The Security Strategist

In the digital age, securing sensitive business information has never been more critical. Microsoft 365 has become the backbone of operations for organisations worldwide, and with that centrality comes an expanding attack surface that many security teams are only beginning to fully understand.
In a recent episode of the Security Strategist podcast, host Richard Stiennon sat down with Rob Edmondson, Senior Director of Product Marketing at CoreView, to unpack the practical realities of Microsoft 365 security. The conversation covered configuration drift, excessive privilege, tenant hardening, and the emerging security challenges posed by AI agents offering actionable guidance for security professionals at every level.
Microsoft 365 Environment
Microsoft 365 has changed significantly from a simple productivity platform into a comprehensive security concern in its own right. As Edmondson points out, the transition from Office 365 to Microsoft 365 marked a pivotal shift in how organisations utilise these tools. What began as a suite of familiar applications, such as Word, Excel, and Outlook, has grown into an interconnected ecosystem of over 60 apps and services, from Teams and SharePoint to Power Automate, Defender, and Purview. That expansion has delivered enormous productivity gains, but it has also multiplied the potential vectors for security vulnerabilities exponentially. Every additional service is a new configuration surface, a new set of permissions to govern, and a new integration that must be secured. Understanding this evolution is the essential starting point for any organisation serious about Microsoft 365 security.
Configuration Drift and Why It Puts Microsoft at Risk
Configuration drift is one of the most pervasive and underappreciated threats in Microsoft 365 environments. It refers to the gradual, often unnoticed divergence of system configurations from their original, secure baseline, which is a slow accumulation of small changes that individually seem harmless but collectively create significant vulnerabilities.
Edmondson highlighted that most organisations lack adequate visibility into how their Microsoft 365 tenant is actually configured at any given moment. Many still rely on manual methods like spreadsheets, periodic snapshots, and ad hoc reviews to track configuration state. This approach is fundamentally inadequate in environments where settings can change daily, sometimes through automated processes or third-party integrations that bypass normal change management controls.
The consequences of undetected configuration drift can be severe. Breaches have been traced directly to unauthorised or unintended configuration changes, a permissions setting quietly altered, an authentication policy weakened, or a data loss prevention rule inadvertently disabled.
Microsoft 365 Security Posture
Excessive privilege is consistently ranked among the leading contributors to security incidents in cloud environments, and Microsoft 365 is no exception. When users, service accounts, and applications hold more permissions than their role requires, the potential blast radius of any compromise — whether through phishing, credential theft, or insider threat — expands dramatically. Edmondson walked through the practical challenge: in large organisations, permissions accumulate over time. A user gets temporary admin access to complete a project, and that access is never revoked.
AI Agents in Microsoft 365
As organisations adopt AI-driven tools and agents within their Microsoft 365 environments, a new and largely uncharted security frontier is emerging. AI agents — automated systems capable of acting on behalf of users, reading emails, accessing files, and executing workflows — introduce permissions challenges that most security frameworks were not designed to handle.
Edmondson was candid about the challenge: many organisations deploying AI agents do not have clear visibility into what those agents can access, what data they are interacting with, or whether the permissions they hold are appropriate. In an environment where an AI agent might have access to the entire Microsoft 365 data estate of a user or a team, the consequences of a misconfigured or compromised agent are significant.
The same principles that govern human access with least privilege, continuous monitoring, and regular review must be extended to AI agents. This requires both the technical capability to enumerate agent permissions and the governance processes to enforce appropriate boundaries. Organisations that deploy AI capabilities without first establishing this control layer are trading short-term productivity gains for long-term security debt.
Microsoft 365 Security
In the fast-moving threat landscape, understanding and proactively strengthening your Microsoft 365 security posture is no longer optional; it is a business imperative. Configuration drift, excessive privilege, and AI agent governance are not edge cases; they are mainstream risks affecting organisations of every size and sector. The insights shared by Edmondson on the Security Strategist podcast provide a practical foundation for addressing each of these challenges with clarity and urgency.
By implementing continuous monitoring, enforcing least-privilege access, hardening your tenant configuration, and extending security governance to AI agents, organisations can significantly reduce their exposure and build a Microsoft 365 environment that is resilient by design. For further insights and tools to support your Microsoft 365 security journey, visit CoreView.
Takeaways
Configuration drift and its impact on security.
Excessive privileges and how to mitigate them.
Tenant hardening best practices.
Managing AI agents and permissions in Microsoft 365.
Strategies for continuous security monitoring.

