About Charlemagne Labs
Charlemagne Labs builds Charley, an on-device AI assistant that helps organizations detect social engineering and prevent sensitive-data exposure across email, texts, messaging apps, and the browser. Grounded in AI threat research and benchmarking, Charley runs locally on each device, preserving business and customer data sovereignty while reducing human-error risk.
Q: Tell us about your background.
I worked in startups before joining Facebook (now Meta) as a product marketer in 2013. I developed a keen interest in privacy and policy-sensitive product work, culminating in a product launch that changed my career. I worked on the optional implementation of end-to-end encryption in Messenger in 2016, and I learned firsthand about the trade-offs a platform like Facebook faces between data privacy commitments and safety and security concerns. I transitioned into product management in trust and safety, and I haven’t looked back.
In 2023, I started a team at Facebook to support creators and influencers who were having bad experiences on the platform, most significantly with impersonation and account hacking. It’s there that I learned about social engineering and the types of attacks that were clearly starting to involve AI. I left to start Charlemagne Labs in 2025 because I grew convinced that commensurate defense against social engineering was technically possible.
Q: Tell us about your business/idea.
The idea is to offer employees and executives an intelligent, privacy-preserving defense against deception. Despite decades of security awareness training, the vast majority of cybersecurity incidents continue to start with human mistakes. In the new epoch of AI, all phishing becomes spear phishing. If humans were the weakest link before, it’s terrifying to consider what the battlefield looks like with adversaries that avail themselves of low cost open models.
Our company sells Charley, a B2B assistant that prevents employees from getting socially engineered by intervening to classify risk before the risky action or click can even take place.
Q: What was the original inspiration for your company/product?
In addition to the work I was doing professionally to protect creators and influencers at Meta, my elderly mother suffered a “vishing” attack. She didn’t succumb financially but suffered irrevocable paranoia about risk and cut herself off from the internet, which has harmed our family. I started a lifelock subscription for her and learned quickly that everything that happens there is reactive. I grew to believe we need proactive, preventative, and privacy-preserving technology to combat the scourge of deception. Sadly, I don’t think there is much daylight between skepticism and credulity – either you’re going to doubt a message, a call, or an email is legitimate or you’re going to walk straight into the propeller. I grew convinced we need to fight fire with fire.
Q: What’s your vision for the future …“What will the market you are pursuing look like in 5-10 years?”
In 5 to 10 years, executives and employees will orchestrate, manage, and optimize the activity of a range of AI agents. These agents will require quality assurance, but also the freedom to improve on their own and interact with other agents inside companies and outside the perimeter at other companies. The market for what we are building consists of a sum total of human decision points. To the extent humans are in the loop, they will exercise judgement or taste. We secure the human decision point.
In the era of AI, 5 to 10 years could see enormous paradigmatic shifts in what work looks like. This means that security priorities will shift, and they will vary by role and function and company size. My vision is that Charley will adapt and grow in personalized ways to secure each person based on their risk profile. If everyone inside the network has a smart, context-aware exoskeleton then the human layer is secured.
Q: How does your business address pressing cyber and data challenges for the commercial sector?
The past six months have been spent scrambling to re-achieve some kind of equilibrium with respect to frontier AI risk in cybersecurity. It’s absolutely appropriate to prioritize code and infrastructure vulnerabilities, but we’re now returning to the status quo ante where the easiest initial access point is to trick a person.
Q: What attracted you to the DataTribe Foundry? Why did you choose to participate in the DataTribe Challenge?
I’m drawn to the connection that the partnership has to national security. It’s quite evident that state actors, with both financial and geopolitical motivation, treat citizens and private corporations as targets, and my documented and validated hypothesis is that social engineering and deception will surge. Defending our nation and defending our free enterprises converge.
Q: What’s your long-term vision for your business?
My vision is that we scale customizable, agile, accurate security assistance for employees and executives across all the devices they use, in such a complete fashion that we protect the integrity of the open internet and maintain the possibility that workers of the future can adopt AI and especially AI agents without creating fatal security risks.
