Monday, November 16, 2026

AI Is No Longer Just a Technology Strategy

AI is a Trust Strategy.

Walking away from the Gartner conference, one message stood out: AI become part of how organizations operate, make decisions, serve customers and manage risk.

 

For the past several years, much of the conversation has centered on productivity. AI making employees perform tasks faster, automate repetitive work, analyze information, generate content or improve customer experiences. While those benefits still matter, the more consequential question is what happens when AI begins acting, not just assisting.

The shift toward agentic AI changes the equation. As systems move from generating recommendations to orchestrating workflows, interacting with other systems and executing actions, the traditional boundaries between technology, business processes, security, privacy and governance begin to blur. This message that was reinforced by the day’s keynote.

 

The next phase of AI is about agency. An assistant can help someone make a decision; an agent may participate in the decision, initiate another process, interact with systems or trigger an action. As such, the risk profile changes accordingly. The question is no longer simply whether an AI model is secure, but leaders must understand:

·      What information can it access?

·      What decisions can it make?

·      Which systems can it interact with?

·      What actions can it execute and within what boundaries?

·      What happens when multiple agents interact?

·      Who remains accountable for the outcome?

·      Can we reconstruct what happened, including why an action was permitted?

 

These are enterprise risk and leadership considerations. Security must evolve with the operating model. Traditional cybersecurity has largely assumed that people initiate actions and technology executes them. Agentic environments increasingly challenge that assumption that makes identity, data governance and authorization foundational. Permissions need to be more granular. Monitoring must increasingly account for intent and behavior not just network activity. Additionally, logging must provide enough evidence to reconstruct not only what happened but how the system reached a decision and why it was allowed to act.

 

A particularly useful point from the keynote was that becoming an agentic organization is not about deploying more AI, it is about deciding where AI can act and under what level of control. A key takeaways described a spectrum from human-led and AI-assisted work through AI-enabled, AI-led and selectively autonomous operations. That is an important distinction since not every workflow should receive the same degree of autonomy. The level of delegation should reflect the nature of the decision, the potential impact and the organization’s ability to govern it. 

 

The principle thar resonated most is that the greater the autonomy, the greater the need for accountability.Strong governance can enable faster adoption because leaders understand where AI can operate, where human oversight is required and which actions should remain off limits.

Jasleen Kaur Sindhu’ keynote, “Harnessing Agentic AI for Organizational Impact,” emphasized G.U.I.D.E. framework as a practical way to think about embedding that governance into the operating model:

·      Govern enterprise context — establish the approved data, knowledge, policies and boundaries an agent can use.

·      Unify decision orchestration — coordinate decisions and workflow routing across agents, people and systems.

·      Invoke approved services — limit access to authorized tools, services and integrations.

·      Deliver controlled actions — apply human approval or guardrail-based execution according to risk.

·      Evidence every outcome — maintain traceable records that support review, accountability and continuous improvement.

 

The framework reinforces an essential point that AI governance cannot exist only in a policy document. It must be part of the architecture and the workflow itself. A practice through identity, permissions, data controls, monitoring, human checkpoints and auditable decision trails. A maturity progression shifts from scoped context and human-approved actions to adaptive routing, guardrail-based execution and continuously evaluated evidence as autonomy increases. 

This also changes how organizations should approach AI initiatives. Rather than approving isolated agent use cases one at a time, leaders should start with enterprise workflows. Considerations around what work is being delegated, where humans retain decision authority, which agents and systems participate, and what controls must follow the workflow end to end. The keynote’s central recommendation is to design for the enterprise, not for disconnected AI experiments. 

 

The traditional build-versus-buy discussion focused on features, cost, integration and capability. With agentic AI, organizations must also ask where context lives, who controls orchestration, where decision logic is encoded and how much control the organization retains. Vendors can accelerate capabilities, but the organization should retain ownership of its context, policies and decision logic. Hence, build versus buy is increasingly an intelligence-control decision, not simply a technology procurement decision. 

