Podcast
A Cybersecurity Wish List for Businesses of Any Size
What actions should businesses be taking to strengthen their cyber defenses – and why does each one matter? Joan Woodward, President of the Travelers Institute, sat down with John Menefee, Vice President and Enterprise Cyber Lead at Travelers, and Fayon Atkinson, Senior Manager of Cyber Risk Services at Travelers, to explore recommended actions and guidance for businesses and why they matter, considering the growing role of AI in cyber tactics used by threat actors.
John and Fayon connect each recommended practice to the real-world risks it addresses, offering practical, actionable guidance to help organizations make smarter decisions about their cyber defenses.
This episode is part of a special cyber-focused series of the Travelers Institute Risk & Resilience podcast. This program is presented as part of the Travelers Institute’s Cyber: Prepare, Prevent, Mitigate, Restore® initiative, which promotes dialogue and education to help leaders prepare for and respond to cyber incidents.
JOAN WOODWARD: Hi, folks, and welcome back to the Travelers Institute Risk & Resilience podcast. I’m your host, Joan Woodward, President of the Travelers Institute. Cyber threats are evolving fast, and staying ahead of them has never been more critical for organizations of every size. This special four-part cybersecurity series brings together cyber experts from across Travelers to unpack the issues that matter most for organizations navigating today’s threat landscape. It’s presented as part of the Travelers Institute Cyber: Prepare, Prevent, Mitigate and Restore® initiative, which promotes dialogue and education to help leaders prepare for and respond to cyber incidents. In this episode, I sit down with John Menefee, Vice President and Enterprise Cyber Lead at Travelers, and Fayon Atkinson, Senior Manager of Cyber Risk Services, also at Travelers, really to explore recommended actions and guidance for businesses and why they matter, considering the growing role of AI in cyber tactics used by threat actors. Take a listen. John and Fayon, I am so thrilled to be joined by both of you for our miniseries finale. Thanks so much for being here. OK, let’s start by understanding your vantage points. Fayon and Cyber Risk Services. You’re advising policyholders on what they need to do to reduce the risk. And John, you’re seeing how those risk decisions affect underwriting and claims. So how do these two perspectives inform each other? Well, what’s the conversation like between your teams?
JOHN MENEFEE: I can start. So our Cyber Risk Services team is at the forefront of helping our policyholders avoid being the victims of cybercrime. And the services that they perform is a big part of the value of buying cyber policies from Travelers. It’s there. The insurance policy is there if a claim happens. But we’re going to try to help our customers as much as we possibly can to avoid being the victims of cybercrime in the first place. And that’s becoming more important than ever in why agents and customers choose Travelers as their cyber insurance carrier. And all of our teams work closely together to identify trends that can inform policyholder outreach to prevent those claims, as well as underwriting and risk selection and even rate plans. It all works together, and the better we get at understanding what’s driving claims, the better we are at underwriting the product and managing that exposure.
FAYON ATKINSON: Yeah, John. Well said. You know, for me and my team, especially, we work really, really closely with the underwriters. And some of our conversations, they happen frequently on a one-on-one basis. So at times, we may work with them to assess the risk of new businesses. And so that would look like hopping on a call with a potential risk or potential new business or someone that’s renewing just to understand, you know, if there were gaps in the infrastructure – if they’ve had an incident, sort of what they’ve done to improve their infrastructure, to prevent something like that from happening again at some point. Sometimes, we have larger meetings between Cyber Services and the Underwriting team. And to John’s point, it’s really around, you know, where are we seeing exposure? What data do we have to inform those decisions on underwriting?
JOAN WOODWARD: All right. Great. So there’s a huge gap between what organizations think they’re doing. And in reality, 90% of those believe they’re patched, but it’s actually closer to 60%, and that’s attackers’ hunting ground. You talk about a 72-hour window specifically for critical systems. Why 72 hours? Why not 48? And what should agents be asking their clients about this?
