Obsidian Security: Agentic Security: Safeguarding AI Agents' Access to Third-Party Apps
Discussion Points
1. Non-human identity, access and privileges
Controlling what AI agents and service accounts can reach, and with what permissions.
2. AI inventory
Knowing which AI tools, agents and connections exist across the business.
3. End-user AI behaviour and risk
Understanding how employees really use AI, and where that creates exposure.
4. DLP controls for the AI era
Keeping sensitive data in bounds as it flows into AI tools and agents.
5. MCP gateway and governance
Governing the connectors that link AI systems to third-party applications.
6. AI run-time enforcement
Applying controls while agents are acting, not just at set-up.
7. AI and SaaS policy enforcement
Turning AI and SaaS policies into consistent, enforceable rules.
Agentic Security: Safeguarding AI Agents' Access to Third-Party Apps
Every enterprise is racing to accelerate its AI journey. To maximise ROI, AI agents need access to the critical company data that typically lives in third-party applications such as Databricks, Snowflake, Salesforce, Slack, ServiceNow and SAP — because agents require that access to make real work happen.
MCP servers are quickly becoming the default connector between AI systems and those applications. That significantly widens the attack surface, and without the right controls, visibility is extremely challenging: it is impossible to manage what you cannot see.
This roundtable brings together senior security leaders to explore how human and non-human identity controls can be extended into third-party applications, so AI programmes can accelerate in a safe, controlled and secure way.
Agenda
Welcomed attendees are integrated into a curated networking space, fostering direct interactions for ice-breaking and relationship-building before engaging in substantive, collaborative discussions.
Shifting to a sit-down discussion with drinks, this promotes open, thoughtful dialogue on a set topic agenda, aimed at fostering collaboration and deepening connections in a relaxed, supportive setting.
The evening culminates with a luxury meal, offering exquisite cuisine in an elegant setting, enhancing networking opportunities, and providing a memorable experience that strengthens professional bonds among attendees.
The finale features a wrap-up and Q&A session, where sponsors address the roundtable, followed by drinks and networking, enabling direct engagement and planning of actionable next steps in a collaborative environment.
About the Hosts
Obsidian Security secures the third-party applications and AI agents where modern enterprises actually run. With business data now living in Salesforce, Workday, ServiceNow and thousands of other SaaS apps — and non-human identities such as service accounts, API tokens and AI agents now outnumbering people — Obsidian gives security teams visibility into what's connected, controls to enforce what's allowed, and clarity on what needs to change.
Built on years of mapping identities, permissions and data flows across the enterprise application estate, Obsidian is hosting this private dinner roundtable to help senior security leaders accelerate AI adoption safely.
Why Attend?
Exclusive Networking
Insightful Discussions
Tailored Content
Strategic Takeaways
Venue Information
Situated in the heart of Marylebone Village, this elegant venue offers a refined and intimate setting for our exclusive event.
Attendees and sponsors will enjoy a sophisticated dining experience in a private room, designed to foster meaningful conversations and connections. The Marylebone Hotel is renowned for its exquisite cuisine, crafted by top chefs to provide a memorable culinary journey.
The hotel's chic, contemporary decor, combined with exceptional service, ensures a seamless and enjoyable evening. Join us at The Marylebone Hotel for a roundtable dinner that blends culinary excellence, elegant ambiance, and the vibrant charm of central London.
