AI Assistant helps you proactively investigate and respond to security events in Web Security Analytics. It combines conversational guidance with intelligent analysis and remediation workflows to help you identify threats faster and reduce manual effort.

AI Assistant includes:

View manager. Build, edit, apply filters, and switch views effortlessly using everyday commands.
Security view analyzer. Get an on-demand analysis of your current Web Security Analytics view and further investigation recommendations based on the displayed data, applied filters, and selected time frame.
Remediable insights. When AI Assistant identifies actionable insights, you can use its guided workflows to mitigate threats. For example, you can directly initiate, build, and deploy custom rules, or set an attack group's action to deny to block malicious requests.
Custom rule initiation. Create and deploy custom rules instantly without manually entering variables. When you accept an Akamai-suggested remediation, the agent either launches an in-chat guided workflow to build and deploy the rule for you, or it opens the custom rule builder so you can review the rule before activation.
Documentation assistant. Get a concise explanation of a feature with links to the relevant documentation for more details.

Access and position AI Assistant

You can open and interact with AI Assistant from any Web Security Analytics view except Attack Insights, where AI Assistant is not available. Simply click the AI Assistant button in the bottom right corner of your current Web Security Analytics view.

AI Assistant remains available as you navigate between Web Security Analytics views. Use the dock icon to dock AI Assistant or move it freely. Use the arrows on the AI Assistant button to collapse it to the right edge of your screen. AI Assistant remembers your docked or undocked settings.

👍
AI Assistant maintains your session when you leave Web Security Analytics and return later, allowing you to continue your previous interaction.

Manage your view

You can build and modify your current Web Security Analytics view by using either suggested prompts or natural language requests. As you interact with AI Assistant, the view updates automatically to reflect your changes.

Use suggested prompts

Click Help me build or edit the view to display suggested prompts. Select a prompt to perform common tasks, such as adding a widget, applying a filter, changing the time range, or switching to a different view.

Enter your own requests

If you prefer not to use predefined prompts, you can type intuitive, everyday commands directly into the agent’s input field. This feature lets you build, customize, or filter your dashboard views without digging through dropdown menus.

Sample commands

You can ask the AI Assistant things like:

Show me traffic from the last 24 hours.
Display only SQL injection attacks.
Add a filter to show only mitigated requests.
Take me to bot analysis view.
Switch back to my first view.
Add a bar chart that shows attack types.
Add a widget showing connecting IP addresses.
Create a custom rule based on the current filters applied.

👍
If AI Assistant can't process your request, try to rephrase the query or break it down into simpler terms. Our team continuously improves the AI Assistant’s understanding of human language, and we encourage you to report any limitations you encounter.

Analyze your currently displayed view

AI Assistant evaluates your currently displayed view in Web Security Analytics (including any applied filters and your selected time range) and generates a contextual explanation of your displayed dataset. It presents you with:

Overall activity trends within your selected time range.
Notable anomalies, abnormal activity, or emerging patterns.
Related CVEs associated with the observed behavior.
Suggested focus areas for further investigation.

📘
View explanations are supported in all Web Security Analytics views except Samples, Attack Insights, and Alerts.

To analyze your active view:

Open any view in Web Security Analytics and apply your desired filters or timeframes.
Open AI Assistant and click the Explain this view to me prompt.
If you change your filters or update the dataset, click New chat in the panel to update the analysis. Next, select Explain this view to me again.

This chat does not support follow-up questions regarding view explanations at this time. However, you can still ask for help building custom views, adding widgets, or locating documentation.

Remediable Insights

AI Assistant proactively surfaces remediable insights when it detects unusual traffic patterns, automatically suggesting remediation actions. These insights help you investigate potential threats and launch guided remediation workflows directly from AI Assistant.

Unlike Attack Insights (which highlight sporadic traffic spikes and provide a weekly geo-industry digest for your independent research), Remediable Insights surface traffic spikes already investigated by the system and offer you guided remediation workflows when a mitigation option is available.

Based on its background investigations, the system identifies the most effective defense for your specific threat and presents it as the Akamai-suggested remediation. You can choose to apply this guidance, for example, by denying an attack group or deploying a custom rule, or you can explore alternative mitigation strategies.

Each remediable insight details the detected activity, explaining exactly what the system identified and why it requires your attention. You can review the technical details, explore the suggested remediation options, and launch guided workflows to apply your chosen defense.

Deny an attack group by updating your existing Web Application Firewall rule. When you start a guided remediation workflow, AI Assistant performs the following actions:
- Identifies the active security configuration.
- Creates a new version of the configuration to preserve the active version while preparing the changes.
- Identifies the Web Application Firewall rule associated with the detected attack.
- Updates the rule behavior to deny requests that match the identified attack category.
- Activates the updated security configuration to production so the changes take effect.
Create and deploy a custom rule immediately, without having to enter the rule variables or activating security configuration. If you choose to create a custom rule, the AI Assistant either launches a guided setup wizard directly inside the chat window to build and deploy the rule for you or takes you to the custom rule builder where you can review the generated rule before saving and activating it. To learn more about completing this process, see Create custom rules with AI Assistant.