 

The goal should not be to create AI that can do everything. It should be to create AI that clearly understands what it can do, what it cannot do and when a person must be involved. The competitive advantage will not come from adoption alone but instead from trusted adoption. Customers must trust how their information is used. Leaders must know where autonomous actions are occurring while employees must understand when AI is influencing decisions. Boards must understand the resulting enterprise risk while the organizations must be able to demonstrate that these systems are governed responsibly. Security leaders play more important role by not simply protecting AI from threats but helping the organization use AI in ways that are secure, responsible, explainable and worthy of trust.

 

Cybersecurity cannot sit outside the AI strategy, waiting to evaluate what the business builds. Security, privacy, technology, legal, risk and business leaders need to design the rules of engagement together. The keynote made clear that the challenge is not just enabling more capable AI but also establishing the controls, accountability and evidence that make greater autonomy sustainable.

Monday, October 26, 2026

Your Perimeter Is a Myth: What I Spoke To and What Was Heard

GDS presentation and discussion on “Managing Third-Party Risk in an Interconnected Digital Ecosystem.”

Third-party risk is no longer a security problem we hand to vendors. It is an enterprise risk we manage through governance, contracts and relationships. The organizations getting it right start inside their own walls.

That was the case I made in my recent session. Every vendor, platform and subcontractor that touches the organization’s data extends an attack surface we do not own and cannot fully see. The firewall no longer marks where our responsibility ends.

Unpacking the 3 part series from my presentation in part 1, what the room added in part 2, and where peers reinforced the position, where they extended it, and where emphasis wrangled.

 

Part 1: What I Spoke To

 

1. Make Third-Party Risk Management (TPRM) an enterprise program, not a security function

Third-party risk belongs in enterprise risk management. That starts with a clear risk appetite statement, tied to security maturity and a deliberate path toward that target state.

It also means a unified process not a disconnected flow. Procurement, legal, privacy, risk and insurance, IT, and security must work together, with stakeholders defined by what data is shared or accessed. Procurement should route through security and other stakeholders from the start and not at signature.

 

2. Get your own house in order first

You cannot assess others successfully until you know what you have. That means an inventory of data, flows, and ownership, up and down the stack. Strong internal data governance and asset management are the fundamentals to making external assessments meaningful.

The biggest gap I hear across the industry is visibility into third as well as fourth parties. We don't know what we don't know.

 

3. Tiering matters but execution is equally essential

No tiering model is universal. Whatever you choose must be applied consistently and weighted by data sensitivity and business criticality.

Let risk appetite set the monitoring model: continuous for critical vendors, point-in-time for the rest, and invest in detection as much as prevention.

 

4. The contract is your strongest initial control

Standard templates should cover transparency, incident notification, breach protocols with real playbook detail, fourth-party disclosure, and best if formulated with credits and penalties.

Build in triggers for material change, such as a vendor acquisition or a new initiative. Given the pace of AI, favor shorter terms and require disclosure of AI use and modeling, including by subcontractors.

 

5. The new perimeter is relationship and trust

Tools are only as good as the people and partnerships behind them. Still, tools, including AI, improve visibility, guardrails and protection, and we should use them.

A vendor's culture and how it operates tell you as much as its questionnaire does. Partnerships that align with matching culture establishes the best relationships. 

 

6. AI is multiplying dependencies

We are moving beyond traditional vendors into AI-enabled services, external models, plugins, and autonomous agents. Vendor-embedded AI can introduce new data flows. Agents can hold identities, permissions and access to enterprise systems. Model providers and subprocessors create downstream dependencies that annual reviews may miss.

We cannot govern what we cannot see. But visibility alone is not enough because we must also challenge what our assessments actually prove.

 

7. Design for failure, not just prevention

We cannot eliminate every third-party failure but we can limit how far it spreads and how long it affects the business. That requires defined isolation and containment procedures, tested recovery and continuity arrangements, clear incident decision rights, and exit plans for single-provider dependencies.

Resilience is proven through tested capability, not documented intent. This is where third-party risk becomes an enterprise leadership responsibility.

 

What's holding us back

Limited team size, time/speed and budget. Competing priorities and limited visibility add to the dependency and complexity. One practical lever often overlooked is reducing the data you share. Less data shared means less exposure to manage.

Emerging pressures

AI guardrails, vendor M&A, cyber insurance requirements and the limits of cloud isolation are all reshaping the landscape. Gateways can help segment higher-risk connections.