FAYON ATKINSON: Listen once upon a time, you know, organizations, when it comes to patching, it would take them, you know, weeks or sometimes even a month before they apply a patch, right? Because in the process, you will have a patch available for vulnerability. Once that vulnerability has been published or that flaw in the software has been published, and then sometimes organizations just sort of drag their feet at it. And it could be a number of factors that result in that. And some of them could be they have some concerns around testing. Is there compatibility with other systems that they have running, or do they need to have a change management approval, like inventory could be a factor. That’s how it was once. But now with these new frontier AI models that completely change, what we’re seeing is that attackers based on, you know, when there’s a vulnerability that’s been published, that window that organization would have to patch has shrunken simply because these attackers, with the use of that AI model, have been exploiting these vulnerabilities at a much, much faster pace. And so that’s why we’re recommending that most organizations take about that 72 window to patch. And to your point, Joan, like, why not 48 if you can do it in 48? Go ahead. The quicker you can get it done, the better it is, right?
JOHN MENEFEE: And I completely agree that the policyholders that can demonstrate timely and effective patching, they’re not only protecting themselves more effectively, they’re demonstrating to their insurer that they’re better risks as patching becomes more important. You know, sometimes these threat actors, they’re targeting one particular vulnerability, and the more effective at patching, the less likely you’re going to end up in the threat actor’s crosshairs. And that’s what we want to try to avoid. We want to avoid being the target of threat actors in the first place. And so carriers are finding it more and more important than ever that insureds patch effectively. And it can affect terms, conditions, organizations that are much slower to patch than other similar policyholders. They might get different terms; they might get different pricing. And those that can effectively patch probably are going to get better terms, better conditions from their insurance policy, from their insurance carriers.
JOAN WOODWARD: OK, great. So if I’m an agent listening right now, is this something I can ask my client about in a 10-minute phone call? Or does it require a deeper conversation?
FAYON ATKINSON: Historically, the questions around patching have really been like a check-the-box or an IT ticket. So the questions are, do you have a patch management policy? Yes or no? And checking “yes,” which is great that you do have it, but I think as an agent, what we should be asking is understanding sort of the metrics around it. So do you have a like, can you confirm there’s compliance or what’s your compliance rate for patching? So I’d like to say making sure that the organization or their client have that patch management policy as a part of their leadership metric, and not just like a quick check the box. Because at that point, it sort of forces them into your question around, you know, that 90%, oh, we think we’ve patched, but in actuality, only 60% of that is true. So making that sort of a leadership metric – I think there’s more eyes on that. And so that’s where a conversation with an agent to their client could sort of force that discussion. Because again, that window of not patching is really, really the sweet spot for an attacker, especially today.
JOHN MENEFEE: One of the interesting things, too, is that, you know, sometimes organizations aren’t aware that they haven’t patched – the inventory of their systems isn’t complete, and they don’t have perfect visibility. And that’s another area where outreach can help, where we might see something that our insureds don’t. And we might reach out and say, have you patched this vulnerability? This is something that we’re seeing exploited out in the wild. And you should patch to avoid being targets. That’s all kind of entangled in the outreach that we do to help our customers maybe see things that they don’t see themselves despite their best efforts.
JOAN WOODWARD: OK. That’s great. Let’s shift a little bit and talk about detection and response. So we know many organizations have detection tools, but those are only as valuable as when someone’s actually responding to alerts, right? So attackers deliberately strike at nights, weekends, when our IT teams are relatively thin. So tell us what is the difference between having a tool and actually having managed detection response or MDR? And John, how do you view MDR and the underwriting process then?
JOHN MENEFEE: Yeah, I can start. MDR was a tool a few years ago that really only the larger organizations could afford. It was kind of a developing technology that now has become more available over time to small organizations, and we see it more widely used, but probably not as widely used as we’d like. And it’s really effective because traditional antivirus, you know, it’s black or white. It doesn’t detect some of the behaviors that go along with ransomware claims, and MDR does. And so that’s why it’s so effective. Effective is, it can detect unwanted behaviors in your system rather than just certain software that exists in your system like antivirus. In fact, it’s a tool that’s often used in our breach response engagements. So when there is ransomware or other activity going on in your systems, oftentimes our vendors are deploying MDR to contain that into certain systems so it doesn’t spread even further. And we’d much rather see it deployed on endpoints throughout the system before the claim to prevent a claim than we do afterward to stop the viruses and ransomware from spreading. And so the more and more organizations are adopting this technology and monitoring it, the better. And we view it very positively on the underwriting side.
