AI Governance matters - want proof?

The temptation to hook AI up to your company’s data platform or an Excel file is hard to resist. Everyone can ask questions and get answers without waiting for support. But what happens when everyone uses a different definition of revenue, an active customer, or growth? Without shared definitions and clear rules for using the data, you can get sideways quickly. Faster answers are only useful if everyone agrees on what’s being measured. This Reddit post describes what can happen when access gets ahead of governance: Read the discussion. Before giving everyone an AI analyst, make sure everyone is working from the same definitions.

Microsoft Copilot Has 4 Different Ways to Work With Excel

Microsoft Copilot is becoming much more useful for Excel, but the different tools can be confusing.

Here is the simplest way to think about them.

Analyst Agent

Purpose: Analyze data you attach and turn it into insights, charts, comparisons, and reports.

Think: “Here is my data. Tell me what is happening.”

Excel Agent

Purpose: Create a brand-new Excel workbook from a prompt.

Important distinction: Excel Agent lives inside the Microsoft Copilot app. You are not inside Excel when you start.

Think: “Build me a workbook.”

Copilot Inside Excel

Purpose: Work directly on the workbook you already have open. It can help with formulas, formatting, PivotTables, charts, sorting, filtering, and analysis.

Think: “Help me work on this workbook.”

Copilot Skills

Purpose: Save detailed instructions for tasks you perform repeatedly so Copilot can follow the same process again.

Think: “Do this job the same way every time.”

The big takeaway: Analyst investigates. Excel Agent creates. Copilot in Excel edits. Skills repeat.

AI Is Coming for Data Catalogs. That Is a Good Thing.

Google’s Knowledge Catalog Enrichment Agent points to a bigger shift in enterprise AI: data catalogs are becoming active, intelligent systems instead of static filing cabinets.

The tool is a command-line AI agent that helps generate Metadata as Code for Dataplex Knowledge Catalog. In plain English, it can read source material like documents, BigQuery schemas, SQL patterns, user feedback, and GitHub repositories, then turn that context into structured descriptions for data assets.

That matters because most companies have a metadata (data about their data) problem. Teams collect huge amounts of data, but the meaning of that data often lives in scattered documents, old queries, tacit knowledge, or code. When metadata is missing or weak, people waste time asking basic questions: What does this table mean? Can I trust this field? Who uses this data? Is this metric still correct?

Google built something to help close that gap using AI

The important part is not just that the agent writes descriptions. It creates reviewable drafts. Humans can inspect, refine, evaluate, and decide when to publish them. That is the right pattern. AI should not silently rewrite your data catalog. It should draft, explain, and prepare work for approval.

Better metadata makes analytics faster. It makes AI systems safer. It helps employees find the right data and avoid misusing the wrong data. As more teams build AI agents on top of internal data, clean context becomes a competitive advantage.

What should you do about it?

1. Audit your most important datasets. Find the tables that people use often but do not fully understand.
2. Gather the context around them. That includes docs, dashboards, SQL examples, business definitions, and feedback from users.
3. Test AI-assisted metadata generation on a small, high-value area before rolling it out broadly.
4. Keep humans in the loop. Treat AI-generated metadata like a strong first draft, not the final truth.

I have always been a fan of data catalogs. They are one of the places where that context will either be organized, or become a bottleneck. Good on Google for building this!

Today’s Practical AI Tip: Let AI Catch Your Forgotten Drafts

Today’s Practical AI tip comes from a mistake I made recently.

I wrote what I thought was a great email, got pulled away, and never sent it. Then I sat there wondering why the other person had not responded. Eventually, I realized the problem was not their inbox. It was my drafts folder.

When I shared the story with my latest Copilot cohort, I found out I was not alone. Several people had experienced that awkward, “Didn’t you get my email?” moment, only to discover the message had never been sent.

So, we turned the mistake into a simple AI workflow.

I created a scheduled prompt in Copilot that checks my drafts every two days. It looks for messages that appear ready to send, tells me who they are addressed to, and recommends whether I should send or delete them based on my sent items and the surrounding context.

