Most IT organizations aren't behind on technology — they're behind on thinking. In this episode, Pops and Steele sit down with Matt Coatney, CIO of a large national law firm, to break down persona-based agentic AI: virtual agents that adapt to *who* they're talking to, not just *what* they're asked. Matt shares how a law firm environment — high-touch partners, urgent deadlines, wildly different workflows across practice groups — makes generic self-service fall flat, and what it actually takes to build role-aware automation that works. They dig into data readiness, the Big Brother problem, tier-zero service desk automation, and where the real ROI conversation with a skeptical CFO needs to start. Matt also gets candid about the risks of over-permissioned AI agents digging up things "security by obscurity" used to hide.

Key Takeaways
- Persona-based AI isn't mainstream yet — the technology exists, but adoption is gated by change management, privacy comfort, and governance, not capability.
- Generic self-service fails high-touch environments. A law firm with hundreds of partner "entrepreneurs" each running practices their own way can't be served with one-size-fits-all automation — urgency and workflow context matter enormously.
- Data exhaust is the unlock. Ticket history, assets, and system usage are already-known data that can power personalization without requiring people to hand over new personal information.
- Big Brother concerns are real but manageable. Personalization lands well when it's baked into an expected workflow (like a service desk that already has ticket history) rather than feeling like surveillance.
- Agentic AI raises new data-handling risks — attachments containing PII/PHI, agents that could store or forward sensitive data, and tools like Copilot surfacing improperly secured internal information that "security by obscurity" used to hide.
- ROI is often about noise reduction. In flat organizational structures, faster, better, more personalized service reduces escalations straight to the CIO/COO — a compelling case even without hard automation-cost metrics.
- The tier-zero cleanup loop: using AI-assisted human agents today (suggested KB articles, predictive closure notes) trains and cleans the knowledge base, setting up cleaner true self-service later.
- Personal accountability doesn't disappear with AI. Ceding too much judgment to an agent ("just read my email and tell me what matters") risks missing what actually counts — and someone still owns the outcome.
- Advice for leaders starting out: network with peers (ILTA, vendor conversations), and personally use the tools — you can't lead AI adoption you haven't internalized yourself.

Keywords / Tags
persona-based AI, agentic AI, ServiceNow, IT service desk, self-service automation, tier zero support, law firm technology, legal IT, CIO leadership, AI governance, data privacy AI, Now Assist, ITSM automation, enterprise AI agents, Copilot security, change management AI, CMDB, knowledge management, virtual agents, AI adoption

If you're an IT leader wrestling with self-service that feels generic, this episode is your blueprint for making it personal — subscribe to The Wired Garage with Pops so you don't miss the next conversation, drop a comment on where your organization stands on the "mainstream vs. science fiction" scale, and share this with a fellow CIO who's still fighting the one-size-fits-all portal.

Support the show