FreezePoints eMagazine

At PMI Orange County’s December dinner meeting on Tuesday, December 9, 2025, Glenn Hoxie delivered a practical keynote on generative AI that balanced two ideas: the real promise of generative AI for project managers, and the guardrails required to use it responsibly.

Glenn has been a PMP since 2015 with two decades in Legal IT (supporting AmLaw 10 and AmLaw 100 firms) and ten years as a Practice Solutions Manager at an AmLaw 10 firm. He opened with the premise that AI is both exciting and something to approach with care. Instead of just giving a keynote from a slidedeck, he facilitated a live, hands-on exploration of three real-world PM use cases using Google Gemini, Microsoft Copilot, and PMI Infinity, with volunteers from the audience trying prompts in real time.

Live Demos

Glenn structured the session around three scenarios, each with a clear learning objective and a short live exercise so attendees could see how AI can accelerate everyday PM work.

Scenario 1:  Meeting Synthesis (Google Gemini)

Glenn gave an audience volunteer a mock transcript and then they used Gemini to convert that raw transcript into structured notes including action items and dependencies while the audience followed along. The demo showed how quickly a messy transcript can become usable meeting minutes and a task list, but Glenn reminded everyone to “check and verify” AI outputs to maintain accuracy. Other audience members shared additional tips: upload meeting minutes, share your screen for review, or use role-play prompts to improve context. This example showed the audience how to use generative AI for meeting synthesis, communication, and risk analysis.

Scenario 2:  Risk Analysis (Microsoft Copilot)

Using a sample Work Breakdown Structure (WBS), a second volunteer used Copilot to draft a Risk Register. Glenn used the exercise to stress an essential point for reliable AI use: be specific. The more precise the prompt (context, scope, assumptions), the more useful and actionable the AI output. The demo showed AI’s usefulness for surfacing risks and brainstorming mitigations, but the PM needs to validate and tailor the results. This scenario showed the audience how to build confidence by practicing simple, reliable AI prompts during live exercises.

Scenario 3: Stakeholder Communication (PMI Infinity)

Glenn used fabricated meeting notes to generate multiple communication drafts tailored to different audiences: C-suite, day-to-day teams, and a direct manager. AI can quickly adapt tone and focus for different stakeholders, but communication outputs must be reviewed for accuracy, sensitivity, and alignment with governance. This scenario helped the audience explore ethical considerations, data governance, and an AI integration roadmap.

Guardrails for Safe and Effective AI adoption

A portion of Glenn’s keynote was devoted to practical guardrails and mitigation strategies for project managers:

1. Hallucinations & Inaccuracy: AI can produce confidently wrong facts.
   Mitigation: Human-in-the-loop validation; fact-check critical outputs; ground AI on trusted data.
2. Intellectual Property (IP) Risk: Outputs might infringe or reveal proprietary training data.
   Mitigation: Use AI for drafts only; edit and verify originality before publishing.
3. Data Leakage & Security: Pasting confidential project data into public tools is risky.
   Mitigation: Use enterprise-approved tools; train teams on a “no sensitive data” policy; anonymize inputs.
4. Inherent Bias & Fairness: AI may replicate or amplify bias (e.g., skewed resource recommendations).
   Mitigation: Scrutinize AI decisions for systemic bias and use human judgment.
5. Lack of Explainability: AI outputs without reasoning leave PMs unable to justify decisions.
   Mitigation: Prompt AI to “show its work”; document the rationale for critical decisions.

Why This Matters

The presentation made two things plain: AI is already useful for routine PM tasks (saving time on meeting notes, surfacing risks, tailoring communications), and adopting AI without governance creates real project, legal, and ethical risks. Glenn’s live demos showed immediate, practical examples while his guardrails gave PMs an actionable framework to adopt AI responsibly.

For PMI-OC members, the message was balanced and empowering: start small, prompt precisely, verify everything, and build the policies and human checks that let AI accelerate work without replacing accountability.