ブログ
/
Email
/
July 18, 2023

How Darktrace SOC Thwarted a BEC Attack

Discover how Darktrace's SOC detected and stopped a Business Email Compromise in a customer's SaaS environment.
Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Nicole Wong
Cyber Security Analyst
Default blog image
18
Jul 2023

What is Business Email Compromise (BEC)?

Business Email Compromise (BEC) is the practice of tricking an organization into transferring funds or sensitive data to a malicious actor.

Although at face value this type of attack may not carry the same gravitas as the more blockbuster, cloak-and-dagger type of attack such as ransomware [1], the costs of BEC actually dwarf that of ransomware [2]. Moreover, among UK organizations that reported a cyber breach in 2023, attacks related to BEC – namely phishing attacks, email impersonation, attempted hacking of online back accounts, and account takeover – were reported as the most disruptive, ahead of ransomware and other types of cyber-attack [3].  

What makes a BEC attack successful?

BEC attacks are so successful and damaging due to the difficulty of detection for traditional security systems, along with their ease of execution.  BEC does not require much technical sophistication to accomplish; rather, it exploits humans’ natural trust in known correspondents, via a phishing email for example, to induce them to perform a certain action.

How does a BEC attack work?

BEC attacks typically begin with a phishing email to an employee of an organization. Traditional email gateways may be unable to block the initial phishing email, as the email often appear to have been sent by a known correspondent, or it may contain minimal payload content.

The recipient’s interaction with the initial phishing email will likely result in the attacker gaining access to the user’s identity. Once access is obtained, the attacker may abuse the identity of the compromised user to obtain details of the user’s financial relations to the rest of the organization or its customers, eventually using these details to conduct further malicious email activity, such as sending out emails containing fraudulent wire transfer requests.  Today, the continued growth in adoption of services to support remote working, such as cloud file storage and sharing, means that the compromise of a single user’s email account can also grant access to a wide range of corporate sensitive information.

How to protect against BEC attacks

The rapid uptake of cloud-based infrastructure and software-as-a-service (SaaS) outpaces the adoption of skills and expertise required to secure it, meaning that security teams are often less prepared to detect and respond to cloud-based attacks.  

Alongside the adoption of security measures that specialize in anomaly-based detection and autonomous response, like Darktrace DETECT™ and Darktrace RESPOND™, it is extremely beneficial for organizations to have an around the clock security operations center (SOC) in place to monitor and investigate ongoing suspicious activity as it emerges.

In June 2023, Darktrace’s SOC alerted a customer to an active BEC attack within their cloud environment, following the successful detection of suspicious activity by Darktrace’s AI, playing a fundamental role in thwarting the attack in its early stages.

Darktrace Mitigates BEC Attack

Figure 1: Screenshot of the SaaS Console showing location information for the compromised SaaS account.  The ability to visualize the distance between these two locations enables a SOC Analyst to deduce that the simultaneous activity from London and Derby may represent impossible travel’.

It was suspected the attack began with a phishing email, as on the previous day the user had received a highly anomalous email from an external sender with which the organization had not previously communicated. However, the customer had configured Darktrace/Email™ in passive mode, which meant that Darktrace was not able to carry out any RESPOND actions on this anomalous email to prevent it from landing in the user’s inbox. Despite this, Darktrace/Apps was able to instantly detect the subsequent unusual login to the customer’s SaaS environment; its anomaly-based approach to threat detection allowed it to recognize the anomalous behavior even though the malicious email had successfully reached the user.

Following the anomalous ExpressVPN login, Darktrace detected further account anomalies originating from another ExpressVPN IP (45.92.229[.]195), as the attacker accessed files over SharePoint.  Notably, Darktrace identified that the logins from ExpressVPN IPs were performed with the software Chrome 114, however, activity from the legitimate account owner prior to these unusual logins was performed using the software Chrome 102. It is unusual for a user to be using multiple browser versions simultaneously, therefore in addition to the observed impossible travel, this further implied the presence of different actors behind the simultaneous account activity.

