Pandemic Fuels New Wave of Surveillance Tech

Whispering, jitteriness and signs of frustration are behaviors compliance teams can pick up on when scanning a trading floor or a team of portfolio managers. With staff working from home due to the coronavirus pandemic, compliance officers in banks and asset managers have to find new ways to monitor employees in the absence of observable behavioral cues.

“You pick up people’s body language, their movements, if they are going to meeting rooms [at strange times]. You overhear things in the kitchen. You don’t realize how much information you absorb about what is going on in the firm [when in the office], if people are anxious, shifty, or whispery,” said Clem Geraghty, head of compliance at Ardevora Asset Management, speaking during a panel at The Summit for Asset Management (TSAM) virtual event last week.

The risk of employees making mistakes or committing market abuse is heightened, as Covid-19 has caused massive market volatility. Portfolio managers, traders and salespeople could panic due to missed targets, and try to claw back losses by committing fraud. Other issues could arise if a salesperson speaks to clients outside of their jurisdiction, or their professional category, said Geraghty. “We’ve learned from the financial crisis that when times get tough, people do silly things,” he said. 

These conditions mean that banks and investment firms have to step up their surveillance measures to better monitor trader behavior and deploy broader coverage of front-office communications to detect any signs of unlawful activity or market manipulation, such as insider trading, money laundering, and market rigging.

Ardevora’s compliance team, for one, is having to monitor front-office communications by sifting through a raft of emails, chat messages, and call recordings. But there is only so much information that can be gleaned from electronic and voice communications; having a literal line of sight into the office is crucial to the surveillance process.

Moana Moore, head of compliance at Trition Investment Advisors, said her firm normally monitors electronic communications. “But we’ve been doing a bit more in relation to e-comm monitoring and watching out for unusual behavior, because we’re in unusual times, and we’re not on the trading floor anymore; we’re working from our homes, and those boundaries blur.

“There are disciplines that we’ve instilled into our investment business professionals while in the office, but they run the risk of blurring when they’re working from home,” said Moore, who also spoke on the TSAM panel.  

Moore said she has been pushing her firm to record more trader conversations and weekly meetings, but has been met with some reluctance internally. 

Blurred Lines: Personal Vs Business

Traders and portfolio managers aren’t allowed to use unmonitored devices for business communications, but working from home has made this increasingly difficult to police.

In response to the challenge, many vendors, such as NICE Actimize, Digital Reasoning, Behavox, and more recently, Symphony, have integrated with messaging applications like WhatsApp and WeChat. Franklin, Tenn.-based vendor Digital Reasoning can overlay surveillance capabilities with the Zoom video calling platform, which has jumped from 10 million to 200 million daily users during the pandemic.

But these communication apps are also for personal use, so overlaying monitoring tools raises concerns around privacy and keeping the division between personal and business matters. A key consideration in this regard is the duration of current lockdown situations in various countries. If they disrupt firms in the long term, traders continuing to work from home will probably have to accept a privacy trade-off.

“If more remote or home trading becomes the medium- or long-term outcome, I think you need to contemplate what that divide is going to be between private and business and what that is going to look like,” says James Kemp, managing director at the Association for Financial Markets in Europe.

“Do we need to have completely intrusive technology in remote environments? Or are we going to find solutions that work differently? Alternatively, rather than trying to think you’re going to solve it all on the front end—which might mean cameras, recordings, etc.—do you rather enhance what you do on the monitoring side for post-trade?” Kemp says.

Banks and investment firms are legally obligated to monitor front-office communications on their own devices and software, under rules such as the Market Abuse Regulation, but in March, several regulators, including the Financial Conduct Authority (FCA), issued forbearance to financial firms on the recording of calls. Yet in those situations where it is not possible to record the communications, firms are expected to take steps to mitigate outstanding risks, such as deploying enhanced monitoring or having the ability to retrospectively review trader activities post-Covid-19.

So if a regulated individual is using a personal device for business interactions, that must also be monitored, but that requires employee consent. 

“One of the things that’s clear is that employees should have no expectation of privacy. If, for example, you are sending an email, the company has the right, in most cases, to read that email. So, you should not be doing it if you don’t want anything observed,” says Lee Garf, general manager, and chief product officer at NICE Actimize.

