Decision Making in the Age of Al – Keeping Humans at the Helm | RSAC Virtual 2025

So I'm just gonna give a little bit of an introduction because I think right now what's happening, there is no shortage of advisors, experts, consultant, content. It is infiltrating everywhere we go. And I always say that you can't even bump into air right now without reading AI articles left and right.

And so I am honored and privileged to be here. But I also wanted to give you some context as to why on earth I am here. Um, I actually 27 years ago was a professional rugby player.

I had zero intention of going into data, AI, security, none of the above. I thought I was gonna play professional sports my entire life. And then what happens as we age, recovery takes longer.

Um, and I'm sure you guys know a good night out or just a pickup game. Heck, sometimes I even wake up in my age and I hurt myself and I'm like, how did I hurt myself sleeping? It just makes, thank you, thank you.

You know what I'm feeling? Um, I'm a lot older than I look and so I kind of needed to get a real job. And so in 1999, the first job offer I got was a data engineer.

I didn't know how to spell data and I barely knew how to spell engineering 'cause I studied chemical engineering at Berkeley. Um, long story short, about seven months into my gig, my manager pulled me aside and said, so we love you. However, you're never allowed to touch a lick of production code again.

You can hack your way through anything, but you cannot write production code and you are creating more work for us than you are alleviating work for us. And I was like, crap, I'm gonna get fired. Thank you.

Someone else has been there as well. The good news is I didn't get fired. What he told me was, is you understand our world, you understand our language, um, and you like people and we don't like people.

He literally said this word for word, you know, and can build relationships and you understand the business. So your job is to go gather requirements from them, bring it to us and translate it into our language and then we're gonna build the stuff and then you're gonna go back and translate it to their language. So I was writing functional specs and text specs before we even called them functional specs and tech specs.

And that was my world. And so I was this what I called a glorified translator and I was like, well, this kind of sucks. Um, but it actually turned out to be one of my greatest mistakes that turned into a career advantage.

And I've had a lot of fun since then. But something pivotal happened because I was a practitioner. I still am a practitioner, tried and true.

Um, I went from being an IC to managing a small team at IBM to a group to a division, then a p and l. And then how many of you guys watched Watson beat Ken Jennings in 2011? Are any of you guys old enough to even have seen that?

I'm just asking question. I'm 47. So like, I was old enough 'cause I was there.

It was a really pivotal moment because if you guys for good or for bad, I still don't think IBM gets enough credit. But Watson was the first commercial artificial intelligence product that the workforce had ever seen. And the pivotal moment was that televised jeopardy moment when Watson beat Ken Jennings, the world's greatest jeopardy player.

And it was officially the moment of machine beat man IBM said, we're ready to take Watson to market. And so we did. Um, well I was in charge of an overseeing the data practice back then.

I went to my boss's boss's boss and gracefully begged and said, you can't do AI without data. I made that up. I had no idea if it was true or not.

He bought into it. And so I actually helped evolve Watson and launch Watson. And my job for about four or five years was to fly around the world, help enterprises build their AI strategy, identify use cases, and help them actually deploy Watson at scale.

Um, no different than what we started doing in 2023. I would just say I've just had a head start and making all those mistakes while a lot of folks are learning and earning of what it takes to do AI at scale. But since then I've had some remarkable positions.

I've worked for some great employers, um, and we've had fun and got to write a book some patents. It's been an interesting journey. But I just want you to know that I'm not a researcher and I'm not a scientist.

We've invented a few things, but my job has always been, I've been hired to build or establish the capabilities across Fortune one hundreds. And that's what I've been doing. So today, while I won't be able to go into everything, it's gonna be a bit of an aggregation of all the mistakes I made that maybe you can avoid.

If we take a look at the history of data. Now, thank goodness I wasn't around until the 1950s. Uh, I, I was born in the seventies so I can pivot more to the right.

If you take a look at the evolution in the nineties, and I would say early two thousands, does everyone remember Cognos, MicroStrategy building cubes? There's a few head nods. Like that was the world of business intelligence.

How are we gonna join the data sets and the taxonomy models to come together so we can generate insights for businesses? And that was really the only application of data really. No one was really talking about it.

And then we had this wave of like master data management became a thing, then it was cloud. And everyone's like, no, I'm not gonna put my date on the cloud. That's just kind of creepy.

