Just Try This #13 — Not Every Question Needs the Biggest Brain
Most of us now have more than one AI model in front of us. A fast one. A thinking one. Sometimes a dropdown full of names we don't recognize. And the easy button “auto”. Most people I talk to never change it. They just figure Auto knows. But it really doesn’t.
I think we skipped a step. Model selection is becoming a basic AI literacy skill. The question isn't which model is best. It's which one fits the work.
I've recently started asking two questions before I decide.
How complex is the work?
How consequential is the answer?
Rewording a note to a colleague? Low on both. Use the faster model and move on.
Pulling a number that's going into an executive update? Simple, but it matters. Fast model, then verify.
Brainstorming options for a new program? Complex, but a bad idea won't hurt anyone. Give it deeper reasoning.
Shaping a recommendation leadership will act on? Complex and consequential. Use the strongest appropriate model, deep reasoning, and your own review before it goes anywhere.
Just try this: before your next five AI prompts, don't start by asking, "What should I ask AI?" Ask, "How much intelligence does this work actually need?"
The model can do a lot. Deciding how much it should do is still a human call.
Gina Schudi
September 2026
Strategic Thinking in a Complex World
Just attended Strategic Thinking in a Complex World (timely!) course and the first in which Copilot was part of the curriculum. Excellent course and a beautiful time to visit Truist Leadership Institute in Greensboro about two hours from Charlotte. The networking with colleagues across the enterprise is such a huge benefit! Such a talented group of leaders and new friends. Heather Williams Alison Lyewski Deepak Kulkarni Molly O'Neal and many more!
Gina Schudi
September 2026
Just Try This #12 — The 3–2–1 Agent Check
If you’re building an agent, don’t try to make version one do everything.
Start with a 3–2–1 check.
3 QUESTIONS
Write down three real questions you would expect someone to ask this agent.
Not the easy demo questions. Pick the questions that would make you say, “If it can handle these well, this could actually save me some time.”
Then use those same three questions to test it.
2 TRUSTED SOURCES
Give it two places where you know the right information lives.
A policy. A procedure. A playbook. A project folder. A reference document.
Keep the source set small at first so you can actually tell whether the agent is finding and using the right information.
1 BOUNDARY
Decide on one thing the agent should not do.
Maybe it shouldn’t make a decision. Maybe it shouldn’t interpret something when the source isn’t clear. Maybe it should stop and tell you it doesn’t have enough information. This one matters.
A useful agent doesn’t just know how to answer. It needs to know when not to answer.
Now build it.
Ask your three questions. Check the answers against your two sources. Make sure it respects your one boundary.
If it works, add more.
That’s the part I think we sometimes overcomplicate with agents. You don’t need to design the perfect agent on day one.
You need to give it one useful job and a small test it can pass.
And the people closest to the work are often the best people to do that. They already know the questions people ask, where the trustworthy information lives and where judgment is required.
Just try this: Think of one repetitive task in your job and write down your 3 questions, 2 sources and 1 boundary.
You now have the beginning of an agent.
Gina Schudi
September 2026
Just Try This #11 — Stop Scaling AI to People. Start Scaling It to Roles.
We have spent a lot of time teaching individuals how to use AI. Prompt better. Build an agent. Automate a task.
That is a great place to start. But I don’t think it is how AI ultimately scales across an enterprise.
The bigger opportunity is the role.
Think about a relationship manager, an underwriter, a project manager or a financial analyst. There are hundreds or thousands of people doing versions of the same work, following many of the same workflows, making similar decisions and producing similar outputs.
What if instead of asking every individual to figure out how AI fits into their job, we redesigned the highest-value workflows for the role?
Now we aren’t just finding isolated productivity wins. We are defining a better way of working that can potentially be reused across an entire population.
And importantly, that doesn’t mean eliminating human judgment or making every job identical. It means figuring out what should be consistent, what AI can handle, and where the human brings the most value.
That feels like the next stage of AI adoption to me:
Task → Workflow → Role.
Just try this: pick one role with a large employee population. Identify the five workflows that consume the most time, then ask what those workflows would look like if they were designed today with AI available from the beginning.
That is where scaling starts.