FAYON ATKINSON: I’m just going to say I like that you highlight the tools and the idea of traditional antivirus and MDR. And I think for questions that we see often is around, does this tool meet the underwriting requirement based on what we were asking for? So I always like to define what each of them are. So when we talk about antivirus, we are talking, OK, what’s already existing, the known threats? And so antivirus tools are the traditional ones that we’ve all had on our devices. So they have what we call “signatures” and it lives in their database, and if they see anything or those bad files based on that signature, that’s when they flag and say, hey, we noticed this; let’s block it. Now, with today’s attackers, that may not always be the ideal tool that we want to see. And so we get to EDR. EDR is the next level up. EDR has the same capabilities as antivirus, but it’s more of a, like, I’d say, provides real-time visibility. So instead of just checking files against the list that they already exist or those known threats, it watches for behavior on your devices. And anything that’s unusual, it sort of flags it and sends an alert. Now, the interesting thing about EDR is it’s just a tool, right? It’s not a service. So even though it’s sending those alerts, it’s asking that, hey, who’s on the other side? Who’s that person that’s looking at that alert to validate it or know what the next action is? And that’s where the MDR comes in. MDR, it puts that human element to the tool itself. So you’re getting that 24/7 monitoring. And if there is a notification or an alert saying, hey, there’s something unusual going on, there’s a human body or like an analyst that can take a look at that and say, OK, what’s next? And so with companies that we see with just that MDR tool, the human error pieces, they could get an alert on, I don’t know, like a weekend or a Sunday or right before a holiday, and no one’s monitoring that inbox where the alert is going to. And so there becomes that lag time when no one’s checking their email. And so an attacker is doing its damage on a Sunday afternoon. And then on Monday, they go in and they’re like, oh my gosh, what’s happening? Right? And so that’s why MDR is super, super important, because it’s 24/7 around-the-clock monitoring. And so that’s why I like to say, hey, maybe get an MDR as opposed to just that EDR tool.
JOHN MENEFEE: And some of these tools, they’re only effective if you properly configure it. They’re only effective if you’re properly using it. And, you know, that’s why some of the consultation services that we offer are so important, to just get a second set of eyes. You think you have something – we can walk you through those proper configurations. We can talk about some of the considerations relative to monitoring some of the gaps that might exist. And it wouldn’t necessarily be, oh, you need this thing; it’s you already have this thing. What are the ways to optimize its use more effectively to prevent being the victim of cybercrime? It’s just so important.
JOAN WOODWARD: Exactly. Let’s shift gears a little bit and talk about backups. So ransomware gangs target backups first to eliminate recovery options, right? And here’s the brutal part of it: Most organizations never test their backups. So tell us about where that became a nightmare. Explain your three-two-one strategy, I think you call it, for someone who’s never heard of the three-two-one strategy.
JOHN MENEFEE: I can start there. And as far as the nightmare scenario, if you’re thinking about these threat actors, they’re all financially motivated and they’re trying to increase the odds of a payday. And a threat actor’s odds of getting a ransom payment go up exponentially if they can encrypt your systems and you don’t have viable backups to recover from. You’re much more likely to pay a ransom if you can’t access any of your files and your business is shut down, versus if you can access your files, you can back up from them. You can recover from them. You have options. And one of those options is don’t pay the threat actor.
FAYON ATKINSON: So what is effective backups? I guess I can chime in on what that looks like, right? I like to say an organization is as good as their backups. That’s the critical part in that business continuity, to John’s point. So we talk about that three-two-one backup strategy. So for me, what that sounds like is redundancy and resiliency. So the “three” is three copies of your data. “Two,” it’s that it’s on two types of medium – so that could be a hard drive or a tape or in the cloud. But of those two, one of that copy should be off-site and kept offline or immutable, meaning that an attacker is not able to make any changes to that backup. They can’t delete it. They can’t write it up. So it’s completely untouchable. That’s the goal. That’s that three-two-one backup. But I’ve talked to organizations and they’ve had their backups, right? But imagine having the efficient backups. However, when and if you’re experiencing a cyber event where you do need to get access to your backup, you’ve never tested your backup. And so there could have been some point in your backup strategy that there’s a failure. So that’s why we like to say, make sure you test your backups too. So while having an efficient backup is number one, testing your backups is number two. An organization can get hit, you know, in the middle of restoration or getting back up and running from those backups. At some point, you know, months ago, the backups were silently failing. So testing, having a real robust test schedule is super critical here. And I like to say sometimes, you know, an untested backup isn’t a backup. It’s more of an assumption that you have a backup.
