Transcription
USEREVIDENCE | THE PROOF POINT | NAVNEET SING
Episode Transcript
This has been generated by AI and optimized by a human.
Mark Huber [00:00:01]:
Opinions are cheap and proof is gold. I'm Mark Huber and this is the Proof Point, a show from user evidence that helps go to market teams, find ideas, get frameworks, and swap tactics.
Navneet Singh [00:00:17]:
In 2025, every B2B marketer needs to be focused on one thing. Building trust. So I went to the hardest industry to pull that off, cybersecurity and and talk to six of the best B2B marketing leaders to learn how they're doing it. On today's episode of the Proof Point, I'm joined by Navneet Singh. He's the VP of marketing of Network security at Palo Alto Networks. Nav and I dove into the big question shaping marketing in his world right now, where generative AI is heading, how it's being adopted in cyber security, and most importantly, how to build trust in a space where skepticism runs so high. This is a great combo. Enjoy.
Mark Huber [00:00:52]:
So, Nav, how are you feeling about the keynote? You got to be pretty excited, no?
Navneet Singh [00:00:56]:
Yes, it is. So it's exciting. Opportunity to talk to audience. It's not very often that you get cybersecurity marketing in a room, a big room at that, with 500 people. So really exciting to be able to talk to them about what I have learned in AI and the discussions that I've had over the last two days have been amazing. I've learned a lot and I hope to be able to use some of that in my talk tomorrow.
Mark Huber [00:01:20]:
I love it. What are maybe the top one or two things that you can share that you've learned since being here at the event? I mean, this is a pretty impressive event.
Navneet Singh [00:01:27]:
Yes, yes. So one, just the importance of in person connection and learning. You know, I get lots of emails from people and I ignore most of them unless it comes from Palo Alto Networks or like bosses specifically.
Mark Huber [00:01:44]:
Yeah, yeah, yeah, I don't blame you.
Navneet Singh [00:01:46]:
Yesterday we went out on dinner, met two or three CMOs, other people working in organizations that organize meetings with C level execs. So just the importance of in real life, learning and connection. I think that's one of the biggest learnings. Yeah.
Mark Huber [00:02:04]:
And this is my first cyber marketing con that I've been to and I've been to a decent number of events this year that are in person, but I think for cyber marketers specifically, in person is just so much more important to them because I feel like it's a safe space where they can share a little bit more openly and freely and not be worried about getting tracked by whatever company is trying to track them and sell them something that's Right. So you recently spoke on the future of generative AI. Can you tell me a little bit about what you're thinking about and kind of where your mind is at with that?
Navneet Singh [00:02:36]:
So this was really interesting. And Palo Alto Networks itself in CyberSecurity is using AI. So we'll talk about that in a little bit more detail. I'm really fascinated with the use of AI in multiple different verticals. Healthcare is one vertical that I'm really fascinated because literally AI can save lives. So I have this example, I have some notes here. The example is there is an LLM called Sibyl. It's been able to forecast lung cancer correctly about 80 to 80, 95% of the time, even before expert human eyes could detect it.
Navneet Singh [00:03:11]:
So imagine if this was a loved one, right? And cancer is, you don't have a cure. The only thing you can do is predict it early enough so that person can be saved. Right? So this is now in further clinical trials at Mass General Cancer Center. So this is really exciting stuff which can actually save lives. Another example, again from healthcare, is Another model called Chief has been trained on 15 million unlabeled the images in five biopsy data sets. Chief achieved 96% accuracy across multiple cancer types. So this is really exciting. And when I look at other verticals, I recently read news that KLM Dutch is using AI to prevent food waste in flights by up to 63%.
Navneet Singh [00:03:58]:
They aim to save up to 200,000 pounds of food that gets wasted. Right? So we have these problems in the world of food shortage and food wastage, right? If AI is able to help in many of these things, it'll be just like a boon for humanity. I think those are the use cases that really excite me. And then you talked about future. So if I can just talk a little bit about that. I think the analogy that I use is that today AI is being used in customer service. So if I lost baggage in the airline, they'll be able to find it faster using AI. But the real use of AI is tomorrow when AI is used so that my baggage is not lost in the first place.