Gina Schudi
September 2026
Just Try This #10 — Your Employees Have AI. Your Management Model Doesn’t.
Your employee uses AI exactly the way you asked them to. They research faster, pressure-test the thinking, improve the work, and put a really solid deliverable in front of you. Something that used to take two days now takes two hours.
Then it sits in your inbox for a week. And when it finally gets reviewed, there are two sentences of feedback.
I think we may have taught managers that reviewing work means finding something to change. If there are no comments, it can feel like we didn’t add value. But in an AI-enabled workforce, that mindset can become the bottleneck.
The value of management is not in putting red ink on someone else’s work. It is in setting direction, applying judgment where it matters, removing barriers, developing people, and knowing when the work is good enough to move.
That also means employees need to make the handoff easier. Instead of sending something up with a vague ‘thoughts?’ or ‘can you review?’, be really clear about what you need.
Just try this: When you send work to your manager, tell them three things: what you need, what you want them to review, and when you need the decision.
Is this for awareness, feedback, or approval? Are you asking them to look at the risk, the recommendation, or the overall direction? And when does the work need to move?
That alone can eliminate a lot of unnecessary review saving the manager time and accelerating progress and innovation.
And managers, before you start adding comments, ask yourself one question: Will my feedback materially improve the outcome, reduce a real risk, or change the decision?
If the answer is yes, absolutely add it.
If the answer is no, ‘Looks good. Send it.’ is a perfectly good management response. Maybe even a great one.
Because workforce transformation is not just about teaching employees how to work differently with AI. It is about helping managers learn how to manage differently too.
Gina Schudi
September 2026
Just Try This #9 — Am I Going to Have to Explain This to AI Again?
I’ve realized I do something with AI that is probably a little ridiculous.
I’ll have a great conversation, give it all the context, explain exactly what I need, work through the details, finally get to a great result… and then a week later I’ll come back and do basically the same thing all over again.
That is a lot of time spent teaching AI the same thing twice.
And once I started noticing it, I saw it everywhere.
Preparing for a recurring meeting. Creating the same kind of update. Reviewing work against the same criteria. Researching the same kinds of questions. Giving AI the same background information every single time before it can actually help.
This is where agents can save you a LOT of time.
If you have already figured out how to get AI to do something well, you shouldn’t have to rebuild that conversation every time you need it. Give an agent the instructions, the right sources, the steps to follow and what a good result looks like — and let it remember the process for you.
It doesn’t have to be a huge automation project either. Start with something you do over and over again.
The question I’ve started asking myself is pretty simple:
Am I going to have to explain this to AI again?
If the answer is yes, there may be an agent hiding in that work.
Just try this: Think about one AI conversation you’ve already had more than once. Open Agent Builder and turn it into something reusable.
Because saving ten minutes once is nice.
Saving ten minutes every single time you do the work?
That adds up fast.
Gina Schudi
September 2026
Just Try This #8 - Protect the Thinking Reps
We are teaching everyone how to use AI. But who is making sure we are still teaching people how to think?
I keep coming back to that question because AI is getting extraordinarily good at doing the work that used to develop us. The first draft. The analysis. The research. The presentation. The hours spent wrestling with a problem before you finally understood it.
We should absolutely use AI to eliminate unnecessary effort. I do it every day. But some of that effort wasn’t waste. Some of it was how we learned judgment, developed expertise and eventually became the people others relied on when the answer wasn’t obvious.
That makes me wonder whether one of the next big workforce problems is hiding inside one of AI’s greatest benefits.
We may be able to give a relatively inexperienced employee the ability to produce something that looks remarkably senior. That’s incredible leverage. But producing senior-level output and developing senior-level judgment are not the same thing.
And I don’t think most companies have figured out what replaces the developmental experiences we’re beginning to automate away.
This is where I think HR, business leaders and AI leaders may have a much bigger job ahead of them than simply teaching people how to use the tools. We may need to deliberately design the experiences that build curiosity, judgment, critical thinking, expertise and the confidence to challenge an answer — including one generated by AI.
EY recently drew attention to the value of distinctly human thinking, and I suspect we’re going to see much more of that. As machine intelligence becomes abundant, organizations may have to become far more intentional about developing human intelligence.
