AI Agents Are Now Executing Phone Attacks | Kevin Tian artwork

AI Agents Are Now Executing Phone Attacks | Kevin Tian

MTS

October 15, 2026

Doppel co-founder and CEO Kevin Tian discusses how AI agents are supercharging social engineering attacks via multi-channel phone and video deepfakes. He outlines Doppel's risk signal ingestion platform and their security defense partnership with the San Francisco 49ers.

0:00

Sophia

Hello, everyone, and welcome back to MTS. Today, we are joined with Kevin Tian, who is the co-founder and CEO of Doppel, an AI-powered cybersecurity company defending against impersonation, fishing and deepfakes. And they just earlier today announced a risk signal ingestion, which combines real-world security data with employee behavior to help companies identify who is at most risk of social engineering attacks. And we're going to talk about their exciting partnership with the 49ers. So, Kevin, welcome to MTS.

0:31

Kevin Tian

Thank you for having me, Sophia. Excited to be here.

0:34

Sophia

Excited to have you as well. And welcome officially. I wanted to talk a little bit about the origin story of Doppel and also just the wider context that's happening in the world that makes this so relevant. We're seeing just improvements in everything from image models to deepfakes, everything that's happening. Tell me a little bit about the wider landscape, what's currently happening right now and why is Doppel important to solve these problems?

0:57

Kevin Tian

Yeah, I mean, well, we'll start off with how this company started, right? So me and Rahul, we were engineers at Uber and in 2022, he was actually roommates with one of the heads of research of OpenAI. So he got a sneak preview of a lot of the technologies that were coming out. Our hypothesis at the time was that the greatest threat with AI was going to be all things, social engineering, deepfakes, impersonation, phishing, fraud, and that's why we're called Doppel, short for Doppelganger. So I think over the past couple of years, we've seen a lot of that come to life. When you take a look at some of the biggest attacks that have happened in enterprise, even looking at sports leagues, and we've seen fake players emerge or interesting disinformation campaigns run.

So this is all the world that we live in today, and AI just makes it a lot easier to execute these threats.

1:49

Sophia

Yeah, definitely. When it comes to AI specifically doing social engineering, we talk a lot about different threats when it comes to how they get in our systems, what they have access to, permissions. But what we don't talk as much about is the social engineering. How do you actually engineer a human to do all this stuff? So tell me a little bit about what the wider landscape of social engineering attacks looks like today that maybe wasn't totally possible a few years ago.

2:16

Kevin Tian

Well, a couple of different things.

I think everyone has probably seen what a phishing email looks like, or have heard about a phishing email. But we're at the point now where we have AI agents in our social engineering defense platform. They're able to call out to people, they're able to shoot SMS messages to people. On average, when they call people, they get nine-minute conversations.

When you think about how does AI make things different, AI has enabled us to do these, what we call multi-channel attacks. So it's not just email, but do it over the phone, do it even on video conferencing. How do you know if I'm a deep fake or not? That's one way the world's changed. I think one other way the world has changed is, we're not just thinking about protecting humans at this point, but we've got agents on the other side. Sometimes an agent's picking up the phone, sometimes an agent's reading an email and writing a response. So how do we protect agents and humans from these social engineering attacks?

3:13

Sophia

Definitely. How exactly does risk signal ingestion play a role here? What exactly is it doing?

3:19

Kevin Tian

Yeah. I think the biggest thing is like if we see you, Sophia, are sharing a lot of things from Google Drive or O365, that maybe you're sharing it publicly, and you're not making sure that's limited to a specific audience, or maybe we see you execute certain behaviors, you don't have great password hygiene, don't use MFA, things like that. Those are all signals that you may be potentially risky.

So of course, you would never do that, but there are people who are executing behaviors within their workplace that make them more susceptible to security attacks. Yeah.

4:02

Sophia

I wanted to ask about this partnership with the 49ers, because typically when we hear about these tools, we're often associating them with, okay, how would this work at the enterprise company? And the typical company impersonating an employer. When you guys talked about the 49ers and how you could almost impersonate actual athletes or the different types of targeted attacks on sports organizations, it gets super interesting. So, I wanted to just hear a little bit more about what this partnership is and what are the unique things that these larger sport organizations might have when it comes to threats.

Read the other 7 minutes of this episode here.

$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.

Not what you expected? Full refund within 14 days.

Thousands of transcripts fetched by people building searchable podcast archives

Fetch transcripts with code

The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000796494380