Blog
/
Network
/
October 14, 2024

How Triada Affects Banking and Communication Apps

Explore the intricacies of the Triada Trojan and its targeting of communication and banking apps. Learn how to safeguard against this threat.
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
Justin Torres
Cyber Analyst
Default blog image
14
Oct 2024

The rise of android malware

Recently, there has been a significant increase in malware strains targeting mobile devices, with a growing number of Android-based malware families, such as banking trojans, which aim to steal sensitive banking information from organizations and individuals worldwide.

These malware families attempt to access users’ accounts to steal online banking credentials and cookies, bypass multi-factor authentication (MFA), and conduct automatic transactions to steal funds [1]. They often masquerade as legitimate software or communications from social media platforms to compromise devices. Once installed, they use tactics such as keylogging, dumping cached credentials, and searching the file system for stored passwords to steal credentials, take over accounts, and potentially perform identity theft [1].

One recent example is the Antidot Trojan, which infects devices by disguising itself as an update page for Google Play. It establishes a command-and-control (C2) channel with a server, allowing malicious actors to execute commands and collect sensitive data [2].

Despite these malware’s ability to evade detection by standard security software, for example, by changing their code [3], Darktrace recently detected another Android malware family, Triada, communicating with a C2 server and exfiltrating data.

Triada: Background and tactics

First surfacing in 2016, Triada is a modular mobile trojan known to target banking and financial applications, as well as popular communication applications like WhatsApp, Facebook, and Google Mail [4]. It has been deployed as a backdoor on devices such as CTV boxes, smartphones, and tablets during the supply chain process [5]. Triada can also be delivered via drive-by downloads, phishing campaigns, smaller trojans like Leech, Ztorg, and Gopro, or more recently, as a malicious module in applications such as unofficial versions of WhatsApp, YoWhatsApp, and FM WhatsApp [6] [7].

How does Triada work?

Once downloaded onto a user’s device, Triada collects information about the system, such as the device’s model, OS version, SD card space, and list of installed applications, and sends this information to a C2 server. The server then responds with a configuration file containing the device’s personal identification number and settings, including the list of modules to be installed.

After a device has been successfully infected by Triada, malicious actors can monitor and intercept incoming and outgoing texts (including two-factor authentication messages), steal login credentials and credit card information from financial applications, divert in-application purchases to themselves, create fake messaging and email accounts, install additional malicious applications, infect devices with ransomware, and take control of the camera and microphone [4] [7].

For devices infected by unofficial versions of WhatsApp, which are downloaded from third-party app stores [9] and from mobile applications such as Snaptube and Vidmate , Triada collects unique device identifiers, information, and keys required for legitimate WhatsApp to work and sends them to a remote server to register the device [7] [12]. The server then responds by sending a link to the Triada payload, which is downloaded and launched. This payload will also download additional malicious modules, sign into WhatsApp accounts on the target’s phone, and request the same permissions as the legitimate WhatsApp application, such as access to SMS messages. If granted, a malicious actor can sign the user up for paid subscriptions without their knowledge. Triada then collects information about the user’s device and mobile operator and sends it to the C2 server [9] [12].

How does Triada avoid detection?

Triada evades detection by modifying the Zygote process, which serves as a template for every application in the Android OS. This enables the malware to become part of every application launched on a device [3]. It also substitutes system functions and conceals modules from the list of running processes and installed apps, ensuring that the system does not raise the alarm [3]. Additionally, as Triada connects to a C2 server on the first boot, infected devices remain compromised even after a factory reset [4].

Triada attack overview

Across multiple customer deployments, devices were observed making a large number of connections to a range of hostnames, primarily over encrypted SSL and HTTPS protocols. These hostnames had never previously been observed on the customers’ networks and appear to be algorithmically generated. Examples include “68u91.66foh90o[.]com”, “92n7au[.]uhabq9[.]com”, “9yrh7.mea5ms[.]com”, and “is5jg.3zweuj[.]com”.

