Skip to the stories
Networking

Enterprise Resilience in the AI Era: Patching, Recovery and Vendor Risk

  • Kelsey Ziser
  • 345 views

CIOs and CISOs are racing AI-driven threats while managing recovery, quantum risk, vendor changes and network resilience. Here's what matters now.

Enterprise Resilience in the AI Era: Patching, Recovery and Vendor Risk

AI has changed the pace of enterprise risk. Attackers use it to find and exploit weaknesses faster, and IT leaders are being asked to keep operations running while they modernize infrastructure and tighten security. Those two goals pull against each other. Here's the thing, though: the leaders who are handling it best aren't treating resilience as a separate project. They're folding threat defense into everyday operations, data management and incident response.

Patching has to match the speed of attackers

The clearest illustration comes from patch management. Bad actors are using AI to shorten the window between a vulnerability's exposure and its exploitation. CISOs and CIOs are answering with the same tools, using AI to find, rank and fix weaknesses before attackers do. Wally Dalrymple, CISO at academic testing companies Educational Testing Service and PSI, put the goal bluntly: "We have to be moving as fast as AI." He said his team is now speeding up remediation because it can patch much faster.

My read is that this shifts the bottleneck. Finding a flaw is getting cheaper. Getting the fix deployed across a messy estate is where teams still lose time, and that's where the operational investment needs to go.

Recovery is more than systems coming back online

Restoring systems and restoring the business are two different milestones, and the gap between them can be expensive. Boston Scientific told investors in September that the fallout from a cyberattack would likely have a material impact on its third-quarter and full-year results, even though it had successfully applied the fix. That's the lesson in one example: a technical recovery can be complete while the financial and operational damage keeps rolling.

For CIOs and CISOs, that means setting recovery priorities in business terms and giving the rest of the C-suite a realistic picture of operational and financial fallout. Teams that do this early tend to avoid the argument later about why the business still looks broken after IT declared victory.

Quantum readiness is a stakeholder problem too

Post-quantum cryptography is the long-horizon item on many roadmaps. The concern is that current protocols could theoretically be broken by sufficiently capable quantum systems. Executives from Mimecast, Everpure and other organizations have been discussing how CISOs can reassure fellow C-suite members and stakeholders that their enterprises are quantum-ready. The conversation is less about a specific deadline and more about showing that the planning is underway and credible.

When the patch depends on someone else

A data breach at the FBI highlighted a problem many enterprises already face. When security depends on third-party vendors, identifying a serious problem doesn't mean you can fix it. Platform ownership is complex, and risks are being exploited quickly. Being able to see the threat without being able to remediate it is a hard position, and it's one that vendor contracts and escalation paths should account for before an incident, not during one.

Vendor ownership changes are a risk category of their own

Airtable's post-acquisition layoffs offer a useful case study. After Bending Spoons acquired Airtable, the Italian technology conglomerate said it would lay off a "large part" of the company's pre-acquisition workforce. That came after promises to invest heavily in the relational database management platform's product and customer support. The story isn't unique, but it raises a practical question: when a provider changes hands, how do enterprise customers work out the new terrain and make sure their needs still get met?

That's a due diligence exercise that doesn't end at signing. Teams in this position typically reassess support commitments, data portability and roadmap dependencies once ownership shifts, because the promises made during a sale don't always survive the first restructuring.

Network resilience and the role of AI

Network resilience has its own version of the problem. Hurricane Isaias, which made landfall in Florida in October, can damage the electrical grid and cell sites, disrupting business and network operations and making communication harder for first responders. John Saw, CTO at T-Mobile, explained how the operator provides temporary network coverage and restores downed cell sites during natural disasters. T-Mobile is also using AI to forecast demand and automate network adjustments during periods of heavy traffic.

How far AI can take network management is still open. Some network experts expect AI to take on more of the operational work with less human intervention. Whether that makes the network itself smarter, or simply makes the management tools smarter, is debatable. That distinction matters when you're deciding how much authority to hand over and how to audit what the automation does.

AI APIs widen the attack surface

The rise of AI APIs brings its own security questions. Pieter VanIperen, CISO at AlphaSense, has looked at how to close the new exposure points and whether reducing privileges for AI APIs can help prevent certain types of exposure. The principle of least privilege isn't new, but AI integrations are often added quickly and sit outside the review processes that govern older systems, which makes them easy to overlook.

Balancing AI value against risk

AI spending keeps climbing, and demonstrating business outcomes tied to those budgets remains difficult. Many companies are moving ahead without guardrails that would limit their exposure. Gartner has suggested that CIOs build a flywheel approach, where AI investment and AI risk are treated as parts of the same system that pushes value forward, rather than as competing line items.

Where to focus next

The common thread across these stories is that resilience now depends on speed, ownership clarity and realistic communication. Patching has to keep pace with AI-accelerated exploitation. Recovery plans have to cover the business, not just the servers. Vendor relationships need contingency plans for ownership changes. If you only take one action this quarter, review your privileged access for AI integrations and confirm who owns the remediation path for your most critical third-party platforms. Those two checks will tell you more about your actual resilience than another tool evaluation will.

Source: Kelsey Ziser · www.informationweek.com

Neighbouring cells