Chapters
00:00 Introduction to Microsoft 365 Security
02:25 The Shift to Security Priority in Microsoft 365
04:30 Understanding Configuration Drift
09:09 Excessive Privilege and Its Risks
12:48 AI Agents and Identity Security
16:20 Tenant Hardening and Common Misconfigurations
18:36 Recommendations for Strengthening Security Posture

Financial crime is no longer a peripheral concern for banks and fintechs; it is a defining operational challenge. The pressure to grow transaction volumes, onboard customers quickly, and keep pace with increasingly sophisticated fraud actors has placed finance and compliance teams at the very heart of business strategy. For many institutions, the question is no longer how to use artificial intelligence in their fraud detection stack, but how to use it responsibly.
In this Security Strategist podcast, hosted by Jonathan Care, Senior Lead Analyst at KuppingerCole, he speaks with Kunal Datta, Chief Product Officer at Unit21, about the changes in financial crime prevention technology and the gaps that remain in the industry.
The role of AI in fraud detection
For most of the past two decades, financial crime prevention operated on one of two tracks. Larger, data-rich institutions invested in machine learning models capable of identifying complex behavioural patterns across millions of transactions. Smaller players, or those entering new product categories with thin data histories, tended to rely on rules-based systems, which are explicit, human-authored logic that flags transactions meeting predefined criteria.
Both approaches have genuine strengths. Rules-based systems are auditable, easy to explain to a regulator, and quick to update when a new fraud typology emerges. Machine learning systems are far more powerful at surfacing non-obvious correlations and adapting to evolving attack patterns, but they require substantial training data and significant engineering effort to deploy.
The arrival of large language models and generative AI has introduced a third paradigm, one that is fundamentally non-deterministic. Unlike a rule that fires predictably on every run, or an ML model that produces a consistent probability score for a given feature vector, a generative AI system may reason differently across identical inputs. This has profound implications for how institutions build, test, and govern their fraud detection infrastructure.
Balancing revenue growth and fraud risk
Perhaps the most underappreciated tension in financial crime prevention is not technical; it is commercial. Every fraud control is also a friction point. A transaction declined as suspicious is, from the customer's perspective, simply a transaction that failed. Every false positive erodes trust, damages conversion rates, and risks losing a customer to a competitor with a more permissive onboarding flow. According to Datta:
“Machine learning excels at identifying complex patterns, but rules-based systems can quickly adapt to new types of fraud that humans can spot with minimal examples.”
This means that fraud teams are never simply optimising for fraud prevention in isolation. They are solving a constrained optimisation problem that is minimising fraud losses while simultaneously protecting revenue, preserving customer experience, and staying within the bounds of what regulators require. AI can shift that frontier, enabling more precise risk assessment that reduces both fraud and false positives simultaneously. But only if it is deployed and governed carefully.
The future of AI in financial crime
Looking forward, Datta sees the trajectory of AI in financial crime prevention pointing towards systems that combine the pattern-recognition power of machine learning with increasingly robust mechanisms for transparency and accountability. The goal is not to choose between a powerful AI and an explainable one — it is to build infrastructure that delivers both.
Several technical approaches are emerging to close this gap. Structured output formatting — requiring AI systems to return decisions in machine-readable formats like JSON, with explicit reasoning chains, makes it possible to audit AI behaviour at scale. Evaluation sets, which establish a curated baseline of labelled cases against which model performance is continuously benchmarked, allow institutions to detect drift and maintain defensible performance records.
The institutions that will lead this space are those treating AI governance not as a compliance overhead but as a competitive advantage. A well-governed AI system is faster to get regulatory approval, faster to deploy new capabilities, and more resilient when regulatory scrutiny increases.
The most striking thread in Datta's thinking is his insistence on placing financial crime prevention within a broader moral frame. Financial crime is not merely an operational risk; it is a conduit for some of the most serious harms in the world: human trafficking, modern slavery, terrorist financing, and the systematic exploitation of vulnerable people. Viewed through this lens, the deployment of better AI in financial crime prevention is not primarily a business efficiency story. It is a contribution to a more just and safer world. Datta says:
“AI should be viewed not only as an efficiency driver but as a tool to address broader societal issues like human trafficking and exploitation. Better detection is a moral obligation.”
This framing matters for how organisations think about investment in financial crime technology. If AI in fraud prevention is purely a cost centre, it will always lose budget battles to revenue-generating activities.
If you would like to find out more, visit: Unit21.ai or read more about Rules vs. Machine Learning: Finding the Best of Both Worlds by Kunal Datta.
If you are looking to strengthen how your organisation identifies and manages risk, you can request a personalised demo with Unit21.
Takeaways
Evolution of financial crime detection over the last decade
Deterministic vs non-deterministic AI systems in fraud prevention
The role of generative AI and context engineering in compliance
Accountability and explainability in AI-driven decision making
Regulatory perspectives on AI and risk management