 

Part 2: Discussion Portion - What Was Heard

The room largely supported the position. Peers added operational texture, extended the governance model and offered one meaningful difference in emphasis.

Where the audience reinforced the position

•      Asset management is foundational. Peers stressed that the attack surface changes continuously. Asset management cannot be a one-time inventory but instead kept pace of change.

•      Detection and protection both matter. The audience echoed that both must be emphasized, not one at the expense of the other.

•      Resources set the ceiling. Peers noted that team size and investment are directly proportional to the intensity and comprehensiveness of assessments. They also reported rising demand in both the volume and complexity of assessments, with time pressure compounding the strain.

Where the audience extended the position

•      Accountability at the top. One model shared had the risk department leading third-party engagement and accountability, with the CEO and other executives signing to engage third parties and signing off on all accepted risks. This puts real teeth behind anchoring TPRM in ERM.

•      A clarification on ownership. I advocated the partnership with procurement through security and key stakeholders from the start, places ownership with risk decision making. These are complementary, not competing: security informs early, Risk owns the process, and executives accept the residual risk.

Where emphasis differed

•      Insurance as a lever, not just a pressure. I framed insurance as an emerging pressure, driven by carrier requirements. The audience framed it as a strategic tool for risk transfer, one that should be a strong position, with coverage levels and limits adjusted as vendor exposure changes. Both views hold. The practical takeaway is to keep insurance top of the TPRM process so coverage tracks the actual third-party risk profile.

 

Part 3: In Closing - The Call to Action

Start internally by defining and socializing your organization’s enterprise risk appetite. Know your data, the flow, your vendors and the AI they bring with them. Put your expectations in the contract, keep insurance aligned with the risk you actually carry, and test how you would respond when a critical provider fails.

 

The room confirmed what I see across the industry: the challenge is less about knowing what to do and more about the capacity to do it consistently. That makes governance, executive accountability, data minimization, and rehearsed resilience even more important.

Vendors will meet the standard you set but only if you set it. Afterall, it’s about partnership and relationships that move us all forward.

 

Thank you GDS, my co-host and everyone who all attended and contributed thoughtful perspectives to the discussion. Your engagement and insights were greatly appreciated. 

Monday, October 19, 2026

Protect Enterprise Value in Accelerating Digital World

Trust at Machine Speed: The New Mandate for Cybersecurity Leadership

The sessions attended covered very different threats and technologies, yet they converged on the CISO panel's message: cyber risk becomes business risk almost immediately and the next frontier is maintaining trust at machine speed.

 

The adversary runs on a different clock

One presentation cited an average eCrime breakout time of 29 minutes and a fastest observed breakout of 27 seconds. Another example showed initial access reaching data exfiltration in 4 minutes, with access handed to a second actor in 22 seconds. In translations, ticket-driven response cannot keep pace.

Machines must collect, correlate and enrich signals at machine speed so people can focus on consequential decisions. Automation should not remove human accountability but instead preserve the limited time in which human judgment still matters.

 

Fundamentals are the floor, not the finish line

With the focus on emerging technology, a theme that kept returning is back to basics: phishing-resistant MFA, eliminating legacy authentication and standing privilege, complete asset and software inventory, least privilege, segmentation and Zero Trust. Vulnerability Threat Management must prioritize the exposed, unpatched and privileged first. Inspecting and addressing just critical- and high-risk vulnerabilities now need to extend to moderate and low-level vulnerabilities. Innovation is essential but automation does not replace fundamentals alone. Otherwise, the cost of getting them wrong is elevated.

 

From alerts to context

Attacks are not bound by product boundaries. The real threat may be spread across identity, endpoint, network, SaaS, cloud and threat intelligence, with no single console seeing the full chain. The value of XDR, SOAR and AI-assisted investigation is not another dashboard. It is the ability to ask one question across every feed: what else happened to this identity or this asset? The goal is a coherent attack story not a larger alert queue.

 

Identity transforms to intent

Agentic AI changes the problem and dynamics. A copilot recommends an action while an agent executes one through enterprise tools and APIs. Identity must now establish who requested the action, which agent acted, what task was approved, what authority was used and what changed.