Navneet Singh [00:04:42]:
And I think that's where AI is going. And there are just many interesting use cases ahead.
Mark Huber [00:04:47]:
So I love hearing use cases like that because of the little, let's say B2B tech marketing bubble that I live in and work in. Everyone is talking about generative AI to automate marketing tasks and outreach and personalization and all that stuff. And those are little problems in the grand scheme of things. But when you reframe the problems and use cases for generative AI using some of the things that you just mentioned. The potential of this is so much bigger than my little brain can wrap its head around. Like it's amazing. It's for humanity. It's not just for, you know, B2B tech marketing.
Navneet Singh [00:05:18]:
Yes, absolutely.
Mark Huber [00:05:20]:
Where is cyber at when it comes to generative AI right now?
Navneet Singh [00:05:22]:
Palo Alt Networks. I'll start with our example because I'm most familiar with that. We said that this is really a brand new fight where our security professionals that we serve, our customers, they are fighting adversaries who use AI and cloud and other technologies, maybe in the future quantum computing. So we said this is a brand new fight. We are going to fight AI with AI. So today we use AI. We've been actually using machine learning, deep learning for a while, several years. We now invested in technology called precision AI because we, we know that AI does suffer from hallucination accuracy.
Navneet Singh [00:05:58]:
So we wanted to make sure that it's very precise because it's really, really important in cybersecurity. You can't be blocking benign traffic. So we invested in precision AI, which means that we invested in data which is really rich. AI is only as good as the data. So first thing was to make sure we have rich, accurate data about security threats. And secondly, we invested in machine learning models that are specific to cybersecurity. They are not general machine learning models. And using that we've been able to increase our accuracy and precision.
Navneet Singh [00:06:28]:
In addition, what we've also seen is industry in general. We've done it as well. We've invested in co pilots. We see that there is millions of jobs that are a skills cybersecurity shortage in this industry. And we talk about AI many times in terms of AI taking away jobs. The way I see it is that it can actually help us fill that cybersecurity skills shortage gap. So we've invested in Copilot so that it's very easy for someone to be able to interact with cybersecurity products which tend to be complex. So you could have somebody just graduate from college and be able to use these copilots ultimately to be able to configure products, deploy products, correctly troubleshoot products and this opens up the industry to a lot more people.
Navneet Singh [00:07:19]:
So I think those are some really use cases where we've been using AI.
Mark Huber [00:07:24]:
So from my own understanding, when you use the term copilot, is that the same or different as AI agents and everything that people seem to be talking about on LinkedIn right now so there.
Navneet Singh [00:07:34]:
Are some similarities and some differences. So copilots are typically, for example, in our case, let's take the example of network security copilot. It's been trained on 50,000 pieces of expert documents that we've written so it can answer questions in natural language. So that's the role of a copilot, to be able to answer questions and to be able to guide users towards best practices, recommendations and so on. When people talk about agents, they're talking about more autonomous cybersecurity or more autonomy in general. So agents who not only answer your questions, but are able to take actions on your behalf. So they are going a step beyond and saying, we are going to take action on your behalf. And I think that's ongoing research and that probably might be the next step in the evolution.
Mark Huber [00:08:25]:
Yeah, I hear and read so much about that right now that I'm trying to learn as much as I can and soak up whatever I can and then also try to figure out, filter out what might be real versus what's noise right now. And it's an exciting space, but I think also a very noisy space because of that excitement right now. So this is my first time at Cyber Marketing Con. Have you been to Cyber Marketing Con before?
Navneet Singh [00:08:46]:
No, this is my first time as well.
Mark Huber [00:08:48]:
So, before we get into it, what do you think? Like, this has been a pretty impressive conference.
