In this week’s episode of the Practical 365 Podcast, Paul Robichaux and I covered everything from the latest Exchange Server updates to the reality behind AI agent hype. New host Bastiaan Verdonk and I also welcomed cybersecurity expert Paula Januszkiewicz, who shared eye-opening stories about security breaches and offered practical advice for organizations of all sizes.
The Experts Conference (TEC) Events
We kicked off the show with news about the upcoming Experts Conference Tech events across Europe, with stops in London, Paris, and Dusseldorf. These events will focus on Microsoft security and AI, with our own Tony Redmond featured as a speaker.
Exchange Server 2025 H1 Cumulative Update
Next, we discussed the latest updates for Exchange Server, focusing on the 2025 H1 cumulative update and Exchange Server Subscription Edition (SE). This update brings several important changes:
- The “Feature flighting” capability allows Microsoft to enable new features across Exchange Server SE organizations
- Support for Windows Server 2025
- The return of certificate management to the Exchange Admin Center
- Blocked coexistence with Exchange 2013
This move to a subscription edition represents a significant shift, with organizations now receiving ongoing updates rather than major releases every few years. We also touched on the persistent challenges many organizations face with email deliverability due to misconfigured DNS records like SPF and DKIM.
AI Agents: Cutting Through the Hype
We examined the current hype surrounding AI agents, particularly within the Microsoft ecosystem. I noted the confusing renaming of the consumer version to “Microsoft 365 Copilot for Consumer,” while Paul and I discussed how quickly the market has shifted from focusing on AI models to declarative agents – following a pattern reminiscent of the Gartner hype cycle.
Paul is skeptical about claims that AI agents can accomplish tasks that are otherwise impossible, while I highlighted the prevalence of advertisements (featuring Matthew McConaughey and others) that overstate AI capabilities. We both cautioned against making technology decisions based solely on hype.
We agreed that AI tools aren’t close to replacing human jobs entirely – even seemingly simple tasks like calendar scheduling aren’t reliably solved by today’s AI agents. The greatest value of AI lies in augmenting human capabilities and making existing roles more effective rather than replacing them. Paul noted the low barrier to entry for experimenting with AI tools but warned organizations should be wary of vendor lock-in.
I emphasized the importance of Microsoft sticking to their “copilot” vision – focusing on making people more effective rather than replacing them entirely. We also discussed the potential economic impact if AI is solely focused on extracting knowledge and replacing jobs, though we concluded with a somewhat lighthearted outlook that fears of AI taking over are likely overblown.
Cybersecurity Insights with Paula Januszkiewicz
The highlight of our episode was our interview with Paula Januszkiewicz, a globally recognized cybersecurity expert and founder of CQURE. As a Microsoft MVP and top-rated industry speaker, Paula brings deep technical knowledge in Windows security, enterprise security, penetration testing, and forensics to her work. She will also be a keynote speaker at the upcoming TEC European Roadshow, next month in London, Paris, and Düsseldorf.
Security Fundamentals for Organizations of All Sizes
Paula emphasized that small businesses are often the most vulnerable to cyberattacks due to limited security measures. She recommended implementing basic protections like multi-factor authentication and email security features to significantly improve baseline security. Even large organizations can be vulnerable, as her work in the oil and gas industry has demonstrated. She noted that phishing attacks using shared documents are becoming increasingly common and effective.
AI’s Dual Role in Cybersecurity
Our discussion revealed how AI is transforming both sides of the cybersecurity battlefield. Organizations can leverage AI-powered security solutions for better threat detection and investigation, but attackers are using AI to automate attacks, find vulnerabilities, and generate exploits.
Paula warned about what she called “productive script kiddies” – less skilled attackers who now leverage AI to create more sophisticated attacks. She emphasized that organizations need robust defensive measures like anti-exploitation solutions, XDR, and event correlation to stay protected. Those with well-established environments and robust monitoring are more resilient, but many organizations still lack basic security hygiene.
Red Team Tales and Social Engineering
Some of the most fascinating insights came from Paula’s red teaming experiences. She shared how gathering information about employees and their habits is a key part of red teaming, and how social engineering can be surprisingly effective. In one example, her team modified a water cooler to gain network access and eavesdrop on conversations.
Her “best” story involved making a domain admin cry after successfully gaining domain admin privileges during a penetration test. Her “worst” involved a production factory crippled by ransomware after a domain admin password was exposed on a public GitHub repository. Despite paying the ransom, the company failed to recover its data, highlighting the critical importance of reliable backups.
Security Hygiene Fundamentals
Throughout our conversation, Paula returned to the fundamental importance of security hygiene. Poor password management practices remain common, with passwords still stored in plain text files. Developers sometimes unintentionally expose sensitive information in public repositories, and test servers get exposed to the internet, creating significant risks.
She noted the tension between startup culture and security, suggesting that security should be integrated into the development process without excessively slowing things down. Paula emphasized that proactive security measures are essential to stay ahead of increasingly sophisticated threats, and organizations should always assume breaches may have done more damage than initially apparent.
On the Next Show
Join us in two weeks as we continue to explore the practical side of Microsoft 365 and meet another TEC speaker who’ll be joining us on the show. Don’t forget to subscribe on iTunes or Spotify to make sure you don’t miss the next episode. Have thoughts on this episode or suggestions for future topics? Let us know in the comments below!