The strongest design pattern presented was to give each agent its own identity, restrict authority to the approved task, authorize every hop, use short-lived identity-bound tokens and keep one linked audit record. APIs are the tentacles of AI. The right question is not, is this agent trusted but is this specific action permitted for the specific task. Once agents act, observability must become accountability. The SOC (Security Operations Center) should be able to determine what happened, who or what caused it, why it was authorized, what data was touched and how to stop it? Without that evidence, we have automation without control.

 

Shift in incident response to decision integrity 

Deepfakes, synthetic voice and fabricated documents target the information leaders use to decide. An attacker does not need universal belief; confusion or hesitation may be enough. Threat actors exploit delays in decision-making, advance planning, testing and stressing decisive action essential. Resilience must include authenticated executive channels, rapid corroboration, coordinated response across Security, Legal, Communications and business line leadership, and tabletop exercises that introduce synthetic evidence. The exercise question is no longer only considering if technology recovery is feasible but can the organization establish what is authentic fast enough to make the right decision.

 

From policies to provable control

Regulatory sessions span automated decision-making, privacy, chatbots, deepfakes and human safety. The direction was consistent that disclosure alone will not be enough. Organizations will need evidence that AI activity is authorized, attributable, observable and controllable. Because technology moves faster than regulation, durable governance should be built around recurring obligations rather than one law at a time. NIST's map, measure manage and govern structure offers a practical bridge from policy to operations. This is the same principle I presented on evidence-of-effectiveness standard that applies to AI.

 

Multiple time horizons

Post-quantum cryptography sessions referenced emerging NIST standards and urged preparation based on data sensitivity, system criticality and migration feasibility. The leadership principle applies equally to AI. Not every material risk must be solved today but every material risk should be understood before it becomes urgent. The level of preparation with Y2K must be applied to PQC (Post quantum computing) based on the timelines we’re not in control, leading to 2030 predictions.

 

Outside the session rooms, conversations with fellow security leaders kept circling back to three practical concerns.

1.        Tool sprawl is not strategy. The crowded market creates its own risk. Peers are pressure-testing whether a capability is genuinely differentiated or merely a wrapper, whether security is embedded or bolted on. Additional considerations pertain to how it fits the operating model and whether pricing is predictable. Buying AI is not a strategy. An investment should measurably improve the ability to see, correlate, decide and act. If it adds another silo or console, it may increase complexity faster than it reduces risk.

2.        Architecture must support change. A case study on a rapid SIEM replacement sparked discussion. The lesson was not that every migration should follow the same timeline but that platforms must allow change, playbook continuity and better enrichment without taking years to deliver value. Familiarity to market leaders is essential however, newer or resounding approach such as cribl.io and vega.io of a fresh perspective. 

3.        Security does not own every domain but it must connect them. Cybersecurity, Technology, Legal, Privacy, Communications, Compliance, Procurement and Risk need a connected operating model since modern technology no longer respects organizational boundaries. That is the shared risk ownership the required and emphasis I mentioned playing out in practice.

 

The executive takeaway is trust at machine speed

InfoSec World 2026 traced a clear focus for security leaders:

  • Alerts to context
  • Identity to intent
  • Policies to provable control
  • Incident response to decision integrity
  • Cybersecurity to digital trust

 

The CISO role continues to evolve and expectations are rising. We must be technically credible, commercially aware, legally informed and capable of influencing decisions at the highest levels of the organization. But this evolution cannot rest on the CISO alone. Organizations must align accountability with authority. Boards must demand transparency rather than reassurance, and we must deliberately develop the next generation of security leaders.

The brakes on a race car are not there to make it go slower. They are there so it can take the curves at speed. Security's job is not to slow the business down but instead to create the trust and control that let the business move faster with confidence. Ultimately, the measure of a CISO is not how many tools we deploy or how many risks we identify. It is how effectively we help the organization make informed decisions, manage uncertainty, protect what matters and pursue its objectives with confidence.

 

This perspective draws on his InfoSec World tracks and my own CISO session.