External Sites Summary Graph showing the rarity of the hostname “92n7au[.]uhabq9[.]com” on a customer network.
Figure 1: External Sites Summary Graph showing the rarity of the hostname “92n7au[.]uhabq9[.]com” on a customer network.

Most of the IP addresses associated with these hostnames belong to an ASN associated with the cloud provider Alibaba (i.e., AS45102 Alibaba US Technology Co., Ltd). These connections were made over a range of high number ports over 1000, most commonly over 30000 such as 32091, which Darktrace recognized as extremely unusual for the SSL and HTTPS protocols.

Screenshot of a Model Alert Event log showing a device connecting to the endpoint “is5jg[.]3zweuj[.]com” over port 32091.
Figure 2: Screenshot of a Model Alert Event log showing a device connecting to the endpoint “is5jg[.]3zweuj[.]com” over port 32091.

On several customer deployments, devices were seen exfiltrating data to hostnames which also appeared to be algorithmically generated. This occurred via HTTP POST requests containing unusual URI strings that were made without a prior GET request, indicating that the infected device was using a hardcoded list of C2 servers.

Screenshot of a Model Alert Event Log showing the device posting the string “i8xps1” to the hostname “72zf6.rxqfd[.]com.
Figure 3: Screenshot of a Model Alert Event Log showing the device posting the string “i8xps1” to the hostname “72zf6.rxqfd[.]com.
 Screenshot of a Model Alert Event Log showing the device posting the string “sqyjyadwwq” to the hostname “9yrh7.mea5ms[.]com”.
Figure 4: Screenshot of a Model Alert Event Log showing the device posting the string “sqyjyadwwq” to the hostname “9yrh7.mea5ms[.]com”.

These connections correspond with reports that devices affected by Triada communicate with the C2 server to transmit their information and receive instructions for installing the payload.

A number of these endpoints have communicating files associated with the unofficial WhatsApp versions YoWhatsApp and FM WhatsApp [11] [12] [13] . This could indicate that the devices connecting to these endpoints were infected via malicious modules in the unofficial versions of WhatsApp, as reported by open-source intelligence (OSINT) [10] [12]. It could also mean that the infected devices are using these connections to download additional files from the C2 server, which could infect systems with additional malicious modules related to Triada.

Moreover, on certain customer deployments, shortly before or after connecting to algorithmically generated hostnames with communicating files linked to YoWhatsApp and FM WhatsApp, devices were also seen connecting to multiple endpoints associated with WhatsApp and Facebook.

Figure 5: Screenshot from a device’s event log showing connections to endpoints associated with WhatsApp shortly after it connected to “9yrh7.mea5ms[.]com”.

These surrounding connections indicate that Triada is attempting to sign in to the users’ WhatsApp accounts on their mobile devices to request permissions such as access to text messages. Additionally, Triada sends information about users’ devices and mobile operators to the C2 server.

The connections made to the algorithmically generated hostnames over SSL and HTTPS protocols, along with the HTTP POST requests, triggered multiple Darktrace models to alert. These models include those that detect connections to potentially algorithmically generated hostnames, connections over ports that are highly unusual for the protocol used, unusual connectivity over the SSL protocol, and HTTP POSTs to endpoints that Darktrace has determined to be rare for the network.

Conclusion

Recently, the use of Android-based malware families, aimed at stealing banking and login credentials, has become a popular trend among threat actors. They use this information to perform identity theft and steal funds from victims worldwide.

Across affected customers, multiple devices were observed connecting to a range of likely algorithmically generated hostnames over SSL and HTTPS protocols. These devices were also seen sending data out of the network to various hostnames via HTTP POST requests without first making a GET request. The URIs in these requests appeared to be algorithmically generated, suggesting the exfiltration of sensitive network data to multiple Triada C2 servers.

This activity highlights the sophisticated methods used by malware like Triada to evade detection and exfiltrate data. It underscores the importance of advanced security measures and anomaly-based detection systems to identify and mitigate such mobile threats, protecting sensitive information and maintaining network integrity.