00:00 Navigating Financial Crime Prevention Challenges
02:54 The Evolution of Fraud Detection Systems
05:55 The Debate: Explainability vs. Performance in AI
08:51 Balancing Accuracy and Regulatory Expectations
12:01 Context Engineering in AI for Financial Crime
15:04 Rethinking Accountability in AI Systems
17:55 AI as a Societal Imperative in Risk and Compliance

Podcast: The Security Strategist
Guest: Rick Wagner, Senior Director, Product Management at SailPoint
Analyst: Jonathan Care, Lead Analyst, KuppingerCole
The identity security market is crowded, but a significant change is occurring below the surface. In a recent episode of The Security Strategist podcast, host Jonathan Care, Lead Analyst at KuppingerCole, sat down with Rick Wagner, Sr. Director Product Management at SailPoint.
In this episode, Wagner pointed out a growing gap between how enterprises manage access and how modern systems operate. As AI and machine identities grow rapidly, traditional models no longer work.
Static Access Reviews Are Breaking at Scale
For years, enterprises have depended on periodic access certifications to manage access. However, such a model is proving to be weak. “Periodic access reviews only look at appropriate access at a point in time,” says Wagner, noting that “certification fatigue results in rubber stamping.”
The challenge is both scale and accuracy. With machine identities often outnumbering humans, governance processes designed for manual oversight are quickly becoming outdated. “Doing those certifications at agent speed is literally impossible,” he adds, emphasising the need for change.
Also Watch: Why AI Agents Demand a New Approach to Identity Security
How is Real-Time Authorisation & AI Redefining Identity Security?
The way ahead is real-time authorisation, which continuously checks if access is appropriate at the moment it is requested. “It’s not only appropriate— is it appropriate right now?” Wagner explains.
This change depends on context, incorporating information such as device health, user behaviour, and risk level. Frameworks like the Shared Signals Framework help enterprises implement this by allowing real-time data sharing across the security ecosystem. This approach leads to more dynamic, policy-driven access that keeps pace with AI systems.
How to Tackle Shadow AI?
At the same time, CISOs face the rise of shadow AI, an expanding network of agents operating with little oversight. “You can’t manage what you can’t see or what you don’t know about,” says Wagner, highlighting visibility as the first line of defence.
The long-term goal is autonomous identity governance, where systems continuously evaluate and adjust access based on risk. “As risk levels start to increase, we might add additional factors up to quarantining that access,” he explains.
In this new framework, identity becomes the core of cybersecurity strategy. As Wagner puts it, the ongoing challenge is urgent – determining “who has access to what—and is that access appropriate right now.”
Key Takeaways
Real-time identity governance replacing static access reviews
AI and machine identities outpace human oversight
“Certification fatigue” is weakening traditional access controls, increasing risk through unchecked approvals.
Non-human identities (AI agents, bots) are now the fastest-growing and least visible attack surface.
Context-aware access decisions—based on risk, behaviour, and environment—are becoming the new standard.
Visibility into agents and their permissions is critical: “you can’t manage what you can’t see.”
Autonomous, risk-adaptive identity security is emerging as the end-state for modern enterprise cybersecurity.