Navneet Singh [00:08:51]:
Yeah, like I was saying, it's very unusual to have cybersecurity and marketing professionals, that kind of community, in one room. It really feels like a community. So I first got in touch with Gianna and Maria through the podcast, their podcast, and it's really amazing what they've done. So kudos to them. This has been very well organized, considering that it's just like couple individuals who are trying to build this community. It really does feel like it like that.
Mark Huber [00:09:19]:
Very impressive. And I think one of the things that I very quickly learned being here at the event is that cyber buyers are generally a skeptical bunch, and I think for very good reason. But why do you think that that's the case?
Navneet Singh [00:09:32]:
I've talked to many customers and buyers who've been burned in the past, and something was done, let's say autonomously, and some action was taken that probably took down the network. So I've talked to many buyers who've said that the day the first legitimate packet is dropped in terms of networking, meaning I block a legitimate transaction, will be the last day of my job. So security is so critical because you can't block benign traffic, because everything today Is digital blocking benign traffic means you no longer will be able to do an earnings call, a financial transaction, airlines. Right. Flights, bookings and so on. So that's why it's really critical and that's why they look at everything very, very critically and their skepticism is justified.
Mark Huber [00:10:21]:
So without giving away the secret sauce of marketing to cyber marketers, what are some of the things you're doing differently in cyber versus what might work in another industry?
Navneet Singh [00:10:30]:
What I would say is that we're building trust in our buyers and cybersecurity is really different from other industries. So if you look at enterprises, we've found that Enterprises invest in 30 plus cybersecurity tools and that's very different from other industries. Like for example, if you ask them how many CRM tools they have, they typically have Salesforce. If you ask them how many IT operations management tools they have typically have ServiceNow. Cybersecurity is like 30 plus tools. Right. So it is very difficult for them to get a holistic picture, a holistic understanding of what's happening in their network, in their organization. So that's why we are working towards platformization.
Navneet Singh [00:11:11]:
Many customers have been able to replace multiple products with our platforms. Let's take the example of network security platform. We built in invested in a tool called Strada Cloud manager that gives them complete end to end visibility from all the users to the traffic that's going to all the apps, what threats are being blocked and why. And we also show them using technologies like digital experience management. If somebody is having a degraded experience, for example zoom is slow or slack is slow, where does the problem lie? Is it because of high CPU consumption on the laptop or because their wi fi is low or the middle mile or maybe we are having some problem. Poll the networks node. So once you give complete visibility to the user, they see it and once they see it, they believe it. And that's the key to building trust with the buyers, to really let them see what's happening behind the scenes.
Navneet Singh [00:12:07]:
Open the curtain a little bit and that's when they start trusting you.
Mark Huber [00:12:12]:
Are there any campaigns that come to mind or marketing stories that you have from recent memory that have landed really well with your buyers?
Navneet Singh [00:12:19]:
Yes. So at the highest level, we invested in a campaign with precision AI that I was alluding to earlier. So we actually had a contract with a famous Hollywood actor. If you want, you can take a look at the ad on our website as well as on our social channels.
Mark Huber [00:12:34]:
I know what I'm doing after this. Yes, so.
Navneet Singh [00:12:37]:
So from there and we hired somebody who is known for fighting the good fight. So from there we think of use cases. So we take the buyers down the journey of use cases. For example, somebody might be looking at automation. Someone else might be saying that I'm struggling with the recent ransomware attack. Somebody else might say, I want to cyber proof my organization because of a board level mandate. So we take them down these journeys and at the lowest level they want to know how what we've done is we've really invested in product tours on our website so they're able to really understand how our products help, how do they work? So this is like taking a test drive, right? You take a test drive before you buy a car.
Mark Huber [00:13:21]:
I hope you do. Yeah.