And then here we are today. But if you take a look at the mid two thousands, another pivotal moment hit our lives that has changed us and our careers permanently, social media catalyzed data acceleration, e-commerce catalyzed data creation and acceleration. And if you think about you combining with these notions around personalization, customization at scale, this is kind of what gave birth to our movement.

And in 2016, the role of the chief data officer was being discussed for the first time. We actually didn't even have chief information security officers as a defined C-suite executive. Believe it or not, CDO came before ciso.

But that's kind of what gave birth to everything. And then as you can see in 2023 there became a resurgence. But we all know AI's not new.

It's the commercialization of AI for ebitda, for margins and for growth. That's what's new. And now there's critical mass behind it.

And if you overlay that, and by the way, this is a super, super, super simplified version, but you've all seen the history of ai. There are other commercial pivotal moments and I, I would still say IBM doesn't get credit for taking this to market, but deep blue beat the world's best chess player that was televised Watson beat the world's greatest jeopardy player that was televised. And then in 2022, during that winter, OpenAI released chat GBT.

But if you overlay data on top of the history of ai, take a look at the last branch, everything is happening literally within the past five to eight years. And I think the other added complication, and this is kind of infamous because I had asked the question, um, in the moderation panel, what do you do to keep sane? Because I am struggling.

I feel like I'm underwater with that straw trying to get as much oxygen as possible because the rate of information is just coming at us so much. And I was like, why are we distracted? Why are we losing focus?

Why can't, why doesn't it feel like I have a single day where I feel I'm caught up outside of the duties associated with our families, our friends, our jobs and our careers? And if you take a look at just these stats in the 19 hundreds, information doubled every a hundred years. Alright, that's a long lead time.

And in the mid 19 hundreds, 25 years and now information doubles every 12 hours and that's across the global worldwide web, you name it. That is a lot. And while we as humans are capable of many, many, many things, we definitely are not algorithmic by any means.

And so we have to pick and choose which information we wanna process, which in and of itself gives us a cognitive load. 'cause we are constantly filtering. So the pace of change is really fast.

So for those of you guys and gals, insecurity of sorts, how many of you are not in security? Show of hands. Okay, what's your occupation?

Development. Development sales. Oh, we have a sales person.

Yeah. Great, thank you journalist. Wonderful.

Okay, so we have sales and journalists. Everyone else is somewhat touching or in the security space, congratulations and condolences. You've probably picked one of the most thankless jobs on earth because you're not gonna get the credits for everything you've avoided on an hourly basis, a daily basis and on a minute basis and a second basis.

But you sure as heck are gonna get that nasty gram or that email or be pulled into a meeting with the one thing that does go wrong. But you have to understand that with the way data is evolving, artificial intelligence is evolving. I mean at this from the bottom of my heart, you kind of are the superheroes of the current world, whether you're getting credit or validation for it or not.

Because what we have to do on a day-to-day basis is not easy. Not only is it not easy, but we are a constant threat and attack. So it does take a bit of a calming stabilized approach, if you will, to be able to handle that amount of pressure.

Now, the state of AI today, again, we hear facts and figures all over the place, but my perspective is very global. It's not us. 85 trillion by 2030, that sounds like a big number.

Show of hands, how much of that is required to solve world hunger? If you think 30% or less of that is required to solve world hunger, raise your hands for me please. Okay, less 50% or less.

Okay. Around 70, 75%. All of that to solve world hunger, okay?

It's about 48% of that total investment. We are now choosing investments into this capability over a major humanitarian issue that we've been trying to solve for half a century. Just to showcase the magnitude and the gravity of where we are today.

But even with those investments, and even though you can't bump into air without hearing AI this or AI that, take a look at the countries that have the highest adoption rates. Now take a look at us. We're on par with a country that still has siesta by the way.

Just like to put it in perspective, why? Why are we still struggling with adoption considering we can't turn around without hearing the latest and greatest? And to depress you a little bit more, I'm gonna give some fun facts.

74% of corporate AI initiatives are either stalled or canceled after MVP. Well that's not fun. 80% are failing due to cost data privacy and security.

Now I'm sure you guys aren't a part of that 80%. So don't worry, 88% of POCs don't make it into production. Now I've done over 200 plus deployments in my life and I would say firmly that about 48 of my 200 plus deployments are are enterprise grade and in production.

Not all my children went to college and they are for a variety of reasons. But I'm gonna share you what they all had in common. And if that's still not impacting you, like think about where we are right now.

