What does secure AI adoption look like when your engineering teams are expected to experiment at the frontier? Mark Strande, CISO at Miro, joins Beyond Vibe Coding to explain why the answer is not a single policy or a single access-control layer. It is a layered security model that combines foundations, agentic principles, hardened integration points, pipeline controls and fast detection.
Mark also challenges the way companies think about agentic work. Reliable, repeatable tasks need clear skills, tests and constraints, not just a more elaborate harness. As agents gain access to systems and data, traditional human-style roles create dangerous combinations of permissions. The next generation of security architecture must make agent identities more precise and segregation of duties more automatic.
The conversation closes with a pragmatic lesson for technology and security leaders: aim for excellence, but do not postpone meaningful protection while pursuing a perfect design. Institutional knowledge, reactive security systems and the ability to learn faster than the threat landscape will shape the winners.
–
Links and References
Beyond Vibe Coding Podcast is a project by Sebastian Heide-Meyer zu Erpen and André Neubauer, partnering with Impala Search. The content is created by us and our guests. Join the discussion on LinkedIn or visit http://www.bvc.fm, where we publish all episodes. For questions and inquiries, feel free to contact us via LinkedIn. Thank you for your time and see you in the next episode.