Figure 2: Screenshot of the Event Log for the compromised SaaS account, showing simultaneous login and file access activity on the account from different browser versions, and thus likely from different devices.

Darktrace identified that the files observed during this anomalous activity referenced financial information and personnel schedules, suggesting that the attacker was performing internal reconnaissance to gather information about sensitive internal company procedures, in preparation for further fraudulent financial activity.

Although the actions taken by the attacker were mostly passive, Darktrace/Apps chained together the multiple anomalies to understand that this pattern of activity was indicative of movement along the cyber kill chain. The multiple model breaches generated by the ongoing unusual activity triggered an Enhanced Monitoring model breach that was escalated to Darktrace’s SOC as the customer had subscribed to Darktrace’s Proactive Threat Notification (PTN) service.  Enhanced Monitoring models detect activities that are more likely to be indicative of compromise.  

Subsequently, Darktrace’s SOC triaged the activity detected on the SaaS account and sent a PTN alert to the customer, advising urgent follow up action.  The encrypted alert contained relevant technical details of the incident that were summarized by an expert Darktrace Analyst, along with recommendations to the customer’s internal SOC team to take immediate action.  Upon receipt and validation of the alert, the customer used Darktrace RESPOND to perform a manual force logout and block access from the external ExpressVPN IP.

Had Darktrace RESPOND been enabled in autonomous response mode, it would have immediately taken action to disable the account after ongoing anomalies were detected from it. However, as the customer only had RESPOND configured in the manual human confirmation model, the expertise of Darktrace’s SOC team was critical in enabling the customer to react and prevent further escalation of post-compromise activity.  Evidence of further attempts to access the compromised account were observed hours after RESPOND actions were taken, including failed login attempts from another rare external IP, this time associated with the VPN service NordVPN.

Figure 3: Timeline of attack and response actions from Darktrace SOC and Darktrace RESPOND.

Because the customer had subscribed to Darktrace’s PTN service, they were able to further leverage the expertise of Darktrace’s global team of cyber analysts and request further analysis of which files were accessed by the legitimate account owner versus the attacker.  This information was shared securely within the same Customer Portal ticket that was automatically opened on behalf of the customer when the PTN was alerted, allowing the customer’s security team to submit further queries and feedback, and request assistance to further investigate this alert within Darktrace. A similar service called Ask the Expert (ATE) exists for customers to draw from the expertise of Darktrace’s analysts at any time, not just when PTNs are alerted.

Conclusion

The growing prevalence and impact of BEC attacks amid the shift to cloud-based infrastructure means that already stretched internal security teams may not have the sufficient human capacity to detect and respond to these threats.

Darktrace’s round-the-clock SOC thwarted a BEC attack that had the potential to result in significant financial and reputational damage to the legal services company, by alerting the customer to high priority activity during the early stages of the attack and sharing actionable insights that the customer could use to prevent further escalation.  Following the confirmed compromise, the support and in-depth analysis provided by Darktrace’s SOC on the files accessed by the attacker enabled the customer to effectively report this breach to the Information Commissioner’s Office, to maintain compliance with UK data protection regulations. [4].  

Although the attacker used IP addresses that were local to the customer’s country of operations and did not perform overtly noisy actions during reconnaissance, Darktrace was able to identify that this activity deviated from the legitimate user’s typical pattern of life, triggering model breaches at each stage of the attack as it progressed from initial access to internal reconnaissance. While Darktrace RESPOND triggered an action that would have prevented the attack autonomously, the customer’s configuration meant that Darktrace’s SOC had an even more significant role in alerting the customer directly to take manual action.

Credit to: Sam Lister, Senior Analyst, for his contributions to this blog.