Credit to: Justin Torres (Senior Cyber Security Analyst) and Anna Gilbertson (Cyber Security Analyst).

Appendices

Darktrace Model Detections

Model Alert Coverage

Anomalous Connection / Application Protocol on Uncommon Port

Anomalous Connection / Multiple Connections to New External TCP Port

Anomalous Connection / Multiple HTTP POSTS to Rare Hostname

Anomalous Connections / Multiple Failed Connections to Rare Endpoint

Anomalous Connection / Suspicious Expired SSL

Compromise / DGA Beacon

Compromise / Domain Fluxing

Compromise / Fast Beaconing to DGA

Compromise / Sustained SSL or HTTP Increase

Compromise / Unusual Connections to Rare Lets Encrypt

Unusual Activity / Unusual External Activity

AI Analyst Incident Coverage

Unusual Repeated Connections to Multiple Endpoints

Possible SSL Command and Control

Unusual Repeated Connections

List of Indicators of Compromise (IoCs)

Ioc – Type - Description

  • is5jg[.]3zweuj[.]com - Hostname - Triada C2 Endpoint
  • 68u91[.]66foh90o[.]com - Hostname - Triada C2 Endpoint
  • 9yrh7[.]mea5ms[.]com - Hostname - Triada C2 Endpoint
  • 92n7au[.]uhabq9[.]com - Hostname - Triada C2 Endpoint
  • 4a5x2[.]fs4ah[.]com - Hostname - Triada C2 Endpoint
  • jmll4[.]66foh90o[.]com - Hostname - Triada C2 Endpoint
  • mrswd[.]wo87sf[.]com - Hostname - Triada C2 Endpoint
  • lptkw[.]s4xx6[.]com - Hostname - Triada C2 Endpoint
  • ya27fw[.]k6zix6[.]com - Hostname - Triada C2 Endpoint
  • w0g25[.]66foh90o[.]com - Hostname - Triada C2 Endpoint
  • kivr8[.]wd6vy[.]com - Hostname - Triada C2 Endpoint
  • iuwe64[.]ct8pc6[.]com - Hostname - Triada C2 Endpoint
  • qefgn[.]8z0le[.]com - Hostname - Triada C2 Endpoint
  • a6y0x[.]xu0h7[.]com - Hostname - Triada C2 Endpoint
  • wewjyw[.]qb6ges[.]com - Hostname - Triada C2 Endpoint
  • vx9dle[.]n0qq3z[.]com - Hostname - Triada C2 Endpoint
  • 72zf6[.]rxqfd[.]com - Hostname - Triada C2 Endpoint
  • dwq[.]fsdw4f[.]com - Hostname - Triada C2 Endpoint
  • tqq6g[.]66foh90o[.]com - Hostname - Triada C2 Endpoint
  • 1rma1[.]4f8uq[.]com - Hostname - Triada C2 Endpoint
  • 0fdwa[.]7j3gj[.]com - Hostname - Triada C2 Endpoint
  • 5a7en[.]1e42t[.]com - Hostname - Triada C2 Endpoint
  • gmcp4[.]1e42t[.]com - Hostname - Triada C2 Endpoint
  • g7190[.]rt14v[.]com - Hostname - Triada C2 Endpoint
  • goyvi[.]2l2wa[.]com - Hostname - Triada C2 Endpoint
  • zq6kk[.]ca0qf[.]com - Hostname - Triada C2 Endpoint
  • sv83k[.]bn3avv[.]com - Hostname - Triada C2 Endpoint
