Insight from Incode: Unpacking the Rise of AI-Powered Scams

In this article, Technology Magazine speaks with Ricardo Amper, Incode’s Founder and CEO, about how autonomous AI agents are changing the economics of fraud. Drawing on Incode’s Agentic Fraud Report, the interview explains how agents now handle target research, weeks of trust building, and money movement on their own, so attacks can run around the clock and at scale. Romance, investment, and business email compromise scams are the most exposed because they depend on long, personalized conversations. The piece also explains why enterprises should treat video calls and voice as one signal among many when they confirm identity. Amper’s advice is to verify identity at the point of entry and keep checking that the person, account, device, and behavior make sense together through every higher-risk action.

Read the transcription of this article, published by Technology Magazine on September 16, 2026.

Founder and CEO Ricardo Amper breaks down a $579bn threat landscape where automated bots execute end-to-end target research, relationship building and thef

The rise of autonomous AI agents has fundamentally broken the traditional economics of cybercrime. 

By removing human labour from target research, multi-week trust building and money movement, fraud operations can now run continuously at a huge scale for a fraction of the cost. 

Incode’s Agentic Fraud Report documents 66 independently sourced incidents – including 44 confirmed AI cases with documented methods – highlighting a landscape where global fraud losses reached US$579.4bn in 2025.

The threat is clearest where scam sophistication meets high automation potential, led by romance fraud and closely followed by investment, imposter scams, business email comprise (BEC) and employment scams. 

To understand how enterprises must adapt to 24/7 hyper-personalised attacks and the collapse of traditional authentication signals, we spoke with Ricardo Amper, CEO and Founder of Incode. 

Incode’s Agentic Fraud Report outlines how AI agents are taking over every step of the scam sequence from initial target selection to managing money movements. How does an attack where AI acts as the operator differ fundamentally from traditional, human-led cyber fraud?

The biggest difference is the removal of the human bottleneck. Traditional fraud still required people to find targets, research them, start conversations and keep them going. An agent can do those things continuously, and can run many attacks at the same time.

That changes the economics of fraud. AI is not simply helping a fraudster write a better message or create a convincing image; it can potentially manage the sequence of actions itself. The fraud operation becomes something that can run continuously, adapt to what the victim does and scale without a corresponding increase in human effort.

The report notes agentic AI lowers the technical barrier so almost anyone with a target list can launch an attack, while pushing labour costs down. What does the rise of off-the-shelf fraud-as-a-service tools mean for the sheer volume and speed of attacks businesses should expect?

We should expect more attacks, and we should expect them to arrive faster. The important change is that the attacker no longer needs to build every part of the operation themselves. The tools, infrastructure and increasingly the intelligence can be bought or rented.

That makes fraud much more accessible. As the cost of generating sophisticated attacks falls so quickly, businesses need to focus on how they can identify and stop them.

Why are investment, business email compromise and romance scams particularly exposed when an AI can run dozens of 24/7 hyper-personalised conversations simultaneously?

These scams depend heavily on conversation and trust, which are precisely the areas where AI can remove the human labour involved. A romance scam, for example, can require weeks of interaction before the victim is asked for money. That used to mean that a person had to maintain that relationship.

An agent can maintain many conversations at once, remember what has been said and tailor the next interaction to the individual.

Investment fraud and BEC have similar characteristics; research and personalisation make the attack more convincing, but historically also made it more expensive to run. AI changes that equation.

With studio-grade voice cloning, photorealistic deepfakes and live video generation becoming accessible, what happens when fraudsters can manufacture live interactions on demand? Are traditional authentication signals like video calls or voice prompts completely obsolete?

Video and voice should no longer be treated as definitive proof of identity. They remain useful signals, but the assumption that a live interaction is inherently trustworthy is becoming much harder to defend.

The more important shift is that identity has to be established using multiple signals rather than one convincing interaction. A video call can tell you what someone appears to look and sound like. It does not, by itself, tell you whether that person is who they claim to be.

If proving humanity online is no longer as simple as checking an email address or a static liveness test, what does a modern identity and trust architecture actually look like? Where should enterprises be focusing their defensive investments today?

The fundamental change is that “human” is no longer the same thing as “trusted”. An attacker can potentially present a highly convincing human identity without being the legitimate person behind it.

Enterprises therefore need to establish trust around the identity rather than just the interaction. They need stronger identity verification at the point of entry, combined with ongoing assessment as the user moves through higher-risk actions. The objective is to understand whether the person, account, device and behaviour make sense together, rather than relying on a single check performed at the beginning.

About Technology Magazine

Technology Magazine is a digital platform from UK-based BizClik Media that connects technology executives at global brands. It covers digital transformation, AI, cybersecurity, and digital infrastructure through news, interviews, events, research, and data services.

Read the full article