Summary

zkjYLPaU7jACmiU4qqtmPjjVGgiL8qnPLQ2bz7PZPHpTLMKq_Z_-G-WxfF3ahbNkl55-_yYna191gOTwN8cam5x13tNxMzABFkFT9wvKiyfaJpz1
Edit Back
Meeting Date
Unknown
Meeting Type
Other
Note Type
Summary
Attendees

1. Meeting Summary

I participated in an AI training session led by Shawn and Gage that focused on our current enterprise approach to generative AI, model selection, prompt engineering, data governance, and the practical distinction between Copilot and ChatGPT. The discussion reinforced that our AI strategy is intentionally split between individual productivity use cases and broader technology-enabled transformation efforts. The team is trying to balance speed, productivity, financial discipline, data privacy, and long-term platform stability rather than chasing every tool labeled as AI.

A major takeaway for me was the practical case for Microsoft Copilot as the preferred enterprise tool for most internal work. Shawn and Gage made the distinction between Copilot basic and enterprise, particularly around the ability to search internal work content such as email, Teams, OneDrive, and shared documents. I also shared my own experience using Copilot for annual review preparation, where it helped surface employee interactions, work history, and themes I would not have found efficiently through manual review. I emphasized to the group that Copilot is worth giving a real chance, especially when used correctly with the separate web versus work context.

The team also covered the organization's evolving AI governance model. The most important policy direction was the emerging red, yellow, green framework for AI data usage. Public data can be used more broadly, internal business data should stay within approved enterprise tools, and sensitive or regulated data such as PII and health information should not be entered into AI tools at all without special handling. This aligns well with the way I think about risk, compliance, and practical enablement - allow useful adoption, but keep clear controls around privacy, security, and exceptions.

The second half of the conversation became more use-case oriented. Marketing shared that Adobe AI is already heavily embedded in their image editing workflows, while ChatGPT tends to outperform for more creative image generation. Jordan described potential sales use cases around field visit summaries, Salesforce updates, and retailer communication workflows. That led to a productive follow-up discussion on how AI and automation could help reduce administrative burden for the sales team, especially around store outreach, reorder prompts, and visit documentation. Shawn indicated that expanding into more departmental use cases is a Q2 objective, and I think sales is a strong candidate for targeted pilot work.

Overall, the session struck a good balance between technical grounding, governance, and practical business application. My view remains that AI is not replacing jobs in the near term, but it is already changing how we work and where we spend our time. The organizations and employees that build the right skills now will be in a much better position as the tools mature.

2. Attendee List

  • Shawn Remick
  • Jordan Haire
  • Gage
  • Scott Warner
  • Chip Otte
  • JKLN

3. Action Items

  • [Shawn Remick / Gage] Continue rolling out AI training, with Copilot Enterprise deep-dive training planned as the next session.
  • [Shawn Remick / Gage] Provide Jordan Haire with a Copilot license after he sends the request email.
  • [Jordan Haire] Email Shawn Remick and Gage to request Copilot access.
  • [Shawn Remick / Gage] Follow up offline with JKLN regarding why SAP for Analysis files are not saving to OneDrive.
  • [Shawn Remick] Follow up offline with JKLN and Chip Otte on image background removal options, including whether credits remain on the previously tested background removal tool.
  • [Shawn Remick / IT] Continue updating the AI policy to the red, yellow, green data usage model.
  • [Shawn Remick / IT] Build and maintain a register of approved AI exceptions and known external tools where company data is allowed by design.
  • [Shawn Remick / IT] Continue evaluation of departmental AI opportunities in Q2, including additional department outreach and idea gathering.
  • [Shawn Remick / Jordan Haire / Neil] Schedule a deeper discussion on sales-focused AI and automation use cases when Jordan is back in town.
  • [Jordan Haire] Identify a sales super user who could pilot field-based AI workflows such as visit summarization and faster Salesforce updates.
  • [All licensed users] Notify IT if AI licenses are no longer being used so licenses can be reclaimed.
  • [All attendees] Share useful tips, tricks, and use cases in the Microsoft Teams AI group.

4. Relevant Timelines

  • Q2:
  • IT plans to engage additional departments to collect more AI use cases and expand pilot efforts.
  • Sales use case exploration should be picked up as part of this broader Q2 expansion.
  • Later this year:
  • IT expects to evaluate whether the current approved AI tool set should be expanded or modified.
  • Additional training resources and more focused use-case sessions are expected to roll out.
  • This week:
  • Jordan Haire to request Copilot access.
  • Shawn Remick and Gage to provision Jordan's Copilot license once requested.
  • June:
  • Microsoft Copilot pricing is expected to increase again, which reinforces the need to reclaim unused licenses.

5. Additional Notes

  • Shawn framed AI strategy around two lanes:
  • Generative AI for individual productivity
  • Technology-enabled transformation using machine learning, RPA, and automation for company-level process improvement
  • Vendor skepticism remains important. A recurring theme was that many vendors market products as AI-enabled without meaningful AI capability.
  • Model selection matters:
  • Copilot is generally better suited for enterprise, fact-based, internal work
  • ChatGPT is generally stronger for creative ideation and image generation
  • Different models behave differently, and users should choose based on the task
  • The distinction between Copilot basic and enterprise is significant:
  • Basic handles general LLM/web use
  • Enterprise can interact with company work context including Outlook, Teams, OneDrive, and Microsoft apps
  • I reinforced to the group that Copilot should be given a fair chance, especially since Microsoft is likely to remain the strategic ecosystem for future enterprise integrations and tool development.
  • I noted one caution that I still do not trust AI to do my Excel work without validation.
  • Marketing is already using AI regularly for:
  • Image cleanup and editing in Adobe
  • Campaign concepting
  • Copy refinement
  • Creative experimentation across Adobe, ChatGPT, and Copilot
  • Sales surfaced several possible AI opportunities:
  • Summarizing customer or dealer visits immediately after meetings
  • Reducing Salesforce admin effort
  • Automating or improving reorder communications to stores
  • Identifying stores with low recent order activity
  • Longer term, potentially supporting planogram or shelf-audit style workflows
  • A broader organizational theme was skill development. The team does not view AI as an immediate job elimination tool, but rather as a capability shift that will reward employees who learn how to use it effectively.
  • Shawn mentioned the IU generative AI course is worthwhile, especially for prompt engineering, although it leans more toward Gemini than our approved enterprise tools.
  • The training itself was well received. Feedback from the group was that it was a good balance and not overly technical for the audience.