  • 9sae7h[.]ct8pc6[.]com - Hostname - Triada C2 Endpoint
  • jpygmk[.]qt7tqr[.]com - Hostname - Triada C2 Endpoint
  • av2wg[.]rt14v[.]com - Hostname - Triada C2 Endpoint
  • ugbrg[.]osz1p[.]com - Hostname - Triada C2 Endpoint
  • hw2dm[.]wtws9k[.]com - Hostname - Triada C2 Endpoint
  • kj9atb[.]hai8j1[.]com - Hostname - Triada C2 Endpoint
  • pls9b[.]b0vb3[.]com - Hostname - Triada C2 Endpoint
  • 8rweau[.]j7e7r[.]com - Hostname - Triada C2 Endpoint
  • wkc5kn[.]j7e7r[.]com - Hostname - Triada C2 Endpoint
  • v58pq[.]mpvflv[.]com - Hostname - Triada C2 Endpoint
  • zmai4k[.]huqp3e[.]com - Hostname - Triada C2 Endpoint
  • eajgum[.]huqp3e[.]com - Hostname - Triada C2 Endpoint
  • mxl9zg[.]kv0pzv[.]com - Hostname - Triada C2 Endpoint
  • ad1x7[.]mea5ms[.]com - Hostname - Triada C2 Endpoint
  • ixhtb[.]s9gxw8[.]com - Hostname - Triada C2 Endpoint
  • vg1ne[.]uhabq9[.]com - Hostname - Triada C2 Endpoint
  • q5gd0[.]birxpk[.]com - Hostname - Triada C2 Endpoint
  • dycsw[.]h99n6[.]com - Hostname - Triada C2 Endpoint
  • a3miu[.]h99n6[.]com - Hostname - Triada C2 Endpoint
  • qru62[.]5qwu8b5[.]com - Hostname - Triada C2 Endpoint
  • 3eox8[.]abxkoop[.]com - Hostname - Triada C2 Endpoint
  • 0kttj[.]bddld[.]com - Hostname - Triada C2 Endpoint
  • gjhdr[.]xikuj[.]com - Hostname - Triada C2 Endpoint
  • zq6kk[.]wm0hd[.]com - Hostname - Triada C2 Endpoint
  • 8.222.219[.]234 - IP Address - Triada C2 Endpoint
  • 8.222.244[.]205 - IP Address - Triada C2 Endpoint
  • 8.222.243[.]182 - IP Address - Triada C2 Endpoint
  • 8.222.240[.]127 - IP Address - Triada C2 Endpoint
  • 8.219.123[.]139 - IP Address - Triada C2 Endpoint
  • 8.219.196[.]124 - IP Address - Triada C2 Endpoint
  • 8.222.217[.]73 - IP Address - Triada C2 Endpoint
  • 8.222.251[.]253 - IP Address - Triada C2 Endpoint
  • 8.222.194[.]254 - IP Address - Triada C2 Endpoint
  • 8.222.251[.]34 - IP Address - Triada C2 Endpoint
  • 8.222.216[.]105 - IP Address - Triada C2 Endpoint
  • 47.245.83[.]167 - IP Address - Triada C2 Endpoint
  • 198.200.54[.]56 - IP Address - Triada C2 Endpoint
  • 47.236.113[.]126 - IP Address - Triada C2 Endpoint
  • 47.241.47[.]128 - IP Address - Triada C2 Endpoint
  • /iyuljwdhxk - URI - Triada C2 URI
  • /gvuhlbzknh - URI - Triada C2 URI
  • /sqyjyadwwq - URI - Triada C2 URI
  • /cncyz3 - URI - Triada C2 URI
  • /42k0zk - URI - Triada C2 URI
  • /75kdl5 - URI - Triada C2 URI
  • /i8xps1 - URI - Triada C2 URI
  • /84gcjmo - URI - Triada C2 URI
  • /fkhiwf - URI - Triada C2 URI