Chapters
00:00 Introduction to Identity Security in AI Era
06:54 Managing Privileged Access Risks
13:52 Real-Time Governance and Joiners, Movers, Leavers
20:14 Strategic Moves for CISOs in Agent-Based Operations
For more information, please visit em360tech.com and sailpoint.com.
To stay updated on B2B Tech front and centre, follow EM360Tech:
YouTube: @enterprisemanagement360
LinkedIn: @EM360Tech
X: @EM360Tech
Follow SailPoint on all its major platforms:
YouTube: @SailPointTechnologies
LinkedIn: @SailPoint
X: @SailPoint
#IdentitySecurity #AIAgents #RealTimeGovernance #SailPoint #IAM #ShadowAI #Cybersecurity #EnterpriseTech #TechLeadership #CIOInsights #DigitalTransformation #MachineIdentities

Over 95 per cent of leaders now say identity security is core to their strategy. A decade ago, this wasn’t even part of the conversation. The awareness is there, but awareness alone isn’t enough. Many organisations feel secure, yet the metrics they track often tell a different story.
In this episode of Security Strategist, EM360Tech’s Trisha Pillay sits down with Craig Ramsay, Senior Field Strategist, and Rod Simmons, VP of Product Strategy at Omada, to unpack the State of Identity Governance 2026 report. Together, they explore why confidence in identity security doesn’t always equal true protection and how AI, non-human identities, and fragmented systems are changing the rules.
Bridging the Gap Between Perception and Reality
Many organisations focus on operational metrics that are easy to measure: provisioning speed, audit readiness, and compliance. These give a sense of efficiency but not necessarily security. Simmons explains: “We can provision identities faster, but that doesn’t tell us about inherent risks. Orphaned accounts, dormant privileges, unmanaged access—these risks often go unseen.”
Ramsay adds, “It’s like home security. You might feel confident, but when was the last time you checked your back door?”
The survey revealed a clear disconnect: strategic awareness exists, but organisations are not always measuring the right things. Security leaders should not only track completed tasks, but they must also understand where risk accumulates and how quickly they can respond to incidents. Risk-based metrics, rather than activity-based metrics, are the key to true governance.
Zero Trust and the Challenge of Integration
Almost every organisation reports adopting Zero Trust principles. The execution often falls short. Policies may exist in pockets, but full implementation requires connected systems that can share signals in real time. Without this integration, Zero Trust becomes a concept rather than a functioning model.
Rod highlights the issue: “It’s one thing to want continuous evaluation, but another to have systems that actually support it. Shared signal frameworks are essential for consistent enforcement across the enterprise.” Until Zero Trust principles are fully integrated across all platforms, access control and identity governance will remain reactive rather than proactive.
Non-Human Identities, AI, and the New Frontline
Identity is no longer just about people. Non-human identities, but API keys, service accounts, and AI agents, are multiplying at unprecedented rates. Some organisations see 150 non-human identities for every human. These identities act autonomously, persistently, and at scale. Simmons explains the challenge: “With human identities, we ask what access they have. With non-human identities, we ask what they can do, and what they’ve done.”
Ramsay adds a crucial reminder: “Artificial intelligence still needs an accountable individual. Human oversight is essential, even as AI agents scale and operate independently.”
These agents create both risk and opportunity. They can automate governance, improve provisioning, and flag anomalies—but without proper visibility and ownership, they become a blind spot. Over 40 per cent of surveyed organisations admitted their AI agents still use static credentials, a simple but serious vulnerability.
One thing is for sure: you cannot govern what you cannot see. Visibility is the foundation. Only once organisations know what exists, who owns it, and how it behaves can they secure identities, human and non-human alike, effectively.
Identity security is no longer a back-office concern—it’s strategic. Organisations must move from confidence to proof, from operational reporting to risk measurement, and from fragmented controls to integrated governance. AI and non-human identities are not just a challenge; they are an opportunity to rethink how identity security can truly enable business, not just protect it.
For more insights on effective identity governance strategies, check out Omada's State of Identity Governance 2026 Report.
Takeaways
Over 95 per cent of security leaders now see identity as a core strategy. Identity isn’t optional anymore.
Feeling secure doesn’t equal being secure. Many organisations track efficiency, not actual risk.
Non-human identities are multiplying fast.
Zero Trust adoption is growing, but integration gaps remain.
AI in identity governance works, but always keep a human in the loop.