A monitoring application is very simple to deploy.  A trader or portfolio manager can easily install an app on their mobile device, the financial firm would then secure that device and the technology would operate behind the scenes, monitoring the user’s communications.

But how does compliance differentiate business conversations from personal ones?

They don’t, says Danielle Tierney, a capital market and financial technology analyst at Greenwich Associates. When a device is monitored, that includes all calls and interactions.

However, Erkin Adylov, founder of Behavox, says the vendor’s machine learning technology can distinguish between personal and business-related conversations.

“The technologies have matured enough where [the machine learning tech] actually knows that this is related to business and this is related to personal. It is able to distinguish that, and as a result is able to silo conversations that are private and put them into a box that says, ‘This is not important, not relevant, don’t read it.’ And anything that’s related to business it would flag up to a different algorithm to check the business communication and make sure there are no compliance breaches,” he says.

Other vendors are building functionality to alert users who are messaging individuals with a tapped device that their conversation is being monitored. The user would get a notification if they messaged a person with a monitored device.

One of the biggest pushbacks from employees subject to surveillance is about who pays for the device in the first place.

“I do hear complaints if the firm does not provide or pay for the device,” Tierney says. “So, they think, ‘Why do I have to put your monitoring [tech] on my device when you’re not even paying for it?’” 

Tierney says an employee could choose to act unlawfully by using an unauthorized device, but the bottom line is that financial services firms must provide their staff with the necessary tools to be compliant.

As a second line of defense against misconduct, financial firms are looking to new technologies that can profile the behavior of front-office teams.

A New Frontier: Behavioral Monitoring

Traditionally, financial services firms have detected market being discussed in voice or chat communications using a lexicon, a list of words relating to bad behavior. But in cases like Libor fixing, traders just avoided using those words to trick the system, or took their conversations offline. The solutions that emerged after this scandal and others include surveillance technologies that used natural language processing and machine learning to identify unusual behavior and understand trader sentiment.

Digital Reasoning captures data from emails, chat applications, and voice calls. It transcribes voice to text, and the machine learning technology is trained to detect language that could be connected to market abuse and understand the context that frames it. The vendor also looks for unusual behavior patterns that could indicate secrecy or collusion.

Tim Estes, executive chairman and co-CEO of Digital Reasoning, says there are three important indicators to look out for when monitoring trader behavior: signs of secrecy, business language, and the parties changing their communications venue.  

“If someone is trying to not be detected because they are talking about something sensitive, that is not wrong. Sensitive could be a personal thing, sensitive could be a professional thing. Once it is a business thing, with business deal language … and they get a little bit concerned and decide they want to take it off-channel, like through WhatsApp, because it’s too sensitive [for the monitored channel], when you have those three behaviors together, that is a real warning sign,” Estes says.

Digital Reasoning built a tool called Cognition, which is used to fast-track a firm’s ability to train machine learning models. The tool uses active learning and interfaces to allow analysts or compliance teams to input training data, such as examples of unusual behavior, without the need for data science skills.

One example is pressure language, when sales teams become too aggressive with clients. Digital Reasoning’s training tool has been used to detect and pinpoint signs of this happening.

Behavox provides an end-to-end data platform designed to help financial firms act on their data. Within that, it uses machine learning technology to detect patterns and risks across voice and text data and supports multiple languages, including English, Arabic, and Russian. The system can then generate specific alerts that are tagged and sent to the compliance team. Some of the tag categories include bribery, insider trading, discrimination, racism, sexual harassment, and corruption. The Behavox platform works by masking the sensitive data and storing it in a location where the financial firm’s compliance team can only access it if they have appropriate authorization.

“It’s stored in a location that is accessible only to the legal team that has the client privileges, but to get that access, [the alert] has to be escalated. So you have to seek permission to it through the Behavox system from the head of legal, so only the general counsel of the company can give you the authorization to go and view what the system deems to be personal,” Adylov says.

NICE Actimize pulls in data from order management systems and execution management systems to capture the trade information and help create a timeline around the execution. The vendor takes communications data from emails, desktop phones, turrets if they are a trader, collaboration tools such as Bloomberg, Symphony or Microsoft Teams, and chat applications like WhatsApp and WeChat.