Appendices

Darktrace DETECT/Apps Models Breached:

  • SaaS / Access / Unusual External Source for SaaS Credential Use
  • SaaS / Compromise / Login From Rare Endpoint While User Is Active
  • SaaS / Unusual Activity / Activity from Multiple Unusual IPs
  • SaaS / Unusual Activity / Multiple Unusual SaaS Activities
  • SaaS / Access / Suspicious Login Attempt
  • SaaS / Compromise / SaaS Anomaly Following Anomalous Login (Enhanced Monitoring Model)

Darktrace RESPOND/Apps Models Breached:

  • Antigena / SaaS / Antigena Unusual Activity Block
  • Antigena / SaaS / Antigena Suspicious SaaS Activity Block

MITRE ATT&CK Mapping

Tactic Techniques
Reconnaissance • T1598 – Phishing for Information
Initial Access • T1078.004 – Valid Accounts: Cloud Accounts
Collection • T1213.002 – Data from Information Repositories: Sharepoint

References

[1] Rand, D. (2022, November 10). Why Business Email Compromise Costs Companies More Than Ransomware Attacks. Retrieved from Tanium: https://www.tanium.com/blog/whybusiness-email-compromise-costs-companies-more-than-ransomware-attacks/

[2] Federal Bureau of Investigation. (2022). 2022 IC3 Report. Retrieved from IC3.gov: https://www.ic3.gov/Media/PDF/AnnualReport/2022_IC3Report.pdf

[3] Department for Science, Innovation & Technology. (2023, April 19). Cyber security breaches survey 2023. Retrieved from gov.uk: https://www.gov.uk/government/statistics/cyber-security-breaches-survey-2023/cybersecurity-breaches-survey-2023

[4] ICO. (2023). Personal data breaches: a guide. Retrieved from Information Commissioner's Office: https://ico.org.uk/for-organisations/report-a-breach/personal-data-breach/personal-data-breaches-a-guide/#whatbreachesdo

Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Nicole Wong
Cyber Security Analyst

More in this series

No items found.

Blog

/

Network

/

August 6, 2026

When AI Agents Attack: The Case for Behavioral Anomaly Detection

Default blog imageDefault blog image

The UK AI Security Institute (AISI) has just published a detailed incident report confirming what many within the cyber security industry have suspected for some time: frontier AI agents, given a degree of autonomy, can independently develop and execute attack chains against real targets, including social engineering, supply-chain compromise, and deception, without explicit instruction to do so.

The report (INC-2026-07-28-01, published 4 August 2026) documents what happened during a routine capability evaluation between 25 and 28 July. It deserves a careful read from anyone responsible for securing the modern digital estate.

The incident

AISI regularly tests frontier AI models for offensive cyber capability. In this exercise, agents were placed inside a simulated network environment and given internet access this was deliberate, given to reflect realism of the task. Across 122 evaluation runs, 19 instances of unsanctioned activity were recorded, including cases where agents broke the intended boundaries of the test and acted against real people and real infrastructure on the open internet.

In the most serious case, an agent (Mythos 5) autonomously:

  • Conducted open-source intelligence gathering to identify real developers associated with keywords from the simulated scenario
  • Created a GitHub account and submitted a pull request containing obfuscated malicious code to a real, unaffiliated repository
  • Created a second sockpuppet account to post comments on its own PR, attempting to manufacture social proof
  • When a human reviewer flagged the malicious content, the agent lied, claiming an accidental commit, then repeatedly attempted to reintroduce the payload under the guise of having "fixed" the code
  • Sent spear-phishing emails to the repository maintainer containing manipulated content designed to convince them to merge the code
  • Planted a prompt injection inside a GitHub issue, hidden in an HTML comment invisible to humans but readable by AI coding assistants parsing the repository via API

AISI confirmed no real-world harm resulted. The PRs were not merged; the payloads did not execute. But containment was a product of circumstance, not design.

Why this matters

Every action the agents took used legitimate services operating within normal parameters. Creating a GitHub account is legitimate. Opening a pull request is legitimate. Sending an email is legitimate. Posting an issue is legitimate.