Chapters
00:00 Introduction to Identity Governance and Security Challenges
02:55 Insights from the State of Identity Governance Report
05:53 The Gap in Security Confidence and Measurement
08:53 Operational Metrics vs. Risk Indicators
11:50 Zero Trust Adoption and Implementation Challenges
14:54 The Role of AI in Identity Governance
17:52 Non-Human Identities and Governance Challenges
21:07 Key Takeaways for Security Leaders

Podcast series: The Security Strategist
Guest: Amit Megiddo, CEO and Co-Founder, Native
Host: Richard Stiennon, Chief Analyst Researcher at IT-Harvest
In the recent episode of The Security Strategist Podcast, Amit Megiddo, CEO and Co-Founder, Native, joins host Richard Stiennon, Chief Research Analyst at IT-Harvest, to discuss a growing challenge in enterprise cloud security. Enterprises are investing heavily in cloud providers’ built-in controls, yet risk persists when those controls are not consistently enforced across complex environments.
According to Megiddo, the problem isn't a lack of tools, but a failure to make them work effectively. Drawing on his experience launching Amazon GuardDuty at Amazon Web Services, the Native CEO explains that enterprises have hit a tipping point. The challenge is no longer about visibility. It is about executing at scale across complex multi-cloud environments.
What is the Execution Gap in Cloud Security?
Cloud providers such as Amazon Web Services, Microsoft Azure, Google Cloud, and Oracle Cloud offer a wide range of built-in security features. Yet, as Megiddo points out, most enterprises are only using a small part of what is available.
“The easy part is turning controls on,” he says. “The hard part is making sure they consistently deliver security results.” This is where many enterprises struggle. Security teams create policies, but platform teams carry them out. In the process, vital context is lost. The result is a disjointed approach where risks are identified but not effectively managed.
Megiddo calls this the “execution gap.” It is a fundamental issue in how enterprises handle cloud security. Even with sophisticated CSPM and CNAP tools, organisations remain mostly reactive. They are relying on detection and fixing problems instead of preventing them.
How to Move From Detection to Policy-Driven Enforcement
The podcast spotlights a key shift in enterprise security strategy – moving from detection controls to proactive, policy-driven enforcement. Conventional methods focus on spotting issues—like unencrypted or publicly exposed data—and then starting remediation processes. However, as cloud environments grow, this method becomes untenable.
Megiddo suggests embedding security directly into the architecture:
Preventing non-compliant resources from being created
Designating approved regions for workloads
Enforcing network isolation rules for sensitive environments, such as AI training workloads

This “secure-by-design” approach turns security from a reactive task into a core operational control. However, implementing this is not easy. Enterprises must translate high-level policy goals into thousands of low-level settings across various cloud providers, each with its own APIs, services, and policy frameworks.
“It’s not just about writing the policy,” Megiddo emphasises. “It’s about safely rolling it out, simulating impact, managing exceptions, and ensuring it stays enforced over time.”
It creates new operational needs such as simulation tools, drift detection, real-time developer feedback, and automated exception handling. Essentially, cloud security becomes a continuous process rather than a one-time setup.
Why is the Unified Control System Critical?
The main takeaway for enterprise leaders is that cloud security is no longer just about managing risks; it is becoming an edge in the market. As major providers continue to invest heavily in native security features, the real differentiator will be the ability to coordinate and enforce those tools effectively.
Megiddo’s vision is straightforward: a unified control system that lets enterprises define security intent once and apply it consistently across cloud and hybrid environments.
In an industry shaped by AI, multi-cloud complexity, and rapid digital changes, this ability could determine how quickly—and securely—enterprises can progress. For CISOs and IT leaders, the message is clear: the future of cloud security lies not in observing more, but in doing more—with precision, consistency, and scale.
Key Takeaways
Shift from detection to proactive, policy-driven cloud security to reduce risk.
Multi-cloud across Amazon Web Services, Microsoft Azure, and Google Cloud requires unified enforcement.
CISOs need tools that turn security policy into automated controls.
Secure-by-design cloud architecture protects AI and enterprise workloads.
Strong cloud security execution drives scalability and resilience.

Chapters
00:00 The Cloud Security Landscape
03:11 Challenges in Implementing Cloud Security
08:00 Transitioning to Proactive Security
12:26 The Evolving Role of Security Leaders
16:42 Future Trends in Cloud Security

For more information, please visit em360tech.com and native.security.
Follow: @EM360Tech on YouTube, LinkedIn and X
Native LinkedIn: https://www.linkedin.com/company/native-security/
#CloudSecurity #PolicyDrivenSecurity #CloudEnforcement #MultiCloudSecurity #SecurityByDesign #ExecutionGap #CISOs #TheSecurityStrategist #NativeSecurity #CSPM #CNAP #EnterpriseSecurity #NativeSecurity #AmitMegiddo