MITRE ATT&CK Mapping

Technique Name - Tactic - ID - Sub-Technique of

Data Obfuscation - COMMAND AND CONTROL - T1001

Non-Standard Port - COMMAND AND CONTROL - T1571

Standard Application Layer Protocol - COMMAND AND CONTROL ICS - T0869

Non-Application Layer Protocol - COMMAND AND CONTROL - T1095

Masquerading - EVASION ICS - T0849

Man in the Browser - COLLECTION - T1185

Web Protocols - COMMAND AND CONTROL - T1071.001 -T1071

External Proxy - COMMAND AND CONTROL - T1090.002 - T1090

Domain Generation Algorithms - COMMAND AND CONTROL - T1568.002 - T1568

Web Services - RESOURCE DEVELOPMENT - T1583.006 - T1583

DNS - COMMAND AND CONTROL - T1071.004 - T1071

Fast Flux DNS - COMMAND AND CONTROL - T1568.001 - T1568

One-Way Communication - COMMAND AND CONTROL - T1102.003 - T1102

Digital Certificates - RESOURCE DEVELOPMENT - T1587.003 - T1587

References

[1] https://www.checkpoint.com/cyber-hub/cyber-security/what-is-trojan/what-is-a-banking-trojan/

[2] https://cyberfraudcentre.com/the-rise-of-the-antidot-android-banking-trojan-a-comprehensive-guide

[3] https://www.zimperium.com/glossary/banking-trojans/

[4] https://www.geeksforgeeks.org/what-is-triada-malware/

[5] https://www.infosecurity-magazine.com/news/malware-infected-devices-retailers/

[6] https://www.pcrisk.com/removal-guides/24926-triada-trojan-android

[7] https://securelist.com/malicious-whatsapp-mod-distributed-through-legitimate-apps/107690/

[8] https://securityboulevard.com/2024/02/impact-of-badbox-and-peachpit-malware-on-android-devices/

[9] https://threatpost.com/custom-whatsapp-build-malware/168892/

[10] https://securelist.com/triada-trojan-in-whatsapp-mod/103679/

[11] https://www.virustotal.com/gui/domain/is5jg.3zweuj.com/relations

[12] https://www.virustotal.com/gui/domain/92n7au.uhabq9.com/relations

[13] https://www.virustotal.com/gui/domain/68u91.66foh90o.com/relations

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
Justin Torres
Cyber Analyst

More in this series

No items found.

Blog

/

AI

/

May 28, 2026

From Efficiency to Exposure: How AI Adoption Is Creating Unseen Vulnerabilities on the Factory Floor

AI in manufacturingDefault blog imageDefault blog image

How AI agents impact the manufacturing industry

Security teams and IT personnel across the manufacturing industry are under constant pressure to protect production, maintain uptime, and safeguard critical assets but the rise of AI is bringing huge new opportunities alongside new cyber risks. Across manufacturing, AI is embedded into workflows, decision-making, and increasingly, autonomous AI agents are acting on behalf of employees and systems.  

Agentic systems are powerful because they can act independently, but that same autonomy also creates cyber and operational risk. Agents have extensive permissions and are capable of carrying out complex tasks, making decisions, and interacting with tools or external systems with little to no human intervention.

Unlike traditional AI models that perform predefined tasks, AI agents use advanced techniques to mimic human decision-making processes, dynamically adapting to new challenges, making decision and taking action based on their own judgement. They look like employees operationally but lack judgment, ethics, or fear of consequences like humans do. This means they can be easily manipulated by cybercriminals, and an AI agent embedded across an OT network creates threats that extend well beyond data exposure. For example, at BMW, AI identifies faults in welding processes as they occur. At its Spartanburg plant, AI monitors the weld of 300-400 metal studs onto every SUV frame to detect misplaced or faulty studs and correct them instantly. Corruption of BMW’s AI system could lead to catastrophic quality control errors.

Adopting agentic AI systems across manufacturing raises some concerns across security teams. New data from our State of AI Cybersecurity survey shows that 78% of manufacturing security professionals are worried about employee use of AI agents – their top concern. That’s followed by employee use of generative AI tools like CoPilot and ChatGPT, a worry for 76% of security professionals at manufacturing organizations. As these tools gain more access to business data and processes, and more autonomy within organizations, security teams, who today have minimal visibility of agent activity in their environments, increasingly have sensitive data exposure (a worry for 60%) and accidental policy and regulatory violations (59%) on their minds.

External AI-powered threats are evolving just as quickly

The same capabilities transforming manufacturing are also reshaping cyberattacks.