Navneet Singh [00:13:24]:
And the same way you should be able to take a test drive, the complex, the product should not be so complex. You should be able to go to a website, take a test drive, decide if this is something that works for you, be ready to have a conversation. So that's the whole campaign that we've invested in.
Mark Huber [00:13:38]:
I love it. So it's funny that you say test drive. I just recently launched a new, we're calling it a demo ranch on our website of a bunch of interactive demos that have very specific use cases and can just show people what it's like to use the tool at their own pace and letting them kind of click through and seeing the tool tips and whatnot. And we've seen, I want to say 250k in pipeline created just by way of these product tours in a very short time frame. So I love that you're talking about letting people take test drives on a site like Palo Alto Networks, which is much larger than our user evidence site, but still amazing nonetheless.
Navneet Singh [00:14:15]:
Yeah, similar principles. Right?
Mark Huber [00:14:17]:
I love it. So are there any results that you can share from this campaign?
Navneet Singh [00:14:20]:
Yeah, absolutely. We measure everything and that's the only way we can justify to the CFO, you know, we might think large company, 100 billion market cap. The CFO still wants to know is it a sound investment or not? So we saw more than 60,000 users watched the entire launch, which was multi hours, more than two hours of launch event. 60,000 people. They are real buyers and prospects who are interested in our products. We had 307 million plus impressions and we saw website traffic grow. Engagement grew by 17%. So this is really critical.
Navneet Singh [00:15:00]:
Growing engagement in such a short span of time is really difficult and we were able to do it because we invested in that campaign that people loved.
Mark Huber [00:15:08]:
So how long did the campaign run for?
Navneet Singh [00:15:10]:
It's actually ongoing. So we started in May of this year and it's still ongoing.
Mark Huber [00:15:15]:
And you know, going into that, did you have a good hunch that this would land as well as it did? Because those are some pretty impressive performance metrics.
Navneet Singh [00:15:23]:
We knew it would be good, but it has been beyond our expectations. It's been really a game changer for us in terms of brand recall, brand recognition, the recognition of precision AI and people wanting to understand our products better.
Mark Huber [00:15:41]:
And if you think back to making the case for this very creative campaign, how did you have to sell that internally? Because I imagine taking a big swing like that requires many different leaders to sign off on and be comfortable with.
Navneet Singh [00:15:56]:
Yeah. The great news is because our CEO is himself been a marketer and has a great marketing mindset. So we collaborate. I report to the cmo, we collaborate very closely with the CEO and we decide, we do a lot of surveys on where we need to invest in and then we combine that with the creative aspects. And we said, this is how we're going to bring it together. And the CEO is very, very supportive. And we also had these measurements in place that this is how we're going to measure. These are the goals, this is the ROI we are looking for, and it has exceeded our ROI expectations.
Mark Huber [00:16:36]:
You make it sound really easy. Many people don't have the luxury of working for a CEO who comes from a marketing background. So do you think that is a good thing? Does it make it harder? Does your CEO ask even more questions?
Navneet Singh [00:16:51]:
There are definitely instances where because the CEO knows so much, he obviously asks, digs deeper. But at the same time, you also get a lot of support. So something, as you said, taking a big swing like this would probably not have been so easy with somebody else. But at the same time, if I think about somebody who has not been A in marketing. Right. So the way to convince that C Suite would be to really start with smaller tests. So like you said that you saw 250k pipeline growth. I think if you do some smaller tests, show some success at a smaller scale, maybe it's one product rather than the whole portfolio of products, you can build the case incrementally as well.
Mark Huber [00:17:36]:
So we're big on experimentation at user evidence, and this is a recent experiment that we've run. Is there a culture of experimentation, would you say, at Palo Alto Networks for marketing? Is that encouraged?