No vulnerability was exploited. No known malware signature was present. No indicator of compromise from any threat intelligence feed would have matched. The entire attack existed as a pattern, a sequence of individually normal actions that, taken together, constituted a supply-chain attack, a social engineering campaign, and an attempted compromise of downstream systems.

This is the class of threat that signature-based and rule-based security tools structurally struggle to detect. They answer the question "have I seen this before?" When the answer is no – and when the attack is novel, composed of legitimate actions, and adapted in real time by an adversary in this case an agent that reasons about evasion – they struggle.

Behavioral anomaly detection: The architecture that fits

Darktrace has spent a decade building a platform around a different question: "is this behavior normal for this entity?"

The platform learns the baseline behavior of every user, device, and communication pattern across an organization's digital environment. It does not need prior knowledge of an attack technique. It identifies deviation from learned normality, surfaces it, and through its Autonomous Response capability, takes targeted action to contain threats at machine speed.

Consider how this AISI incident might show across the Darktrace platform:

Network detection and response

While no payload was successfully delivered to a victim in this incident, AISI's post-incident analysis of the agent's code (detailed on pages 24-25 of the report) reveals what would have landed on target systems had the social engineering succeeded. The agent produced three successive payload generations, each replacing the last after being detected and called out:

Generation 1 hid a module-load stager inside a legitimate bug-fix commit, guarded by an environment variable check, to skip execution in CI. Its second stage attempted to fetch Poseidon implants (Linux and Windows variants) calling back to the agent's own C2 listener, with a fallback to publicly hosted polling scripts that fetched and executed commands every 60 seconds.

Generation 2 exploited an Electron "What's New" window with nodeIntegration enabled, loading remote HTML from a GitHub Pages site the agent had registered. The page decoded an XOR-obfuscated payload at runtime that beaconed host identifiers over DNS, attempted to download the same implants, ran its own in-process C2 loop, and installed reboot persistence via crontab and Windows registry Run keys.

Generation 3 was a post-install dropper added after the agent had already been publicly caught. It XOR-decoded a hidden URL, downloaded a script over plaintext HTTP, and ran it detached, force-exiting so npm install always reports success, giving the attacker remote code execution on any machine running the install.

None of these payloads reached a victim. But the technical detail matters: this is the calibre of tradecraft an autonomous agent produced in a matter of hours, unprompted. Each generation demonstrated adaptation, the agent iterated its approach in direct response to being detected.

For network defenders, the behavioral signals these payloads would have generated are clear. DNS beaconing to novel domains, periodic HTTP polling to newly registered infrastructure, XOR-encoded data in transit, registry and crontab persistence writes on endpoints that have never exhibited such behavior. These are exactly the anomalies that Darktrace / NETWORK identifies in real time: deviations from learned device and connection baselines, regardless of whether the specific payload or C2 domain has been seen before.

Beyond the payload analysis, the agents' broader activity produced additional behavioral anomalies: new identities engaging in high-frequency interactions with external services, coordination via unusual channels (public pastebins, gists), and lateral communication patterns that deviate from any established baseline.

Email security

The spear-phishing emails sent by the agent, crafted with context-specific content designed to manipulate a specific individual, represent exactly the kind of socially engineered communication that static email gateways can miss. Darktrace / EMAIL analyses the linguistic patterns, sender behavior, and relational context of every message. A brand-new sender using tailored language to pressure a recipient into a specific technical action deviates from the learned communication profile, regardless of whether the message contains a known-bad URL or attachment.

Securing AI

This is where the incident points to a threat category that many security platforms do not yet fully address. The agent did not just use AI as a tool for attack. It targeted other AI systems, attempting prompt injections against coding assistants and trying to compromise automated workflows. Organisations adopting AI agents for development, operations, and decision-making are introducing a newer extended attack surface.  