JOHN MENEFEE: Those are all really good points. And the recent backup, you know, if you’re in a position where your data is backed up but it’s three months old and you’re faced with losing three months’ worth of work, then your options might be limited again. It’s, do we back up our data instead of paying the extortionist for the decryption keys? And you have to gauge whether or not you’re OK with three-month-old backups or paying the ransom for the decryption keys. And paying for the decryption keys isn’t always a silver bullet. It takes a while sometimes to decrypt data from those keys. And so, you know, you could be down for a significant amount of time during the decryption process, when if you had a good backup plan that was tested regularly, the backups are resent. You practice at recovering. It just makes a bad situation so much better when you’re back up and running sooner without having to pay an extortionist for decryption keys.
JOAN WOODWARD: All right, let’s talk a bit about AI governance and Shadow IT. So employees are pasting sensitive company data into AI tools every day really without thinking about it and Shadow IT is everywhere. So give us a concrete story of what could go wrong when that happens. And what does AI acceptable use policy actually look like, you know, for a small business?
FAYON ATKINSON: I can jump in on some of what could potentially go wrong should you know an employee just shared sensitive data with the use of AI. I think everyone is sort of getting hip to it and using it. And so if we have a scenario where an employee is using maybe like their personal AI tool and this hasn’t been approved by the company or it’s not the enterprise version, and they could be using it for something as small as, you know, just, hey, draft this client email or summarize this contract. They’re essentially putting data, and this data may be now feeding like a new model because essentially what AI is doing is they’re constantly taking the data. You’re feeding them to train how to respond to that. So that could constantly be happening. And so even with it not being like a company-issued or enterprise version of that tool, there’s a lack of visibility. And so there’s no monitoring happening from the enterprise side. And so now you’re providing sensitive data out in the ethos, and who knows what can happen with that. So those are some of the risks that I can think of. John, I don’t know if you’ve seen anything else.
JOHN MENEFEE: Yeah. You know, this is an interesting developing risk. And we just released the Cyber Risk Index that we do every year. And more and more questions related to AI were included. And one of the things that I was most interested in was related to AI acceptable use. So 89% of companies reported that at least some of their workforce uses AI tools, but only 59% have established AI business practices in place. And it’s probably immature that 59% isn’t quite as mature as you’d like to see it. And even there, that 30% gap really stands out. Companies are very focused and concerned about what cybercriminals are doing with AI and what they’re potentially capable of. And that concern is justified. However, you know, companies should be equally focused on the risk created by their own AI use when the right practices aren’t there. And AI can be a really effective tool, but it’s important for employees to know what they can and can’t do with it so that they aren’t creating unwanted business risks.
FAYON ATKINSON: Yeah. that’s interesting. You talked about the level of maturity. So for us, we do offer an assessment on an organization’s AI maturity. It’s a consult. It’s our AI risk assessment that we offer to our cyber policyholders. Is there a review from the human eye, and sort of like any new tools that are being considered? But yes, I like to tell organizations, you know, there’s no point in banning AI because it’s being used.
JOAN WOODWARD: John, we actually did an entire episode previously about the AI risk assessment. So for those that are listening, I do recommend that you go ahead and take a listen to that. You learn a lot more about the acceptable use policy there. So AI-powered phishing is getting more convincing and personalized. What makes an AI-generated phishing email different from ones we’ve been getting for 20 years? I mean, the ask is to move away from SMS codes toward authenticator apps or hardware keys, but how do you get organizations there when you can barely get employees to use the SMS consistently? And, Fayon, maybe you know in your advisory work: How difficult is device-level verification to actually implement?