Low adoption rates great, but in terms of critical mass and and in that adoption curve, did you know 50% of fortune one thousands are still having discussions on what to do with ai? Most haven't activated this yet. 12% have actually been able to push what they started in POC into production.

And then those that make it into production, only 14% of the 12% actually scale. So here's a fun fact. Just 'cause something's in production does not mean it has scaled what determines scale adoption.

We are struggling with adoption at this point in time. And of those 14%, by the way, 11% have to start over. Now I know right now it's not sounding very positive and very happy, but I promise there's good news in all this.

So in our world, we hear about these amazing technology companies and what they're doing. We hear about these startups and their amazing billion dollar valuations and we're like, oh my gosh, what's going on? And meanwhile, in our world, this is what it feels like.

We're just stitching things together, trying to stay afloat, trying to stay above water. All right? We're gonna get into why it is this simple data and workforce.

And they both by the way have something in common. At the end of the day, when you are trying to deploy AI at scale within any enterprise, actually very little has to do with the infrastructure and the technical capabilities, believe it or not. And even with the data stuff, thanks to all your hard earned effort, they can be overcome.

Now granted, it's exhausting 'cause you have to have 11 one-on-one discussions before you get the one group discussion so that you can get alignment to move forward. I know that's our world and it's very, very draining. You just can't move fast enough, but still all solvable.

But what we tend to do and where budget always gets cut, and I know we call it change management, that's kind of a woowoo term because it's not change management. This is funded mentally. Something very different is unrealistic expectations.

Oh, we have bad customer data. Use ai. With all due respect, your sales rep is not gonna write better data into your CRM system that is not gonna correct data entry by doing ai.

It is not a light switch. We turn on two. There is fundamental resistance because how are we going to market with artificial intelligence right now?

It's the new digital workforce. We're gonna increase productivity and capacity, which means we are going to do what decrease teams, be more lean. And you're starting to see a lot of that happen.

Well, why on earth would any knowledge worker document what they know into A PDF to be able to ingest it into a large language model? You're threatening their longevity in their future. Why would they?

So yeah, you can buy a few licenses, a copilot and say you're doing ai if that's your version of ai or if you're embedding it operationally and redefining business processes and workflows and work functions, it's a completely different ball game. And it's unfortunate that we just don't do enough right now to be able to help our workforce prepare. So what do data and the workforce have in common?

I want everyone to think about the word security at its core. Take it out of our work environment. What does it mean to feel secure?

Freedom of fear. Confidence in our safety protection, whether it's with our data or whether it's with jobs in the workforce. Everything's being threatened right now.

We don't feel secure because I think we mentioned before, we are flying the plane while we are building it and whether folks are willing to admit it or not. In those implementations that I did at the end of the day, the ones that struggled to gain adoption, I had to spend a lot of time with a lot of people trying to figure out, Hey, what's going on? Why aren't you using this?

Or, okay, you're using it to summarize notes, but there's more you can do with it. Why aren't you doing it? And it was very much a human to human interaction.

And then I would spend time with the compliance team, the legal team, the security team to understand why so much resistance. And it came down to the very fact that it was threatened. Individual security, family security, occupational security, enterprise security, and all of it has a cost associated with it.

And that's why fundamentally there was just a lot of lag and drag and we weren't able to move forward as quickly as we wanted to. And so as you guys are going through your day to day, I always say there's five dimensions of security that we must fundamentally hold near and dear to our chests. Stability is important to us at all.

Like it's funny saying that especially in today, when the pace of change is faster than it's ever going to be, today is faster than yesterday and tomorrow's gonna be faster than today. Stability, oddly enough, is still really important to our enterprises. The companies we serve, serve to us and our families.

Sovereignty, the ability to have control and make a choice. An agent's not gonna undo what we did or that choice isn't going to be taken from us. Even if you think about, I remember in 2016 when GDPR came out, I called it gush darn pain in my rear, by the way, that was the nicest version acronym that I had for it.

Thank you. You're laughing at my jokes. I really appreciate it.

Thank you. Um, we had a choice. We had to remember the double opt-in, the single opt-in.

Now the default is you are opted in and you have to actively go in and opt out. LinkedIn is using all your comments, likes, posts, and engagements as a default to train. Its LLM model.

You have to go in and actively opt out. You upgrade your phone to iPhone 16, you have Apple intelligence. You can't opt out of that.