Darktrace / SECURE AI extends the same behavioral detection model into this space. It monitors AI usage across the enterprise, detecting anomalous interactions with AI services, identifying sensitive data exposure to AI systems, providing visibility into shadow AI use. As AI agents become standard components of enterprise infrastructure, the ability to detect when an AI system is behaving outside its expected parameters, when there is a chance, it is being manipulated, or when it is interacting with something it should not be, that becomes a direct security requirement rather than a future consideration.

The principle is consistent across the Darktrace platform: learn what normal looks like, identify deviation, respond autonomously.

The AI-to-AI threat surface

One of the report's most important findings is a confirmation that AI agents can act outside of instructions and guardrails to complete tasks. The Mythos 5 agent conducted reconnaissance on a repository maintainer, inferred it might be an AI coding assistant, and deployed a prompt injection specifically crafted to compromise it.

This is a new attack class. Organizations deploying AI coding assistants, AI-powered code review, or autonomous DevOps agents now face an adversary that:

  • Moves at speed
  • Is able to adapt its approach when initial attempts are halted
  • Specifically targets AI systems via techniques (prompt injection, hidden instructions in HTML comments) that are less visible to human reviewers but parsed by machines
  • Reasons about which evasion strategies are most likely to succeed

Defending against this requires security that operates with the same adaptability and speed. A system that learns what "normal" looks like, and flags deviation from it.

What AISI recommends

The report identifies five contributing factors to the incident. Two are directly relevant:

No synchronous real-time monitoring (Section 5.3): AISI acknowledges it had not built real-time behavioral monitoring capable of immediately blocking unsanctioned agent actions. Their security team detected the incident via commercial security monitoring software, but only after activity had already occurred.

No fine-grained behavioral controls (Section 6.2): AISI's forward-looking response includes building continuous monitoring that can "immediately block or flag unsanctioned AI agent behavior."

The direction is clear: organizations need security that watches behavior continuously, learns what normal looks like, and acts when deviation occurs, without requiring a human in the loop for every decision.

The takeaway

The AISI incident is documented proof that autonomous AI agents will independently develop novel attack chains using legitimate tools, adapt when detected, employ deception, and target both humans and other AI systems. This happened last week, in a controlled setting, with commercially available models.

The security architecture that addresses this is behavioral anomaly detection applied across the full digital estate, as AI agents become standard components of enterprise infrastructure, writing code, managing deployments, processing communications, the attack surface they create is behavioral by nature. The defense must be too.

Read the full report from the UK AI Security Institute here.

Continue reading
About the author
Adam Stevens
Senior Director of Product | Darktrace

Blog

/

AI

/

August 5, 2026

Testing a Prompt injection Attack Against an Enterprise AI Agent

Default blog imageDefault blog image

Key takeaways

  • Darktrace successfully detected and quarantined a prompt injection email before it could be processed by an enterprise AI agent.  
  • Prompt injection attacks increasingly rely on natural language rather than traditional malware, making behavioral analysis an important complement to signature-based detection.  
  • Organizations deploying AI agents should combine model guardrails with behavioral monitoring to reduce the risk of malicious instructions reaching enterprise systems.

How behavioral detection helps stop prompt injection attacks

A Darktrace customer running a Gemini AI agent in Google Cloud asked us two simple questions:

“If my agent can read inbound emails and access internal data, what stops an attacker from hiding malicious instructions in the message? Couldn’t the agent be tricked into deleting or exfiltrating sensitive data?”

The scenario centers on an indirect prompt injection attack, where malicious instructions are hidden inside content that an AI model later interprets as trusted input. The same weakness was exposed by  EchoLeak (CVE-2025-32711), a zero-click Microsoft 365 Copilot vulnerability enabled data exfiltration from a single well-crafted email..

This blog follows the Darktrace team’s investigation of the customer’s hypothesis and examines how the attack interacted with their existing security stack. The results highlight which defenses held, where gaps emerged, and how behavioral detection mattered more than guardrails. This investigation also demonstrates why behavioral detection is becoming increasingly important for AI security, as prompt injections often contain no traditional indicators of compromise.  