AI is enabling attackers to automate reconnaissance, refine targeting, and adapt in real time. What once required time and manual effort can now be executed continuously and at scale. Manufacturers are already seeing the impact. According to manufacturing security professionals we surveyed, 76% are already being impacted by AI-powered threats and 90% see AI increasing the success of social engineering attacks.

And the techniques themselves are evolving. Concerns across the manufacturing sector show growing anxiety about the range of AI-powered attack routes, most pressingly of adaptive malware that evolves in real-time – a prospect half (49%) of manufacturing security professionals we surveyed are worried by, a full 9% more than the average across industries. AI adaptive malware is followed by:

  • Automated vulnerability scanning and exploit chaining (48%) which has become even more pressing as Anthropic’s new Mythos AI Model supercharges vulnerability discovery
  • Hyper-personalized phishing campaigns (46%), which remain a mainstay in hackers’ arsenals, and AI has amplified their effectiveness by making phishing emails more convincing and harder to detect.

This is not just an increase in volume, it is a shift toward threats that evolve as they unfold - often faster than static defenses can respond.

Despite rising awareness, many manufacturers are not yet equipped to manage this shift. More than half (51%) say they are not adequately prepared for AI-driven threats, and only 37% have formal policies governing AI deployment.  

Securing AI through visibility, context, and guardrails

Addressing this challenge does not require manufacturers to slow innovation. It requires a different approach to security, one that can operate at the same speed and scale as AI. Three specific priorities are emerging for manufacturers looking to take advantage of the power of AI.

Visibility is foundational.  

Organizations need to understand where AI is being used, what it can access, and how it behaves across both IT and OT environments. Without that, risk cannot be measured or managed. It is no surprise that Darktrace’s research found that 91% of manufacturing security professionals said that they need to understand how AI makes decisions before trusting it. This is even more critical in operational settings where disruption has safety, environmental, financial, and reputational impacts.

Context is what turns visibility into action.  

In environments shaped by AI, normal behavior is constantly shifting. Detecting threats requires a behavioral approach; understanding patterns of life across the organization and identifying subtle deviations in real time – a step change in organizations’ traditional approach to security and risk management.

Guardrails ensure that agency does not become exposure  

As AI systems take on greater responsibility, organizations need clear boundaries around what they can do and when they can act independently. These controls must be embedded into systems themselves, not applied after the fact.  

Securing AI Agents Across Manufacturing IT and OT

The rise of agentic AI is transforming manufacturing - powering next-generation operations while reshaping the security landscape. This is not just an increase in threats, but a shift to autonomous systems, continuously evolving behaviors, and risks moving at machine speed. For organizations trying to grapple with the challenge of enabling AI while managing the risk, visibility, context and guardrails should be foundational.

Darktrace helps manufacturers build secure AI approaches by making those foundations possible. It provides visibility and real-time detection and response to unusual activity across IT and OT environments and allows organizations to understand AI activity from the prompts employees use and the agents they build to how those agents are behaving across the environment. For manufacturers scaling AI, this delivers a foundation for innovation without sacrificing control.

Continue reading
About the author
Oakley Cox
Director of Product

Blog

/

AI

/

May 28, 2026

How to Evaluate AI Vendors: 5 Key categories for AI Adoption

Default blog imageDefault blog image

Understanding the AI buyers’ market

AI adoption has become a central topic of discussion in boardrooms, drawing growing interest from business leaders. Ultimately, organizations hope that an investment in AI technology will have tremendous returns. However, the process of buying an AI solution is not as straight forward as it appears on the surface.  

While business leaders may be eager to improve productivity across their operations, practitioners responsible for evaluating and selecting AI solutions may not always have the visibility or technical understanding needed to make the right decisions for their business. What is typically marketed as a holistic solution to their most critical problems is usually followed by uncertainty when AI tools are finally operationalized in real environments.