This is a great example of practical AI. It is not about replacing your judgment. It is about creating a safety net for the small mistakes that cause confusion, delays, and embarrassment.

Sometimes the best AI use case is not flashy. It is simply making sure a forgotten draft does not become a missed opportunity.

Today's Practical AI Lesson: PowerPoint Edition — Let Me Count the Ways

I found seven (!!!) different ways to create PowerPoint presentations using Copilot. Yes seven, and yes they all created different outputs.

Two live inside PowerPoint itself. Five can be found/created inside Copilot on the Web. The outputs ranged from unusable to what I will officially rate as “not bad.” The best results inside the PowerPoint app itself came from using “New Slide with Copilot” button in the toolbar (pictured below). The worst output across all seven options came from clicking the Copilot icon on the toolbar in PowerPoint and asking for a full presentation in the chat window sidebar, weird, right?

The very best results came from a new PowerPoint agent currently in beta. If you have early access, go to the web version of Copilot, click Frontier (left side , and look for PowerPoint Presentation. You need the paid version of Copilot and may need to speak to your O365 administrator to turn this on. This Powerpoint agent asked clarifying questions, planned the structure, and only then generated slides. Same prompt. Same underlying AI. Completely different outcomes.

This is why giving teams Copilot licenses does not automatically improve processes. Interface choices matter. Planning matters. Training matters.

If you want to go deeper into how to actually use Copilot in real workflows, join me on February 27th for my all-day Copilot program with Fisher College of Business Executive Education program where we dive into this and more!!

#ArtificialIntelligence #MicrosoftCopilot #Copilot #AITraining #AIInBusiness #OSU HappyAIPath

#GenerativeAI #AIProductivity #DigitalTransformation

#ExecutiveEducation #BusinessAnalytics #FutureOfWork

Brands Are Building ChatGPT Apps — And I Told You They Would

Today’s practical post: an “I told you so” moment.

Two months ago, I wrote that brands should start building apps for ChatGPT. I thought it would take longer to matter — but on that point, I was wrong.

In just a few weeks, major companies have launched ChatGPT apps that let consumers interact directly with their products and services through AI. This isn’t theory anymore — it’s happening:

• Target — ChatGPT app live. Discuss design preferences and see results.
• Zillow — find homes through AI.
• Instacart — plan groceries by conversation.
• Uber — book rides via ChatGPT.
• StubHub — discover events directly inside chat.

What brand do you think should build a ChatGPT app next? The shift is happening faster than most expected, and it’s reshaping how people interact with brands.

According to new research from PYMNTS.com, based on a survey of ~2,100 consumers:

• Over 60% of consumers used a dedicated AI platform in the past year — AI is mainstream.
• More than one-third of Gen Z and “Power Users” now start their personal tasks in AI, not search.
• 42% of those users rely on traditional search engines less often.

This isn’t a gimmick. Forget “SEO for AI.”

Integration into AI — with customizable, frictionless, conversational transaction enablement (phew, what a phrase 😄) — will be the expected starting point for consumer engagement in less than a year.

People are talking to AI and expect brands to talk back.

Follow HappyAIPath for more practical AI insights on how it’s reshaping the way companies work — in real time.

P.S. All dashes were made of my own free will 🤣

Microsoft Ignite 2025: Copilot Becomes the Manager

Microsoft used Ignite 2025 to make one thing clear. AI is moving from the background to the center of how we work.

Copilot is no longer an add-on. It is the system. In 2026, about 80% of Office 365 enhancements will involve Copilot. That means AI will not just draft your email or summarize a meeting. It will schedule your calls, manage your inbox, and prioritize your day before you even open your laptop.

Agent 365 ties it all together. It gives every AI agent an identity and guardrails so companies can trust automation without losing control.

The message from Microsoft is simple. You will not just work with AI. You will work through it.

Labels: AI, Microsoft, Ignite 2025, Copilot, Office 365, AI Agents, Productivity