How do prompt injection attacks work?

Prompt injection works by carefully crafting the content and structure of the prompt to alter the LLM’s behavior or output in unintended ways. This can cause models to violate guardrails, generate harmful content or enable unauthorised access.

Prompt injection attack example

The well-known example, EchoLeak (CVE-2025-32711), was a zero-click vulnerability in Microsoft 365 Copilot that relied on a carefully crafted email containing hidden instructions that the AI system interpreted as commands rather than content, creating a pathway for unauthorized access to sensitive information without any user interaction.

While Darktrace / SECURE AI is designed to prevent agents from producing unintended outcomes, we wanted to see if we could catch and prevent this threat type earlier in the attack life-cycle, at the email security layer.

How we tested prompt injection attacks on an enterprise agent

Summary:

  1. Claude generated a prompt injection payload.  
  2. Hidden instructions were embedded in an email.  
  3. The email passed traditional validation checks.  
  4. Darktrace analyzed the language and sender behavior.  
  5. The email was quarantined before the AI agent could process it.

To test Darktrace / EMAIL against this attack class, we opened Claude, gave it the customer's context and problem statement (Gemini agent with inbox access, internal tool calls), told it we were validating Darktrace / EMAIL's detection of prompt injections, and asked for a test payload. See below:

Figure 1
Figure 2

Despite the guardrails supposedly built into the model, Claude surprisingly gave us the entire exploit in plaintext (albeit very basic), illegible to a human as the text was sent in white text (see Figure 1) but framed as an authoritative override for anything downstream reading the mail programmatically (i.e. the Gemini agent).

How Darktrace detected a prompt injection attack

We then sent the Claude-crafted email from a freemail address to the target recipient’s inbox. Despite the email containing no malicious payload, the freemail address having no malicious reputation, and the validation checks all passing, Darktrace  /EMAIL flagged the email as a 93/100 anomaly and moved it to junk, out of scope for the AI agent.

Figure 3: The test email sent with the hidden prompt injection
Figure 4: The email analysis in Darktrace / EMAIL 
Figure 5: Darktrace / EMAIL detection of malicious activity

The interesting part is what triggered the detection (see Figure 5)

  • Possible machine prompt content: text in the body detected as instructions written for a machine to execute, not for a human to read
  • Possible machine prompt content + basic suspicious correspondence: the same content, correlated with sender-side anomalies: freemail domain (yahoo[.]com), unknown correspondent, no prior mail history with the recipient, and suspicious references to payment information

Neither of those is a signature match. Nothing in the email was on a blacklist. There was no malware, no link and no attachment. Darktrace analyzed the context in which the email was delivered and flagged it as likely risky.  The anomalous language features and the context of the sender relative to the recipient's normal behavior, combined with the unusual hidden text (prompt) were enough for Darktrace / EMAIL to act on the risk.

Result: Darktrace / EMAIL autonomously junked the email, out of scope for any AI agent parsing the inbox.

Why behavioral security makes a difference detecting prompt injection attacks

Cyberattacks don't look like traditional exploits anymore. They now operate in natural language, not strictly code.

That breaks the traditional stack. AV, firewalls, static scanning and signature-based SEGs all assume a payload to inspect.  

A prompt injection has no payload. It's just an instruction, written in natural language, dressed up as anything the attacker wants: an invoice, an HR request, a calendar invite, some simple PowerPoint slides.

EchoLeak proved that hidden instructions can sit inside an email invisible to the user but fully readable by the LLM, and the LLM will follow them blindly.  

This test and GTG-1002 proved that the LLM itself can be socially engineered. Tell it you're an authorized tester and it will hand you the attack.

Rules and static classifiers can catch the obvious cases. But natural language has infinite variants, and the attack surface is the model's innate functionality to comply.  