This guide is intended to support security leaders who are under growing pressure to adopt AI tools while navigating complex terminology, vendor claims, and increasingly crowded buying cycles. Ultimately, the goal is to help organizations evaluate and adopt AI in a safe, effective, and well-governed way. To support this, we’ve structured the evaluation framework across five key categories:

  1. Governance, safety, and data controls
  1. Data gathering and training
  1. Model and technique choice
  1. Performance and accuracy validation    
  1. Interpretability, adjustability, and transparency    

What buying AI looks like in cybersecurity

While investing in AI can bring immense benefits to your security team, first-time buyers of AI cybersecurity solutions may not know where to start. They will have to determine the type of tool they want, know the options available, and evaluate vendors. Research and understanding are critical to ensure purchases are worth the investment.  

With acceleration in AI adoption, accompanied by the recent boom in agentic AI and autonomous agents, CISOs must look “beneath the hood" of these tools to understand how they work, how they are governed, and to ensure the system is secure and compliant with internal policies.

Challenges in the AI buyers’ marketplace  

The AI security software market is buzzing with hype and flashy promises, which, understandably, needs to be addressed with due diligence. Potential buyers, especially in the cybersecurity space, are hesitant when it comes to allowing AI autonomous capabilities across their workflows, and a lack of vendor transparency can exacerbate those feelings.  

Reinforcing this sentiment, research from this year's Darktrace’s State of AI Cybersecurity report shows where confidence and hesitancy emerge amongst potential buyers. On the one hand, security professionals agree that they have good visibility into the logic and reasoning processes their AI solutions use. However, they lack the explainability and trust to allow AI to take independent remedial action.

  • 89% say they have good visibility into the reasoning behind the outputs generated by AI solutions
  • 92% say they need to understand how a defensive AI tool makes decisions before they can trust it
  • Only 14% say they allow AI to act independently, performing autonomous actions without human approval
  • 74% say they are limiting the autonomy of AI taking action in their SOC until explainability improves

Given the desire for trust and explainability we are seeing from buyers, it's important for them to be equipped with the right questions to ask vendors during an assessment or POV of AI tools in order to demystify marketing hype from real operational outcomes.

Below is a list of categories in which buyers can assess AI vendors or AI Service Providers (AISPs) to help reach safe adoption and maximize their ROI.  

5 categories of AI vendor assessment

Darktrace groups these AI-related questions into 5 categories: governance, data and training, model and technique choice, performance validation, and interpretability and adjustability. By asking questions regarding each of these 5 categories, buyers can gain a deeper understanding of how an AISP’s systems work and whether they suit their business requirements.

Governance, safety, and data controls

Governance of AI systems is critical for all AISPs. Whether their platform is based around a single model, or is a more complex, composite AI solution, strong governance is essential to ensure the system is safe, robust, and reliable.

A simple question you could ask is:

What AI governance policies and frameworks do you follow, and/or certifications do you currently maintain?

For more questions you can ask vendors, download the full guide here.

Darktrace is certified to the ISO/IEC 42001 standard, the world’s first AI Management System (AIMS) standard. ISO/IEC 42001 addresses the unique ethical and technical challenges AI poses by setting out a structured way to manage risks such as transparency, accuracy, and misuse. This includes a commitment to ethical AI development, and effective management and monitoring of AI systems both prior to and continually after release.

Data gathering and training

Accurate, meaningful, and unbiased data gathering is the first important step in producing any AI system. An AI model trained using inaccurate, unbalanced, or poor-quality training data will fail to perform optimally.

To alleviate concerns regarding training data quality, a question you could ask is:

What steps do you take to prevent bias in your AI models and training data?

For more questions, download the full guide here.

AISPs should be able to provide information about the steps taken, workflows followed, and auditing performed to reduce AI bias where appropriate. While it’s sometimes impossible to fully remove bias from an AI model, appropriate actions should be taken to mitigate or reduce bias where relevant.

Model and technique choice

Different AI techniques are optimal for different tasks. For example, research from Gartner suggests that relying on a single “one-size-fits-all" model can lead to data gaps, especially in highly specialized domains.

To achieve more accurate and robust AI solutions, AI leaders should move beyond using just one model or technique, embrace composite AI practices, and adopt a holistic AI system perspective.