Thank you to the event partnership and building strong relationship: CyberRiskAlliance.com, Snyk.com, myriad360.com, adaptivesecurity.com, onspring.com, gritsolutions.com

Friday, October 16, 2026

From the CISO Chair: InfoSec World 2026

The definition of cybersecurity is getting bigger while the time available to respond is getting smaller. A CISO viewpoint and what InfoSec World 2026 made clear about the road ahead.

 

Joining my fellow CISOs on stage at InfoSec World, the conversation had already moved past controls. Nobody asked how many tools we run or how many vulnerabilities we closed. The questions were about liability, accountability, board confidence and the kind of leaders we are shaping. That shift framed everything else I heard that week. Across sessions on ransomware, AI agents, deepfakes, regulation and post-quantum readiness, the same executive question kept surfacing: can organizations preserve trust and control while everything around it moves faster?

 

My answer and the theme of this piece, is that it can but only if security leaders do two things at once. We must promote the business and its brand and we must protect the enterprise. Those are not competing mandates but the job in one.

 

The central message was simple: the CISO conversation has shifted from controls to risk. Credibility with executive leadership now rests on two things together, evidence of effectiveness and confidence of resilience. Below are the positions I shared on stage and expanded to illustrate that accountability without authority is exposure. CISOs face growing personal liability for governance, disclosure and reporting, including decisions made both before and after a breach. The lessons are already public. Uber incident taught us not to conceal. SolarWinds taught us that internal control and external statements must match. DORA and government mandates are tightening disclosure rules further.

 

The answer is not to avoid accountability. It is to establish clear decision rights, defensible governance and an environment where material risks surface and get addressed. That requires alignment across people (who does what), process (proof and audit) and technology (visibility, access and recourse).

 

CISOs need the access, resources, governance and escalation paths required to meet their responsibilities. We cannot eliminate risk alone however, we can minimize avoidable surprises through transparency, disciplined escalation and defensible decision-making.

 

Boards need risk intelligence, not a technical inventory nor more technology experts. They need advisors who translate technical exposure into business impact, options, recommendations and clear ownership. The question is always why it matters.

A board presentation succeeds when it improves the quality of a decision, not when it shows how much the security team knows. And boards should demand transparency over reassurance.

 

Security is a leadership responsibility across the enterprise, not an obligation delegated entirely to the CISO. Risk ownership must be shared. A gap between business strategy and the execution team will show up in security. Peers must be partners, working vertically and horizontally against the same goals.

The most influential CISOs follow a consistent discipline:

  • Start with the business objective
  • Frame risk as business consequences
  • Present options, not objections
  • Build relationships before you need a decision
  • Follow through on outcomes

The best CISOs do not just explain what is wrong. They clarify what matters, what the options are, what they recommend and what happens next. We must be the Department of KNOW, not the Department of NO. Showing risk reduction while enabling business speed is table stakes.

 

The CISO role is about governance, not technical competency alone. Technical expertise earns credibility. Executive influence earns budget, authority drives organizational change, and together they determine whether a security strategy succeeds.

A model referenced by others before me is to develop business leaders who understand security, not just security experts who learn to present. That means building financial acumen, legal and regulatory awareness, executive communication and adaptability alongside technical depth. Through progressive exposure to business decisions, risk ownership and board-level communication, is how we each continue to grow and accelerate. While certifications demonstrate knowledge, leadership requires judgment, credibility and the ability to decide under uncertainty.

 

I emphasized that learning through watching other leaders in the boardroom and the aptitude to learn and adapt is success. A few practical lessons from my own path:

  • Reporting structure is a learning opportunity – Independence and objectivity are essential but proximity to other executives accelerates growth.
  • Mentorship runs both ways – Sitting in on committee and board meetings helps but sound coaches and supervisors are what set people apart.
  • The CISO is a talent builder – Retaining and growing scarce cyber talent matters more than any tool you buy.
  • The CISO is a negotiator – We constantly broker trade-offs between speed and safety with business units, vendors and the budget.
  • Do not wait – Develop the next generation deliberately before you need them.

 

Finally, the fundamentals still matter. Foundational standards remain the building blocks that must be done comprehensively and effectively before we can soundly address new threats and trends. That point also echoed through nearly every session I attended.