Palo Alto Networks bets on ‘precision AI’

Palo Alto Networks is betting on “precision AI” – a combination of machine learning, neural networks and generative artificial intelligence (GenAI) – to ensure 100% accuracy in security decisions for customers.

The company’s regional vice-president for ASEAN, Steven Scheurmann, said this high bar is crucial in the cyber security industry, where a single mistake can lead to breaches. “We can have no line of error in anything we do, because if we make a mistake one time, someone could compromise the endpoint, exfiltrate data, pick up your personal information or data, and compromise your payment system,” he said.

To achieve this level of precision, Palo Alto Networks has invested heavily in data collection and analysis, ingesting 36 billion events and 7.6 petabytes of data per day from the company’s next-generation firewalls, cloud security and endpoint protection offerings.

But it’s not just the volume of data that sets the company apart. Scheurmann emphasised the importance of the data’s context, noting that Palo Alto Networks collects data from a wide range of customers across various industries, including public sector, banking, telecommunications and retail.

“The contextual piece is very important, and being in an organisation that only lives and breathes cyber, we pretty much know the context,” said Scheurmann. “All those pieces put us in a unique position to make very good decisions on the data set.”

Unlike some cyber security vendors that offer payouts or guarantees if their products fail to prevent attacks or aid in recovery, Palo Alto Networks does not. Instead, it focuses on proof of value through pilot projects, allowing customers to achieve the outcomes they want.

“And don’t forget, it doesn’t matter how good your system is,” he added. “If you look at some of the incidents that have happened, it’s not because of the tech, but because processes are not implemented, or the processes are broken. There’s a lot of human error that happens above and beyond the technology.”

Indeed, one of the biggest challenges in the rapidly digitising ASEAN region is not the technology itself, but the ability to properly implement, maintain and operate security products, said Scheurmann.

“The problem is a lot of people have never had to run and operate security for applications in the cloud and services at the edge – this is an absolute new paradigm for them,” he added.

To address this, Palo Alto Networks takes a hands-on approach, actively engaging with customers in testing, training and operationalising its products. “For example, I was just in the Philippines with a large government agency where we were involved in all the testing, training, use cases and playbooks, because unless they operationalise and use it, it’s a waste of time,” said Scheurmann.

On security consolidation, he argued that the risks of fragmentation outweigh the potential drawbacks of a platform approach. “Every security incident now is due to too many products not talking to each other and the inability to correlate, run and support multiple products – and hackers take advantage of all of that,” he said.

“Are you willing to take that risk? Or try to consolidate to have a better view, using generative AI, machine learning and neural networks to make the right decisions? This is a far more important outcome than worrying about the risks of consolidation.”

That said, Scheurmann is not advocating for customers to adopt all of Palo Alto Networks’ platform offerings, recognising that they may still have a need for some point products.

“It would be extremely arrogant for me to think that a customer is going to be 100% on Palo Alto Networks,” he said. “It would be nice, but the fact of the matter is consolidation, for some of our customers, means reducing the number of products from 37 to 10, not one.”