When Farmers Meet Firewalls

Tony Taylor talks about AI the way a competitive athlete talks about training: pragmatic, relentlessly focused on practice, and aimed at giving people the tools they need to win. As Chief Information Security Officer at Land O’Lakes, a member-owned agricultural cooperative that touches 50 percent of the harvested acres in the U. S., Tony blends three decades of infrastructure experience with a clear view of how cloud, SaaS, and generative tools are reshaping risk and defense across an unusually wide business footprint.

Tony walks through where he sees real AI value for defenders today, from helping junior SOC analysts make decisions to automating repetitive work that nobody wants, and where he’s urging caution. The result is a pragmatic playbook for security leaders who must protect both physical supply chains and cloud-native services.

Security at Agricultural Scale

Land O’Lakes operates far beyond the grocery shelf, and Tony uses that breadth to frame the scale of his security challenge: “But, we touch 50 percent of the harvested acres in the U.S.… we process about 12.5 billion pounds of milk on an annual basis and we feed 100 million animals every day.”

That operational scale includes not just physical infrastructure and logistics, but also cloud and SaaS environments that have radically changed the attack surface. Tony described how the move to public cloud means previously internal misconfigurations are now public exposures in seconds: 

“Before we could make a simple mistake inside of our network… it wasn’t on the Internet. And now everything’s on the Internet… within seconds you’re going to be known that you’ve got an opening and you’re going to have data exposed. That’s probably the biggest concern I have.”

He also emphasized that the shift to cloud and SaaS requires broad education across teams, because not all administrators have experience with public cloud tooling and continuous updates.

AI in the Hands of Attackers

When the conversation turned to AI’s impact on threats, Tony was frank about what he’s already seeing in the wild: “There are no longer bad grammar errors in the emails… I don’t think I’ve seen a bad form of phishing email recently.”

That change has contributed to exponential growth in phishing and other attacks, because generative tools allow individuals with minimal expertise to launch sophisticated campaigns: “You don’t even have to be a professional in this business anymore to launch an attack with AI and ransomware as a service… Anybody sitting at their home can go do it.”

In Tony’s framing, criminals don’t need advanced skills anymore. They only need organization, funding, and off-the-shelf tooling, which makes the threat landscape larger and more accessible than ever.

AI as a Force Multiplier

While Tony warned about attackers tapping AI, he’s optimistic about the role AI plays for defenders, especially for expanding the capacity of security teams and accelerating the development of junior talent. With a security team of about 22 people supported by an MSSP, Tony noted that:

“We bring in our younger people… they have the academic background, but they don’t have the real-world background… we partner those folks up with engineers… and they eventually go through our engineering groups.”

Here’s where AI becomes an accelerator: instead of expecting new analysts to learn solely by doing or pair programming, AI can help answer contextual questions, reinforce playbooks, and even enhance existing playbooks where none exist: “…if I don’t already have a playbook for it, it should even add to and enhance my own playbook.”

This not only lessens the burden on senior engineers but also helps newer team members build experience faster, addressing what Tony sees as a chronic talent shortage: “We need to be able to bring in less mature or less experienced people and let them build that experience because the workforce isn’t there… I can’t go hire senior engineers all over the place because they’re not available.”

AI Potential in the SOC

Tony also shared his thinking about how AI might shape Security Operations Centers in the years ahead. For him, the key variable is trust: knowing what an AI engine can do and when it can be relied on. “I think at the beginning… it takes the rudimentary stuff and the low-hanging fruit… and… you don’t even have to see it. Let those people work on higher threat hunting, more value-added activity…”

Yet he also foresees a future where AI can present recommendations with confidence scores, letting humans approve or tweak actions based on environment context. “Maybe it does 80 percent of the work and says, here’s what I found… and we look at it… Yeah, go ahead and do that… or no… that’s not right.”

He sees this as an evolution toward customizable, adaptive AI, not black boxes, but partner tools that reflect the enterprise’s operational realities.

AI Wins and Skepticism

One of the clearest examples Tony shared of security automation paying off came from an investment in phishing prevention tied to financial processes: instead of relying on training alone, his team implemented controls that stop the malicious messages before they ever hit employee inboxes.

Over five years, that investment prevented recurring losses. “That was a really good investment… saving us X amount of dollars a year because for the last five years we saw that money go out the door…” 

At the same time, Tony expressed healthy skepticism about overblown AI marketing claims. He wants tangible, demonstrable value before enabling or buying. “You can’t talk to a vendor today where they don’t talk about their AI… Everybody’s got an AI engine. For me… I go, well, show me how that AI works.”

And from a privacy standpoint, he stressed that sensitive data must stay inside enterprise boundaries, not be exposed to public models: “if you’re going to take my data… that’s got to stay in my domain.”

Adapting With the Business

Part of Tony’s leadership philosophy as CISO is not to be draconian or a gatekeeper, but a partner. He credits his long tenure (20+ years at Land O’Lakes) with helping him build relationships so people include security early in innovation discussions: “I know I’ve made an impact… when more people come to me, versus me trying to hunt people down…”

He emphasized that modern security must be visible, business-oriented, and collaborative,  not locked away in an IT back room.

The Workforce Shift

Tony’s view on AI’s impact on the cybersecurity workforce is clear and optimistic: AI will reduce workload, especially repetitive tasks nobody wants to do, enabling teams to focus on more strategic work. “It’s going to allow people to work at a higher level… and it’s going to help them educate and teach them.”

He believes this makes cybersecurity a more attractive field, not one hollowed out by automation. He also anticipates new roles and skills emerging, especially around data analytics and AI-enabled environments, where humans will guide models and curate data, ensuring outputs remain trustworthy and contextually relevant.

Leadership Lessons

Tony distilled several practical insights for security leaders:

• Don’t underestimate the amount of time you need to spend with the business. Build relationships up front.
• If you’re a CISO and you’ve got your hands on the keyboard, you’re doing the wrong things. Focus on people and partnerships.
• To stay current, consume information broadly: read widely and listen to daily briefings like SANS Stormcast and CyberWire.
• On the basics: “if somebody doesn’t have MFA… shame on you.”
• On the fear that AI will replace people: “I don’t think it’s going to be the great displacer of people… it will provide opportunities.”

AI as an Amplifier

Tony Taylor’s vision of AI in cybersecurity is neither alarmist nor starry-eyed. It’s practical, measured, and grounded in enterprise reality: AI helps defenders manage noise, accelerate learning, and scale people without replacing them. It elevates the work they do and strengthens collaboration between security and business. And like any powerful tool, it requires demonstrable value, clear guardrails, and contextual trust before being adopted widely.

As Tony reflects on the competition of cybersecurity, where defenders must get it right 100 percent of the time, AI does not change the stakes, but it changes the odds.