NICE Actimize can also integrate its technology into mobile phones through the carrier under a tri-party agreement with the provider and the financial firm.

“We’re capturing [the data] through the carrier so some mobile providers have support for that, like we have one client that has, globally, seven different mobile providers, and we’re integrating with each of those,” Garf says.

NICE Actimize can also capture location data from the monitored device to see if the individual is trading or working where they say they are.  

“While they are making a phone call, it’s recorded and we can recapture that and then analyze it. Text has a similar idea but there is also attribute information about those communications, like time and the location attributes, that we can capture,” adds Garf.

Data can also be pulled in from human resources systems to triangulate abnormal behavior; for example, if a trader is on vacation but the system shows they are actively trading.

Cloud adoption has been a significant driver in being able to access these data channels, says Garf. Prior to the cloud, normalizing data on-prem has proved complex and time-consuming.

Advancing the Scope

Regulators are looking for financial firms to better detect areas of misconduct. As a result, banks and asset managers are taking steps to broaden their scope beyond trade surveillance to include workplace analytics and human resources-related monitoring.  

As mentioned above, firms are looking at ways of identifying and stamping out discrimination, racism, indicators of sexual harassment, or bullying in the workplace. Another sign being monitored is employee morale.

Some believe Covid-19 has exacerbated the need to deploy technologies that can more closely monitor and flag these types of instances.

“Pre-Covid, it was easier to enforce those rules. Now it is more difficult to enforce, and how do you ensure that there is no ‘bro’ culture happening in the company? What we’re seeing across our clients is a spike, an increase in the sexist, racist, and discriminatory language that people are using in their communications between each other. There is also a rise in frustration from the work-from-home environment,” Adylov says.

This is done by monitoring communications and analyzing text and sentiment from surveys and complaints, both from the financial firm’s employees and clients. This could detect potential flight risks, where an employee is planning on leaving, or a severe customer complaint that could lead to litigation.

Estes says these kinds of analytics have to be used by a trusted party within an organization to ensure they are dealt with appropriately and effectively.

“I think the audience for that has to be immensely qualified, like HR professionals who know the law, who is going to handle that kind of stuff immensely and sensitively, and not just any peer/supervisor,” he adds.

Adylov also says banks, hedge funds, and regulated entities are further extending their monitoring capabilities to cover their entire employee count. Adylov adds that regulators are encouraging this level of surveillance as the technology becomes more available. 

He says two of Behavox’s clients—a very large Japanese bank and a Canadian bank—are both looking to extend monitoring to all their employees over the next three years.

“So the industry is changing from sampling, from randomized monitoring and limited monitoring, to monitoring of everybody, because it’s the best thing to do, because the regulator is demanding us to do it and technology is available to do so,” he adds.

Orwellian Issues

Systemically important institutions, like major banks, are heavily regulated because of their role in supporting the economy. Regulatory scrutiny could lead to all employee business interactions and communications being tracked in the future. That surveillance could spill over to employees’ private lives. 

Tierney says that financial firms are obligated to monitor activity that is tied to trading. Anything beyond that should not be considered part of market surveillance.

Some have raised concerns that tech solutions go far beyond what is necessary for normal monitoring. Kemp says some providers are starting to deploy camera technologies for recording the user to make sure they are in front of the screen. If they move away, a message pops up to ask: Where were you and what were you doing?

Estes says providers have a responsibility to ensure they do not overstep their position such as by accessing the data or making employees feel uncomfortable with the technology being deployed.

“The area of workplace analytics has to be done as a very close and thoughtful development with a client or a bank because you could be getting into almost Orwellian issues really quickly. It’s very important to find the line between listening and awareness so that you can be a better employer and coach a better workforce, [without] everyone feeling like they are being spied on,” he says.

Regulators also need to give clarity around what they require firms to do, especially if work practices will remain disrupted for an extended period of time due to the pandemic.

“There are probably some enhancements to guidance that are required as well. As we start to rapidly deploy some of this technology that we already have, like Zoom and chat, what are going to be the rules around that? I can see that being an area where some clarity is going be needed,” Kemp says.