Navneet Singh [00:17:47]:
Yes, absolutely. I'm also going to talk about what we did for AI. AI is a transformative shift. It really requires us to build a culture of AI and to build that culture, what we did was we had AI summits where all the senior leaders of the company came together every couple months. We invited industry luminaries like people who work on Google Vortex and people who work on IBM. Watson came in, people who had startups working on AI products came in. We learned from them. And then we said we are going to experiment with AI and see what we can do in marketing, what can we do in it? What are the use cases we can solve in sales, what can we solve in finance? And all of those use cases that we experimented with, we came back and the leaders demonstrated that to all of the leaders cross functionally and got feedback, brainstormed, debated, discussed.
Navneet Singh [00:18:47]:
So that's how we got started on our journey on AI.
Mark Huber [00:18:52]:
So we are right in the thick of planning for next year fiscal planning at User Evidence Are there any personal use cases that you are most excited for next year at Palo Alto with Generative AI and how you and your team uses it?
Navneet Singh [00:19:07]:
I would say from the customer perspective we have products that help customers use AI very safely. For example, how do you use ChatGPT but safely while preventing source code from being leaked out? How do you use the other use cases? How do you enable your developers to build custom applications which are AI applications? Third, I mentioned Copilots. All of these three use cases are really interesting and we are really excited to see more and more customer adoption. And on the internal side, we are excited to see AI being used in making launches much more agile by using IT for messaging, by using AI for, let's say, press releases, blog posts and so on. There are other use cases like our IT team uses AI for answering basic questions that employees have so that they can actually focus on higher priority roadmap items. So all of these use cases within the company are really, really exciting I think.
Mark Huber [00:20:12]:
You know, I used ChatGPT quite a bit now. I was very intimidated buy it at first because I spent very little time on my prompts and what I very quickly realized was prompt engineering really is kind of the name of the game when you're using it for personal use in your day to day. And it's only as good as the amount of work that you put into the prompts and then the output increases or the quality of it significantly. So I use it all the time in my day to day and it's something that really has freed me up to do more of the work that I want to be doing and usually find excuses to not find the time to do so. It's been an absolute game changer for Me, especially in a small marketing team.
Navneet Singh [00:20:52]:
Yeah. And what we also see is that I was listening to somebody who was saying every person must behave like a manager now because they have an assistant, even though it's a virtual assistant.
Mark Huber [00:21:03]:
Yep.
Navneet Singh [00:21:04]:
And it requires skills for you to be able to train somebody who's reporting to you. So every person needs to develop those skills in addition to prompt engineering writing, you know, using the frameworks. So it is just a very different world, very interesting world.
Mark Huber [00:21:20]:
One of my favorite prompts is basically to ask ChatGPT to, you know, respond back based on all the different interactions that you've had instead of just what you've been, you know, prompting in that given conversation. And it's amazing some of the breadth of the responses and the depth that some of those responses get into. And I feel like I probably know this much on how to use it right now, but next year, by the time the year's over, I hope I know that much. And then I hope I kind of keep learning and growing the more that I use it.
Navneet Singh [00:21:47]:
I think that's really critical, like using AI. I learned from people I was just talking to somebody, they said that they use. They've created custom prompts, and they've trained it by giving them some of their own writings. And now they say if I have to ask, that's what I've been doing. If I have to ask that AI to that custom model to write for them, in many cases, they can't distinguish if it's been written by them or by AI.
Mark Huber [00:22:15]:
Yeah. One of the things that I've been doing recently is I've been using the. I have a Mac, I use the desktop app, and then basically just have a conversation with it using my own voice. And it speeds up just prompt engineering in general and makes it a little less intimidating because then you don't really have to worry about how do I type this out perfectly so that ChatGPT understands it. You can just have kind of a natural conversation. And it gets me, you know, it's never. Right now, you know, something that I'm comfortable using, you know, 100% of the time. But for some of these smaller tasks that do take up time, it gets me 75, 80% there.
Mark Huber [00:22:52]:
And if I can spend the remaining 25% tweaking it in less than an hour versus doing it a couple hours, that adds up considerably over time. And it's just been a huge time saver for me. It's amazing.