A straightforward question you could ask is simply:

What type(s) of AI model(s) do you utilize in your solution?

For more questions, download the full guide here.

While specific detailed information about custom systems used by AISPs is likely proprietary, buyers should expect vendors to be able to provide an overview of the broad techniques used. This will allow you as a buyer to determine if the type of model is appropriate for your use case.

Performance and accuracy validation  

Testing and evaluation of performance is essential for all AI systems. Performance analysis should be performed both before release and continually after release to identify potential data or model drift.  

A question you could ask to understand an AISPs testing workflow is:

How do you audit, test, evaluate, verify, and validate your AI model outputs?

For more questions, download the full guide here.

Testing workflows will likely vary depending on the type of model – measurements relevant to one system may not always be relevant to others. Assessment of systems should also extend beyond these standard accuracy and robustness tests, and should also feature physical performance, such as latency and resource consumption.  

Interpretability, adjustability, and transparency  

AI systems are typically a black box, simply providing an output without an explanation of how that output was attained. Interpretability and transparency are critical to ensure that both SOC teams and end-users trust the outputs of a system to be accurate and meaningful.

A question you could ask is:

How do you promote a trust relationship between human analysts and AI outputs?

For more questions, download the full guide here.

In the context of cybersecurity, trust and interpretability are even more essential. This is particularly relevant for generative AI-based systems (including most AI Agents), where the risk of hallucination can reduce trust in responses.

Cybersecurity systems often need to perform autonomous actions to block incoming threats – an email filtering system may hold potentially dangerous emails; a firewall may block malicious inbound connections. If SOC teams can’t trust these systems to perform accurately, these systems may be limited or disabled, critically reducing their defensive power.

Darktrace as an AI-native cybersecurity vendor

Darktrace has been building and applying AI in cybersecurity for over a decade, developing its capabilities alongside an increasingly complex and fast‑moving threat landscape. This experience has resulted in a mature, multi-layered approach to AI, which continuously learns the normal patterns of each organization to understand behavior, interpret context, and identify meaningful deviations — without relying on predefined rules or known attack signatures. Over time, this has enabled a proven behavioral understanding that helps uncover subtle signals of risk that may otherwise be missed.

With the backing of our ISO/IEC 42001 certification, stakeholders, customers, and partners can be confident that Darktrace is responsibly, ethically, and safely developing its AI systems, and managing the use of AI in day-to-day operations in a compliant and secure manner.  

Explore the principles behind Darktrace’s responsible AI approach, informed by collaboration with global experts in academia and governments, detailing how accountability, explainability, and continuous validation are built into its cybersecurity technology.

How Darktrace secures AI systems

Darktrace now brings these capabilities to monitor and respond to risk generated from AI systems across organizations with Darktrace / SECURE AI. This solution analyzes how prompts, agents, and systems are used within the context of each organization, bringing every AI interaction into a single view. This unique approach helps teams understand intent, assess risk, protect sensitive data, and enforce policy across both human and AI agent activity.

Stay up to date

Sign up for the Secure AI Readiness Program here: This gives you exclusive access to the latest news on the latest AI threats, updates on emerging approaches shaping AI security, and insights into the latest innovations, including Darktrace’s ongoing work in this area.

Ready to talk with a Darktrace expert on securing AI? Register here to receive practical guidance on the AI risks that matter most to your business, paired with clarity on where to focus first across governance, visibility, risk reduction, and long-term readiness.  

Further Reading on AI in cybersecurity

When deciding to invest in an AI solution, it’s important to understand what this means for you and your organization. The questions presented here are only a starting point in understanding an AI solution and whether it is appropriate for your use case.  

Gain deeper knowledge on applications of AI in cybersecurity and Darktrace’s multi-layered AI in the AI Arsenal White Paper.

[related-resource]

Continue reading
About the author
Jamie Bali
Technical Author (AI) Developer
Your data. Our AI.
Elevate your network security with Darktrace AI