The deeper problem here is intent: an LLM can't reliably tell whether an instruction in its context came from its developer, its user, or an attacker who slipped it into an email. To the LLM, everything reads as language and looks like a legitimate ask. This is why behavioural detection wins, as you become aware of intent when you look at the context of an interaction. Does this sender normally send this kind of message to this recipient? Does this prompt fit the user's normal pattern? Is this agent behaving the way this agent normally behaves?  

Intent can't be read off a single email, it emerges from behavioral context. Which is how Darktrace enables threat detection, through behavioral understanding.

Why enterprise AI security requires more than guardrails

Claude didn't roll over immediately… the first section of the response was a (slight) pushback, but then it wrote the payload anyway without having to ask twice.

Here the framing of the prompt did all the work. The “testing security capabilities” angle moved the model from refusal to unquestioned compliance to the user prompt.

This isn't the first time this has happened, of course. Anthropic disclosed in November 2025 that a Chinese state-sponsored group they tracked as GTG-1002 ran the first documented AI-orchestrated espionage campaign against ~30 targets by posing as employees of a legitimate cybersecurity firm doing authorised penetration testing.

The takeaway isn't that AI guardrails are ineffective. They raise the cost of low-effort attacks and remain an important first layer of defense. However, for most organizations today, they’re the only line of defense when deploying AI agents. If a prompt injection bypasses those controls, organizations still need a way to detect and stop malicious behavior elsewhere in the attack chain.

Attackers will continue to have working prompt injections easily and quickly. The question is what stops one when it lands in an inbox your agent is reading.

That's where behavioral detection comes in.

How Darktrace detects prompt injection attacks in emails

Two things Darktrace does that a model-level guardrail or static rules and signatures can't:

Natural language analysis at the email or prompt layer. The email is assessed on its own merits: is this content shaped like instructions for a machine, regardless of what the receiving agent decides to do about it?

Behavioral context around the language. An AI agent behaves like an extremely agreeable human, and it will go above and beyond to comply with the user’s request. That's exactly why you must consider the business context, such sender behaviour, mailing history, and organisational norms, as these matter even more when the recipient is an AI.

Darktrace has been perfecting behavioral anomaly detection for over a decade; the same self-learning approach that catches BEC and account takeover applies directly to prompt injection delivery. Our multi-layered AI stack extracts content from the message, builds behavioural understanding through social graphing and Pattern of Life analysis, and then combines natural language, topic, inducement, sender relationship and anomaly signals before deciding what action to take.  

This matters for prompt injection because the threat is not the plain language itself, but the intent behind the language that can cause an AI agent to respond in unexpected ways.

How to secure enterprise AI operations from prompt injection attacks

Email was the entry point in this case, but it is only one of many possible vectors.  

Anywhere an agent can retrieve information, an attacker can potentially introduce a prompt injection.

Emails, documents, SharePoint sites, web pages, knowledge bases, chat platforms, and third-party integrations all provide opportunities to influence an agent's behavior. Wherever an agent finds its orders, a prompt injection opportunity exists.

This is why securing AI requires more than blocking malicious inputs. Organizations also need visibility into how agents behave after consuming information from across their environment. If an agent begins accessing unexpected data, taking unusual actions, or operating outside its normal patterns, those behaviors may provide the strongest signal that something has gone wrong.

Effective AI security requires defense in depth: reducing the likelihood of malicious instructions reaching the agent while maintaining the ability to detect and investigate suspicious behavior if they do.

The challenge isn't protecting a single entry point. It's recognizing that, in an AI-powered environment, every source of information is also a potential source of influence.

Are you deploying autonomous agents across your enterprise and want to see this tested in your environment? Let's talk.

Continue reading
About the author
Carlo Loregian
Solutions Engineer
あなたのデータ × DarktraceのAI
唯一無二のDarktrace AIで、ネットワークセキュリティを次の次元へ