Fable Security

Would your help desk reset MFA for a deepfake of your CEO’s voice? For years, the advice for a suspicious request was to pick up the phone and verify. That doesn’t work when an attacker can clone an executive’s voice from a conference talk or podcast clip in minutes and call your help desk with it. Most help desks have a verification process, but a familiar voice, a believable story, and a little urgency are often enough to get an agent to make an exception. Most teams know this, but few get to test it often. Phishing simulations stop at the inbox, and phone-based testing tends to happen once or twice a year. Fable’s AI vishing lets you test your shared phone lines, like your IT help desk or customer support, using authorized deepfake voices, including your own executives with their consent. Each call is graded against your actual verification policy. When the agent either does the risky thing or firmly refuses, the AI reveals it was a simulation and coaches them on the spot. You get the transcript and evidence to share with whoever owns the policy. Most people don’t take deepfakes seriously until they hear one. Book a live demo and we’ll place a deepfake call while you listen in. Get a demo from our experts here:https://lnkd.in/g7tFt-qc/?utm_campaign=Boost_Post_Engagement_Oct19,2026&utm_source=linkedin&utm_medium=paid&utm_id=1233405416&hsa_acc=514925310&hsa_cam=1233405416&hsa_grp=914811006&hsa_ad=1648685076&hsa_net=linkedin&hsa_ver=3

Attackers can now clone a familiar voice or join a video call as your CFO, and when the caller sounds exactly like someone you trust, recognizing their voice is no longer proof of who they are. Today we're sharing how Fable is helping security teams prepare for AI-driven social engineering: - AI vishing tests on the shared phone lines attackers go after first, like the IT help desk (now in early access) - Deepfake video training in Fable AI Studio, so finance teams, executive assistants and help desk analysts can rehearse the moment an "executive" asks for something urgent - Human-led red-team engagements for the complex, multi-channel scenarios automation can't cover Every test produces evidence, and every finding turns into targeted follow-up for the people who need it, all on one platform. Jacob Berry breaks down what's changed and how we're approaching it in our new post. Link in the comments.

Text, voice, and video attacks don't all work the same way, and lumping them together as one "AI threat" hides what's actually changing. On October 20, our Field CISO Jacob Berry sits down with Stephanie Carruthers of Coalfire for a field guide to generative AI social engineering. Stephanie brings real-world observations from Coalfire client environments, redacted and drawn from across industries. Together they'll cover how attackers use AI differently in each channel, how high-volume AI campaigns differ from targeted human-led attacks, and why one-off phishing simulations struggle to keep up. It's a 45-minute conversation with 15 minutes of live Q&A, focused on what's showing up in the field rather than broad AI fear. Save your seat: https://lnkd.in/gZjEuHfp

Insurance carriers know how to underwrite risk: Gather data, model exposure, monitor, and reduce loss. So why is human risk still managed with annual training and click-rate metrics? Join Fable for a live look at: Six attack patterns hitting insurance employees now. Fable's four-stage model: Identify, Quantify, Intervene, Track. How documented behavior change supports compliance. Underwrite Your Human Risk: Wed, Sept 23, 10:00 AM ET

It's live. Our Cybersecurity Awareness Month game is here, and you can play it right now! Six minutes. Five real workplace moments, built on the AI-enabled threats we're tracking today, not last year's playbook. Play it, then share it with your team, or anyone you want to help get more cyber secure. If you own OCAM planning at your company, request admin access and your whole campaign comes already assembled: reporting on exactly who completed it, the option to customize the deepfakes with your own team's voice and video, and swag shipped straight to your office. No integration work required, and free to use. Links to play the game and how to request admin access are in the comments 👇

Ask your security team what share of the company is using AI, and they'll probably answer from what they can monitor: the ChatGPT tab, the Claude login, the Cursor session. That number is real, but it's the floor, not the ceiling. We spent 90 days analyzing detected employee behavior across roughly 290,000 employees at more than 30 organizations to build The AI Behavior Index. This is our first look at how AI-enabled employees behave differently from everyone else, broken down by role, and what that means for where security controls need to go next. A few findings stood out. AI-enabled employees are measurably better than average on the controls security teams have spent a decade funding, and measurably worse on a behavioral layer most organizations have never instrumented. That risk doesn't spread evenly across the org, either: the function most likely to write your AI policy is often the one most exposed to it. We're publishing Volume 1 today. Download at the link below.