deletion and provider data-use terms.
Validate SSO, user provisioning, access roles and reporting integrations against your organisation’s systems.
Test answer grounding, unsafe outputs and language quality; retain traceable evaluation and audit records.
Proposed priorities, with availability confirmed only after each capability passes validation. No release dates are committed here.
Scenario drafts, multilingual review, evidence-linked reports and simulation insights.
A learner pilot with adaptive follow-up, a source-grounded policy coach and reviewed campaign planning.
Analyst-assisted email triage, authorised voice exercises and validated channels for a wider organisational rollout.
Define the audience, policies, languages and integration requirements for an agreed pilot scope.
Read the AI security awareness guide