I haven't upgraded my phone yet. That sovereignty is being threatened. And then if you take a look at sanctuary, it's about creating safe spaces to a certain degree, but we are also being monitored quite a bit.

And then if you take a look at safety and surveillance, this is all what's prohibiting our ability to actually scale artificial intelligence for the good of humanity because we still fundamentally are having a resistance towards it. And so my ask of you all is, is you have a very, very important job whether you feel it or not, whether you're getting credit for it or not. And you could be stuck in the minutia of the day-to-day, but you have to have self sovereignty that you are fundamentally in a position to shape what's going to be happening in the next year or two.

I don't even think we have two to five years that was mentioned earlier in the moderation. I I do think it's a, it's a game of nine months to maybe a year and a half as an individual and as proxies and advocates for your enterprise surveillance monitor for safety and not invasion, right? That's something that's gonna be fundamentally needed.

Stability in your language and how you're approaching things and encourage your peers, encourage your colleagues. The goal is to be able to protect who we are, our systems, our architecture from outside threats, that's the name of the game in the business that we're in. Sovereignty, give people that choice.

It may be an extra feature, it may be a pain to be able to add that. But it's important because the more we feel that we have a choice and that we have influence, actually oddly enough, the smaller the hurdle of adoption is going to be safety, protecting against vulnerabilities, right? That's kind of what we're all doing on a day-to-day basis.

And for sanctuary, whether it's your home, whether it's your safe space, whether it's within a company, at the end of the day, what everyone is craving for and what everyone feels apprehensive about when it comes to artificial intelligence is, is there going to be a place of safety moving forward? Nest cameras, appliances, roombas, phones, but you all are all, you're all being hired to be able to develop what I call that creepy zone. How much is too much where it's borderline creepy and how much is enough?

Because that's what's gonna protect your ability from being threatened or attacked in a number of ways. So we're wrapping up, but as security and cyber specialists, and I am totally gonna geek out because I'm a marblehead, you are the new age guardians, like without a doubt. And you are being stretched and bent beyond belief because you not only have to bring credibility to the corporations that you serve, the startups that you represent, the institutions and entities and universities that you advocate for, but also thought leadership.

This isn't just about security systems and architecture, any of these capabilities, if they are meant to scale, it's fundamentally ring fencing and being guardians of safety. And everyone is kind of on edge when it comes to this stuff. It really, really is a humanitarian problem.

And so going back to your question around the certifications, listen, I, who knows what's gonna happen tomorrow and who knows what's gonna happen in two to three months, but if I were reading the tea leaves and if I had to put a bet, there's more to this. But these are the six biggest takeaways I would say are really important right now. But the last one, oddly enough in my past and prior experience is what prepared us the most.

I deployed this several, several, several years ago after making mistakes. 'cause we had great documentation, we had great policies, we bought the latest and greatest in tools. It's not until your call to action that you actually know how all these things come together.

So we pivoted from one-time training courses, online portals. I would actually run real-time simulations of cyber attacks and hijacks. And I would never tell my team when I was doing it, it was, I knew when it was happening.

And then I would inject it and I would see how the team responded. Could we leverage all the latest and greatest tools that we were, that we had bought? 'cause we thought it was gonna help us.

Could we detect in real time? Could we respond in real time? Could we send an email off to our stakeholders and executives about what's happening, why it's happening and what we are doing?

It was not about our documentation, it was not about our policies. It was not about that video of 45 minutes of just, ugh, I gotta do this again. It's end of year.

It was actually putting the team to test to see how we even operated and connected together. So just, if you guys can do this a few times a year, you'll be solid. And then last but not least, I do think anyone who's in this space, you guys are the new super superheroes.

It's a tough job. There's no doubt about it. If you think about AI scaling within enterprises, it really does come down to data and the workforce and both of them are hooked on the notion of security and safety and how much is at risk right now we're still in the infancy stages of adoption.

So for those of you guys who may have FOMO or feel like you're falling behind, here's a little secret everyone. I shared some stats. If that doesn't make you feel better, at least know that everyone is somewhat behind.

There's just too much information to be able to ingest. And then more than ever, I would rethink this career choice because it's not just a job anymore, it just isn't. It may feel like it and I know some days are worse than others.

Um, but you have no, I don't know if you can comprehend how much of an influence you have on how this stuff shapes for the future. Alright, last but not least, if you guys have any questions, my email is there. Let me know.

Um, I appreciate this time. Thank you so much.