FAYON ATKINSON: So with the changes that we’ve seen with AI-generated phishing, in the past, I’m sure we’ve all gotten some email at some point where it could be pretty obvious that it’s a phishing email. They’ll have weird formatting, you know? It’s pretty obvious that it’s from the wrong sender. It could have bad grammar, but now with AI, it eliminates most of that, right? These are very, very sophisticated phishing campaigns. And they’ve increasingly become much, much harder to detect. And then to your other question around, you know, moving away from SMS codes, my pitch is not all MFA are created equal, right? So that second layer of authentication, it can be in the form of SMS or email, but I think those are the two weakest links, especially SMS codes. We try to say, hey, whenever I speak to a policyholder about this, like you want to consider those phishing-resistant MFA methods, especially for those in leadership or have high-risk accounts, and the good thing about hard keys and authenticator apps, they’re less likely to be phished. So that’s the good thing about that. And when I talk to organizations about, like, how can we convince everyone to adapt to this method as opposed to the SMS? I can say it’s more of an education. Because I think there’s a misconception around the SMS versus hardware keys or authenticator app. And that misconception is technically like a push notification is less friction than, you know, an SMS. Because if you think about it, you get a text with a code. Now you have to switch out of your app to get that code. And then you have to move quickly to get that code into whatever platform you’re using before it expires, or you have to do that process all over again, as opposed to like a push notification. It’s just, you know, a push of a button. So I think if organizations could sort of change the idea around that, it may change how the users will receive it. The other thing I would suggest is maybe target those high-ranking accounts. So if someone is dealing with finance or has access to sensitive information or like admins, I would say let’s prioritize those types of users first. And then when you do that, maybe stack it with other system changes. That way it feels more like one job as opposed to like an extra ask for everyone who’s involved or for part of their day-to-day. So it’s more of a psychology, I’d say, than a technical change.
JOHN MENEFEE: Joan, we used to say that any MFA was better than no MFA at all.
JOAN WOODWARD: Yes.
JOHN MENEFEE: And I think that’s still true, but it’s becoming less true over time. And there was a lot of work that was done recently, you know, since 2020, I’ll say, and I’m just throwing that date out there, where lots and lots of organizations implemented MFA that didn’t have it before. And that’s a lot of work. It was inconvenience on the workforce. It was a lot of work. And now we’ve reached a stage where it’s becoming less and less true that any MFA is good enough. And so organizations just need to stay on top of these risks to understand where the weaknesses are. The phishing-resistant MFA technology is going to be more and more important over time. And phishing in general, it’s very important still to train your workforce. It’s getting harder and harder to identify what is phishing, what’s not. But there are still signs. People still need to be hyperaware, and organizations need to kind of create the backstop, the technologies that can protect the organization if someone falls for it, but still train employees on phishing identification, you know, what to look for, what might be indicative of a phishing attack, you know, what could be risky versus not all of that is still, you know, organizations can’t just sit and be content with what they’ve done. They need to keep evolving as the threats evolve.
FAYON ATKINSON: Yeah. And I’ve talked to some other insurers or different organizations that talk about sort of how they’ve incentivized phishing trainings. And I’ve heard some very, very creative ones. And I think, you know, I get those phishing campaigns and I get, you know, this click rates varies, but I wouldn’t keep going. With time, like you said, it gets much more difficult. And I remember those days back in 2020 when MFA was so hard to implement. And it would take time, but now it’s built into just about everything. So again, it’s really training and just moving with the time. I guess that’s the beauty of cybersecurity. It changes so rapidly.
JOAN WOODWARD: OK, let’s talk about leverage and action then. So six security domains is a lot. For someone feeling overwhelmed, what’s the most important thing they can tackle first? And really, our agents are in a unique position with hundreds of relationships. What should they be doing differently maybe starting next week on this?
FAYON ATKINSON: Yeah, I can start that one. So if I were to think about this, I would say backups would be sort of that one thing I would say to tackle first, right? Because your backups, again, it addresses that resiliency, the redundancy and the testing. So if all else fails, right? I’ve heard before, you know, it’s not about if, it’s about when. So let’s say your MDR catches something a little too late or, you know, there were some issues and you were too late to patch, right? If all else fails, having really good backups could prevent or minimize that downtime for an organization. So again, making sure that you do have that three-two-one backup strategy. And then that backups are being tested, and making sure that you have updated backups. I’d say that’s my No. 1 priority.