Navneet Singh [00:23:04]:
Right. It is also useful for creativity. So, for example, if you're Launching something new and you're looking for different names.
Mark Huber [00:23:13]:
I do this all the time. Keep going. Yeah, names.
Navneet Singh [00:23:16]:
Or you're launching a campaign and you're searching for taglines. Right. So it just gives you those angles. Oftentimes we talk about biases in AI, but at the same time I was talking to somebody, they said that it actually helps them uncover biases in themselves because they would not be able to come up with some of the taglines, some of the names, because of their cultural background. And using ChatGPT or AI in general, you would be able to get some more creative ideas that you would not think of.
Mark Huber [00:23:47]:
So you mentioned taglines, and that's something that I've used with ChatGPT recently, where I feel like I have some idea that might be good, or there's something there, but it's not fully baked yet. And then I'll throw it into ChatGPT and prompt it a little bit more. And then ChatGPT will return a bunch of different options that are kind of playing off of that same concept. Sometimes do I find the exact tagline or whatever it is that I want to use? Definitely. Other times it'll then return something that I didn't think of, and it'll help me just connect the dots faster. And it's kind of fun to use. And it's something that I just was never really expecting out of a tool like that this early. It's crazy.
Navneet Singh [00:24:24]:
Yeah. I also found another interesting use case where one of the product marketing people that I work with, he had actually used AI to train a model to behave like a cio. So to say that this is your background, this is your education, this is where you have worked, this is xyz, everything about the aspirations, goals set for the next year and everything, and we can use that model for message testing. To say you are a virtual cio, how does this tagline resonate with you? So we don't always have to look for customers or we do customers, and this is additional source of testing for us. Yeah.
Mark Huber [00:25:02]:
One of the things that I saw recently was a prompt. I forget who posted it in their LinkedIn feed. I think it was an AE. And what he suggested doing was using this prompt and then taking call transcripts from any of the calls that you've had with, you know, let's say three to five people on the buying committee. And then in that prompt, you're looking at the role that that particular person is in, and then there's a couple bullets underneath the prompt that shows what they might be Most interested in where kind of they are raising objections, where their concerns are. And again, it's not going to be perfect yet. Maybe it will be perfect in time, but it at least speeds up the process to then inform how you need to go about that next call and really use the right message for the right person in the buying committee in a fraction of the time. It's been amazing.
Navneet Singh [00:25:49]:
I also want to come back to and you talked about skepticism, right? So definitely this does happen in the security industry. It reminds me of a user that we were doing a user study. The interviewer asked the person who was asking for I need this control button, I need this button in the UI and so on. So the person asked, you must like a lot of control. He said yes. The interviewer asked, do you have a stick shift? He said, no. He said, why don't you have a stick shape? You like a lot of control in your life. Said that, well, cars, I trust them, they've been around, I trust them.
Navneet Singh [00:26:26]:
So automatic is okay. So I think it's really about being able to build that trust and that's why there is this movement towards explainable AI, which is really a set of processes, the methods that allow human users to comprehend and trust the results because they are explained how some result was arrived at. And I think as more and more of that happens, I think there will be more trust in AI and very soon it'll be in all walks of life. It'll be part of daily life and we won't question it.
Mark Huber [00:26:55]:
Exactly. Well, Nav, this was great. I'm very excited for your keynote. I can't wait to probably be in maybe the first or second row. I'm excited for it. It's gonna be awesome. Thank you for coming on the proofpoint.
Navneet Singh [00:27:05]:
This has been great. Thank you.
Mark Huber [00:27:06]:
Thanks for listening to the proofpoint. If you like what you heard during this conversation, you probably will like Evidently, my bi weekly newsletter where I share my biggest hits and get honest about my misses as a first time VP of Marketing. You can subscribe using the link in the show notes here. In this episode, the proof point is brought to you by User Evidence. If you want to learn more about how our customer evidence platform can help you build trust and close deals faster, check out userevidence.com