Apr 28, 2026 - Timo Rüppell - Most security teams have accepted a hard truth in recent months. AI has already arrived in your organization. It resides in your codebase, runs in your cloud infrastructure, and is likely open in a nearby browser tab right now.The question is no longer whether to let AI in. It is whether you have the visibility, controls, and audit trail to prove you have it under control.Our Q1 2026 release is our most comprehensive platform update to date, built specifically to answer that question. Here is a look at what is new, why we built it, and what it means for your team.Enhanced AI Discovery for Code, Cloud and WorkforceFind Every AI: Discovery Gets a Major UpgradeYou cannot govern what you cannot see. This remains the fundamental problem in AI security. It gets harder as AI spreads across more environments, platforms, and areas of your codebase.‍This update significantly expands where and how FireTail finds AI across three dimensions:In Your CodeIf your development team uses Azure DevOps, FireTail can now scan those repositories directly. We have also added LangChain AI agent detection. If your engineers are building agentic workflows, FireTail surfaces them automatically during analysis. We have also extended Google Vertex AI SDK detection to Java, Go, and JavaScript.The highlight here is our new Software AI Agents view. You can now see every AI agent discovered in your codebase, complete with a direct link back to the relevant source code. For teams trying to get ahead of agentic AI risk, this is a major advantage.In Your CloudMicrosoft 365 is now a supported integration. This brings AI discovery from your M365 environment directly into FireTail. We have also added OpenAI usage log collection. Your OpenAI API activity now flows into the same visibility layer as your other cloud AI services. For teams running multi-account AWS setups, we resolved a long-standing limitation in Bedrock scanning. All accounts are now covered consistently.In Your WorkforceMost organisations have the biggest visibility gap in this area. We have invested heavily here. The new Workforce Applications view gives you a centralised picture of every AI tool your employees are actively using. New anonymous consumption tracking goes even further. It surfaces AI activity not tied to a specific user account, such as chat completions, image generation, and file uploads. You can now see the full scope of AI usage, including previously invisible activity.‍Unrivaled logging of all AI UsageLogging That Actually Tells You SomethingKnowing that AI activity is happening is one thing. Understanding what is sent to and from AI models is another. This release makes significant advances on both fronts.A Redesigned Employee Logs ExperienceThe Employee Logs page has been completely rebuilt. Log rows are now clickable. They open a detailed view with full chat history, device information, a resources panel, and a raw data toggle. This allows your analysts to inspect the exact request and response payloads. The initial message is now surfaced as a dedicated column. This makes it far faster to triage suspicious activity at a glance.PII, Secrets, and Prompt Injection Detection EverywhereSensitive data scanning now covers prompt content and chat message fields directly. Prompt injection detection, suspicious Unicode analysis, and URL scanning have been extended to additional fields in workforce logs. This closes gaps where certain content was slipping through unanalysed. We have also improved false positive rates for IBAN detection. Adding checksum validation alongside pattern matching resolves a significant source of noise in some environments.Ask in Plain English, Get Filters InstantlyOne of our favourite features in this release is AI-powered natural language filter generation. Analysts can now type a plain-language description of what they want to see instead of manually building filter logic. The platform automatically converts this into a ready-to-use filter. It sounds simple, but it meaningfully reduces the time between a desire to investigate and seeing the relevant data.A Full Audit Trail: Automatically RedactedA new Audit Logs page gives you a detailed, searchable record of every action taken within your FireTail organisation. Sensitive data is automatically redacted. This makes the audit trail safe to share with compliance teams and external auditors without manual review.‍New policy capabilities and compliance frameworksGovernance That Keeps Pace With How AI Is Actually UsedWe are most proud of the governance section in this release. The volume of new capabilities is significant, but the philosophy behind them matters more. Effective AI governance is not about blanket blocking. It is about giving the right people access to the right capabilities with the right guardrails in place.Guardrails for AI Content PoliciesGuardrails are now available as a first-class feature within workforce policies. You can browse the library of available guardrails and use them as templates when creating new policies. This dramatically speeds up the process of getting sensible baseline controls in place. Pre-made policies are also available for the most common governance configurations.Control What AI Can Do, Not Just Who Can Use ItNew capability policies let you control which AI features specific users and groups can access, rather than just which platforms they can reach. For example, you might allow your creative team to use image generation while restricting the same capability for the finance team. You can also set device-based policies to control which devices are permitted to access AI tools.Content blocking capabilities round this out. These prevent specific types of content from being sent to or received from AI services at the policy level.New Compliance Frameworks: ISO 42001, OWASP Agentic AI, NIST AI 600, and OWASP LLM Top 10Compliance mapping has never been more critical with the EU AI Act enforcement deadline arriving in August 2026. FireTail now supports four additional security frameworks: ISO 42001, OWASP Agentic AI, NIST AI 600, and the OWASP LLM Top 10. You can filter findings against any of these frameworks. This makes it significantly easier to demonstrate your compliance posture to auditors, boards, and regulators.Posture Management: Always Current, Never StaleSecurity findings in FireTail now update automatically when underlying resources change. Your posture view reflects reality, not just the last time you ran a scan. We have also fixed several accuracy issues that caused findings to be flagged as updated unnecessarily. Full functionality has been restored to auto-remediation. The result is a posture management experience that is quieter, more accurate, and more trustworthy.Platform Polish: The Improvements That Add UpBeyond the headline features, this release includes a number of platform-wide improvements:The navigation has been reorganised for a more logical layout. "Workload" replaces "Inventory", and Alerting is repositioned accordingly.The default dashboard is now AI Usage, keeping the most relevant information front and centre when you log in.Filters across the platform now show logos for services, model providers, platforms, and devices. This makes it much faster to find the right filter value.The Integrations page has been redesigned with a cleaner tab-based layout.Bulk actions are now available for integrations, AI models, and prompts. This saves significant time when managing larger inventories.The Bigger PictureEvery feature in this release was built with the same conviction. Security teams should be able to say yes to AI. Governing AI adoption confidently requires visibility to know what is happening. It requires controls to enforce what is allowed. Finally, it requires an audit trail for proof.That is what full-spectrum AI security looks like in practice. This release is the most complete version of that vision we have shipped to date.All these features are live now. If you want to see them in action, schedule a demo today. We would love to show how FireTail gives you unparalleled visibility, security and control across your entire AI ecosystem.‍

The post From Shadow AI to Full Control: FireTail’s Q1 2026 Updates – FireTail Blog appeared first on Security Boulevard.

FireTail - AI and API Security Blog

Source: Security Boulevard
Source Link: https://securityboulevard.com/2026/04/from-shadow-ai-to-full-control-firetails-q1-2026-updates-firetail-blog/