JOHN MENEFEE: And to answer your question about what agents can do, one of the things that’s unique about the relationship between the insurance company, the agent and the customers, the agent is so often the point of contact with the end customer. They’re the point of trust. And in these circumstances, where the cyber insurer is identifying risky software or a risky VPN or whatever the alert might be, that customer might not be used to getting an email directly from their insurance company. And so when we can, we partner with our agents on these alerts. And the agent is often making phone calls. They’re often reaching out to customers and saying, you’re getting an alert from your cyber insurer. That’s a pretty high bar to alert policyholders. They alert their customers when they’re seeing active exploits in the wild. You should patch this vulnerability as soon as you can. You should take care of this as soon as you can. You should communicate with the insurer in this circumstance as that point of trust. I think that’s something that is becoming more and more common as partnering with agents for these alerts to increase the speed at which action is taken. And it’s important to take action fast. You know, the window is closing in terms of the opportunity to remediate. And the faster remediation happens, the less likely a claim is going to happen. And that’s where the agents can really help. The process is being that point of trust, understanding the process and helping us in this alerting process. It’s a partnership.
JOAN WOODWARD: Well, I think on that note, John and Fayon, that’s the perfect place to leave off today and close out our cyber miniseries, we call it. Thanks so much for being here. And to those listening, I hope you learned as much as I did, and I hope you feel better prepared for your cybersecurity needs. Thanks so much for listening, folks. Thank you for joining us for this insightful conversation. Check out our show notes for more information about the Travelers Institute and to sign up for our mailing list and LinkedIn newsletter. If you enjoy our content, please be sure to subscribe, rate and review on Apple Podcasts and Spotify. You can also contact me by following me on LinkedIn or emailing [email protected] to let me know what you really think and what else you’d like to see from us and hear from us at the Travelers Institute. Thanks again so much for listening.
This episode is also available on Apple Podcasts®.

Host
Joan Woodward
President, Travelers Institute; Executive Vice President, Public Policy, Travelers
Key takeaways
- One of the most consequential factors in whether an organization becomes a target is how quickly it patches a known vulnerability. AI has compressed the window that attackers need to exploit a disclosed flaw, and 72 hours is best treated as a ceiling for patching critical, internet-facing, actively exploited vulnerability.
- There’s a meaningful difference between having a detection tool and having managed detection and response. An alert that no one sees until Monday morning isn’t protection – 24/7 monitoring with a human analyst on the other end is critical to help close that gap.
- An untested backup isn’t really a backup. Ransomware groups routinely target backups to eliminate recovery options, and organizations that can’t restore quickly are more likely to pay. Make sure backups are resilient (multiple copies, at least one immutable) and test the capability to restore these backups regularly.
- 89% of companies report that at least some of their workforce is using AI tools, but only 59% have established AI business practices in place. While organizations rightly focus on external threats driven by AI, that gap points to the need to address the risks that come from their own use of AI.
- AI has made phishing emails significantly harder to spot, and text-based authentication is increasingly the weakest link in the multifactor authentication chain. Phishing-resistant methods like authenticator apps and hardware keys are becoming the standard worth moving toward.
Travelers Institute Risk & Resilience podcast
Subscribe on LinkedIn® (Opens in a new window)
The Travelers Institute Risk & Resilience podcast is available on Spotify and Apple Podcasts®. Subscribe today to be alerted when new episodes are released.
Related content
Episode one
AI’s Impact on Cyber Vulnerability and Defense
Lauren Winchester, Head of Cyber Risk Services, Travelers, joined Jesse Sugarman, AVP, Travelers Institute®, to explore how AI is reshaping vulnerability discovery, detection and patching – for both defenders and attackers – and what steps security teams can take to stay a step ahead.
Episode two
Closing the AI Governance Gap
Lauren Winchester, Head of Cyber Risk Services, Travelers, joined Jesse Sugarman, AVP, Travelers Institute, to explain what the gap between AI adoption and governance maturity means for businesses – and how leaders can get ahead of the risk.
Episode three
Dark Web Intelligence and the Access Broker Economy
Joan Woodward, President, Travelers Institute, was joined by Josh Doguet, Senior Manager, Threat Intelligence, Travelers, to explore how dark web intelligence can disrupt the access broker economy and stop ransomware before it starts.
In the news
PropertyCasualty360
Digital Insurance
UserFriendly 2.0
Join our mailing list
Invitations, event announcements, replays and more — delivered to you.
Sign up