Access remediable insights

You can access and review remediable insights through external notifications or directly within the Web Security Analytics interface:

External notifications. When you receive an alert through your configured channel and click the provided link, the system redirects you to Web Security Analytics. The dashboard automatically applies the necessary filters to display the exact data associated with the problem, allowing you to review the issue and select a remediation action. Learn how to set up remediable insight notifications
In-app AI Assistant. When you open AI Assistant in Web Security Analytics and it displays active insights as a list of quick-action bubbles. The Web Security Analytics dashboard dynamically adjusts based on your current configuration context:
- Active remediable insights in your current configuration. You see a prompt that says Show me remediable insights. Click it and then select one of the insights to automatically adjust your Web Security Analytics filters, load the pre-curated threat data, and view your remediation options.
- No active insights in your current configuration. You see a prompt that says Show me remediable insights in other configurations. Click it to view up to five other security configurations that currently have active threats. The list displays the configuration names. They are sorted by the number of total remediable insights in descending order. Click the configuration name to switch the view to a different configuration and see remediable insights generated for it.

Remediable insight statuses

Each remediable insight has a status that indicates its current state and helps you track investigation and remediation progress:

New. AI Assistant generated a remediable insight but you have not viewed or acted on it yet.
In progress. You have started a remediation workflow for the insight, but have not completed the remediation action yet.
Remediated. Remediation completed successfully.
Retrying remediation. Last remediation attempt was unsuccessful. The system is automatically retrying remediation.

📘
Status visibility in AI Assistant
The agent automatically displays all your active and new remediable insights without explicit status labels. You only see these status labels when you open an insight using a direct link from your SIEM.

Set up remediable insight notifications

You can configure the system to alert you about new remediable insights through your preferred communication channels:

In Web Security Analytics, click All actions.
Select Set up remediable insights notifications.
Choose your target email group and notification channel. Learn how to configure Security Notification Settings.

After you configure the notification settings, you receive automated alerts detailing potential threats directly to your specified channels or mailbox.

Find relevant documentation with AI Assistant

AI Assistant provides contextual links to official technical documentation directly within the chat, helping you understand product features and behaviors without leaving your workflow.

When you ask about a specific feature or interact with tools like remediable insights, AI Assistant evaluates your query and provides a direct link to the relevant Akamai TechDocs article.

Feedback and opt-out

Feedback. Click the thumbs up icon if you are happy with a result, or thumbs down to share feedback. Your inputs remain completely confidential.
Opt-out. If you want to opt out of this feature entirely, click the Opt out link available in the AI Assistant footer.

Data protection, privacy, and model information

When you use AI Assistant, Akamai protects your data through strict security frameworks, robust encryption, and isolated processing environments.

How the AI models work

AI Assistant combines Akamai-developed machine learning models with industry-leading large language models to assist you:

Llama 3.3 70B. Powers the primary agent functionality, including event analysis, custom prompts, and guided remediation workflows.
GPT-4o. Handles documentation searches and feature explanations.
GPT 5.4. Analyzes dataset views, filters, and time-frame contextual trends to power your Web Security Analytics view analysis.

These models simplify how you investigate threats by translating natural language into database filters and commands. The AI evaluates your filtered data to generate deep insights securely, ensuring all analysis happens exclusively inside Akamai’s environment so your customer data never leaves the Akamai controlled system.

Data privacy and storage

Akamai never uses your customer data to train the underlying AI models. The system strictly uses your data to generate outputs during an active session.

To improve your future interactions, Akamai stores your prompts and the system’s responses for up to 120 days.

Secure isolation. To maintain strict security, the architecture stores prompts and performance data completely outside of the active AI processing engine.
Operational use only. Akamai retains this data solely to enhance future responses, such as remembering your specific filter logic. For example, if you ask for WAF-triggered events from the past week, Akamai stores your prompt and the system’s logic (such as Attack type == WAF and time range == past week). The models do not store actual customer data or Web Security Analytics results.
Anonymized fine-tuning. We fine-tune the agent based on positive feedback, exclusively using anonymized data.

Security and compliance protocols

Akamai’s Legal and InfoSec teams strictly govern AI Assistant. The platform actively protects your information at every level through the following measures:

Access control and data security. Robust access controls and strict segregation ensure you only query the specific security configurations you manage. No user can access another user’s data.
No sensitive data processing. AI Assistant protects your information by segregating user input and enforcing strict security measures, including robust access controls and encryption for your prompts and outputs both in transit and at rest. While you can query and view all security configurations within your account, the system strictly verifies your user permissions before it allows you to make changes. For example, you can only execute a remediable insight if you have the explicit rights to modify that specific configuration.
Encryption. The system secures your stored data using AES 128-bit encryption at rest and protects data in transit using TLS 1.2.
Internal processing. Akamai processes all your data internally and never transfers your information to third-party vendors.
Compliance standards. The application fully complies with GDPR, HIPAA, and FIPS 1.42 frameworks, alongside the EU AI Act provisions.
Continuous security. Akamai builds the agent using secure coding practices, incorporating strict image signing and hardening to guarantee software integrity and minimize vulnerabilities. We also run regular vulnerability assessments (DAST and SAST) and enforce real-time monitoring to detect and alert you to abnormal behavior.
Input classifiers. The system segments your input into strict classifications to prevent data manipulation. If a query falls outside these classifications, the agent blocks it and returns an error message.

Akamai customer data is not stored or shared with OpenAI or Meta. Learn more

📘
Akamai's AI governance and compliance
Akamai confirms that it is operating its business in compliance with all applicable laws, including AI regulation and guidelines.
Akamai has put in place a comprehensive principle based AI Governance Program to balance innovation and technical excellence with responsible use and human control
The Akamai AI Governance Program and the Ethical and Responsible AI Principles are publicly available in the Akamai Trust Center.
This program is built upon 8 principles that aim for technical excellence based on human accountability. AI use case risk assessments are performed by Akamai Global Data Protection.