Two Incidents, One Lesson in Trust

Javvad Malik | Nov 20, 2026

Evangelists-Javvad MalikOptimizely unfortunately suffered two incidents in nine months, so clearly they must be doing something wrong and the answer to that is to buy a bigger firewall.

But is that really the answer?

One incident started with a phone call. The other started with a flaw in a signup form and ended with malicious code served to visitors of the BBC, CNN, Zoom, and HSBC. On the surface they have nothing in common. But dig a little deeper and you’ll notice that criminals didn’t get in through brute force or exploiting a zero-day vulnerability. They exploited trust. And when trust is the attack surface, your workforce is both the target and a critical line of defense.

February: It's Good to Talk (To the Wrong Person)

On February 11, a threat actor got in touch with Optimizely to say they already had access to its systems. They weren't bluffing. According to the company's own breach notification, the way in was a "sophisticated voice-phishing attack." Someone picked up a phone, and the person on the other end said the right things.

Optimizely says the attacker couldn't escalate privileges, install software or plant backdoors, and what left the building was basic business contact information. SecurityWeek reported that the systems touched included internal business platforms such as Zendesk and Salesforce. Although Optimizely didn't name the group, the tactics seemed consistent with that of ShinyHunters.

While it’s easy to downplay that it’s “just” contact data, it’s valuable information that can be used to craft even more convincing phishing emails or phone calls. A breach of business contacts isn't the end of an attack. It can often be the reconnaissance phase of the next one, aimed at somebody else.

November: The Script Everybody Trusted

On November 16, Scott Helme and the team at Report URI spotted malicious JavaScript in Optimizely files used by 178 organizations, including the BBC, CNN, Zoom, Salesforce, and GitLab. Optimizely told BleepingComputer the attacker had abused a flaw in its account signup flow, which let them insert code into customers' Project JavaScript.

The payloads included card skimmers, crypto wallet address swappers, and routing that pushed iPhone users towards an exploit kit.

Not one of those 178 organizations was breached in the traditional sense. Their own systems were fine. They had simply trusted a vendor's script.

To be clear, there's no public evidence that the February and November incidents are connected. One was social engineering, the other a software flaw. But that’s kind of the point. They are two very different routes to one destination.

The Inbox Is Not the Only Door

For years, security awareness was quietly treated as a synonym for "don't click the dodgy link in the email."

Email is still the most popular avenue, but look at where the attacks this year on Optimizely actually came from. A phone call, a helpdesk and CRM platform, and a trusted third-party script running in a browser. Not one of those is an inbox.

That's the reality for most organizations now. Work happens across email, Teams, Slack, Zoom, SMS, voice calls, shared documents, and a long tail of SaaS tools, each with its own notifications, its own login prompts, and its own sense of "this is internal, so it's safe." And that’s before you factor in the increasing number of AI agents.

Vishing groups ring the helpdesk pretending to be an employee. They message staff on Teams pretending to be IT. They follow up an email with a phone call to make it feel real. The more channels we use, the more places trust can be borrowed.

This is why collaboration security matters. Protecting all channels across wherever people talk. They need to have the same defenses to protect them against social engineering whether it comes through an email, a chat message, or a voice note.

Why Training Matters More Than Ever

The February breach is the obvious case of a convincing voice and a person who wanted to be helpful. While you want your helpdesk to be helpful, you need to equip them with the skill to pause when a call, a message or a request seems a bit odd and verify it through a trusted channel.

The recent November breach initially appears like a purely technical problem. Which is partly true. Consider what happens next when customers of 178 organizations now have reason to be worried about their card details. Criminals know that a worried person makes for an easier target. Expect "your account may have been affected, please verify your details" emails, texts and calls, because every high-profile incident is followed by a wave of opportunistic phishing riding on its coattails. Stolen business contact data, like the kind taken in February, makes those follow-ups more targeted and more believable.

Therefore training has to change shape. Annual click-through modules about suspicious emails won't be enough when the threat is an authoritative and polite voice claiming to be from IT, or a Teams message from a "colleague" asking you to approve an MFA prompt. Simulations need to reflect the channels people actually use. Coaching needs to happen at the moment, not once a year. The goal isn't to turn everyone into a security expert. It's to build a culture where checking is normal and nobody feels silly for saying "I'll call you back on the number I've got."

But training isn't a magic bullet either. People will make mistakes, or overlook issues, especially when the attack arrives on a platform they believe is safe. That's why the human layer and the technical layer have to work together to catch what they can across every channel, and trained people who catch what slips through, with each one feeding the other.

What Security Leaders Can Do This Week

  • Train for all your channels, not just the inbox. Run vishing simulations, especially for helpdesk and support teams who are paid to be helpful and are therefore prime targets.

  • Make call-back verification a habit. Any request to reset credentials, approve MFA, or move money gets verified through a known number or channel, never the one the request came in on.

  • Extend protection to collaboration tools. Treat Teams, Slack, and other platforms as attack surfaces, with the same scrutiny you give email.

  • Brace for the follow-up phish. After any big public incident, warn staff and customers that scam messages exploiting it are likely, and tell them how you will and won't contact them.

  • Know your third parties. Inventory the scripts and SaaS platforms you trust, monitor what they do, and rotate credentials and tokens when a vendor has an incident.

  • Make reporting easy and blame-free. The colleague who says "I think I just spoke to a scammer" within five minutes should be celebrated as a win, never shamed for making a mistake.

Trust Is the New Perimeter

Optimizely isn't unique, it's simply a well-known name. The lesson is that trust is now the thing attackers exploit, whether it's trust in a voice on the phone, a message in Teams or a script on a web page.

That's why at KnowBe4 we think about securing the whole workforce. Attack Simulation & Training builds the instincts people need across every channel, Email & Collaboration Security protects email and the platforms where work really happens, and Agent Security looks after the AI agents that increasingly sit alongside your staff. Each layer makes the others stronger.

None of it replaces a person who stops, thinks and says "hang on a minute."

Secure the Digital Workforce: Human + AI

KnowBe4 empowers the modern workforce to make smarter security decisions every day. Trusted by more than 70,000 organizations worldwide, KnowBe4 is the pioneer of digital workforce security, securing both AI agents and humans. The KnowBe4 Platform provides attack simulation and training, email and collaboration security, and agent security powered by AIDA (Artificial Intelligence Defense Agents) and a proprietary Risk Score. The platform leverages 15 years of behavioral data to combat advanced threats including social engineering, prompt injection, and shadow AI. By securing humans and agents, KnowBe4 leads the industry in workforce trust and defense.