So here’s the question I’m wrestling with: If AI allows junior employees to produce senior-level work much earlier, how will we make sure they still develop senior-level judgment?
I don’t think we have the answer yet, but we should start designing for it now.
Just try this: Look at one piece of work your organization is excited to automate with AI and ask a second question: What did doing this work used to teach someone?
That answer may matter more than we think. How are you thinking about this?
Gina Schudi
September 2026
Just Try This #7: HR, This Might Be Your Moment
AI transformation is quickly becoming workforce transformation.
And that means HR has a much bigger opportunity than helping people adopt the technology.
As AI removes tasks, compresses work and changes how decisions get made, someone needs to ask a more important question:
How do we use this moment to improve the human experience of work?
What happens to career paths when expertise can be augmented instantly?
What happens to learning when AI can teach in the flow of work?
What happens to performance when one person can suddenly do far more?
What happens to meaning, contribution and growth when the work itself starts changing underneath us?
This is where I think HR can lead.
Not just by managing change.
By helping redesign work so people gain something too.
More time for meaningful work. Faster learning. Better opportunities to grow. More creativity. Stronger judgment. Less friction. More room for the things humans do best.
Because if AI creates enormous value for the enterprise but the human experience gets worse, we missed something important.
The real opportunity is to redesign both at the same time:
Better work for the company.
A better experience for the human.
That feels like one of the most important leadership opportunities HR has had in a long time.
Just try this: The next time AI creates an opportunity to remove work, don't stop at asking what the company gains.
Ask a second question:
How could this make work better for the human too?
Gina Schudi
September 2026
Just Try This #6: Get Smarter Faster by Talking to AI
One of the most underrated ways to use AI is also one of the simplest:
Talk to it.
Not just to get an answer. Use it to pressure-test an idea, challenge your assumptions and expose what you may be missing.
Say you have an idea for a new workflow and need to present it.
You could ask AI to build the deck.
Or you could use it to make your thinking better first.
Describe the workflow. Explain what is broken today. Ask where the idea could fail. Ask what your audience will challenge. Ask what evidence would make the case stronger. Then ask how to tell the story more clearly.
A few rounds later, you may still have the same idea.
But now you understand it better.
You can defend it. Refine it. Explain it. And probably improve it.
That is the part of AI I think we undersell.
The goal should not be to outsource human thinking. It should be to amplify it.
AI can give us more information, more perspectives and more possibilities.
But humans still bring the judgment.
We decide what matters. We recognize nuance. We imagine what should exist next. We make the call.
That human thinking is the secret sauce.
The real promise of AI is not a world where machines do more of our thinking for us.
It is a world where more humans have the tools to think deeper, learn faster and make better decisions.
And that could make all of us better.
Just try this: The next time you have an idea to develop, don't hand your thinking to AI. Bring AI into your thinking. Let it challenge you, stretch you and sharpen your point of view — then make the call yourself.
The future doesn't need humans who think less because of AI. It needs humans who think better because of it.
Gina Schudi
September 2026
Just Try This #5: Creating Human Capacity
Just Try This #5 - Creating Human Capacity - old school tricks.
I keep thinking about one of the stranger contradictions of the AI era.
Companies used to buy the coffee. Then we got sophisticated. The coffee expense got allocated across cost centers. The cost centers started questioning why they were paying for it. Eventually, in a lot of large companies, the coffee disappeared.
Well, technically, the coffee didn't disappear. We kept the machines.
We just stopped stocking them.
And now companies simultaneously spend billions of dollars on AI to make those same employees more productive.
Copilots. Agents. Automation. Workflow redesign. Capacity creation.
We fund the models. We'll fund the infrastructure. But now humans bring the coffee. The tea. The creamer. The sugar. The cup. The stir stick.
I mean….that's kinda funny.
Here's what I believe. The companies that get AI right will be the ones that actually remember the human, improve their relationship with the human - while they work to build the machine.
People need energy.
They need time to think.
They need other people to learn from and have relationships with.
And they need small signals that say, we actually thought about what would make your day a little better.
Coffee is not a strategy, but it's a strong signal.
It says you matter enough that we didn't optimize you out of the kitchen. And yeah - we're guessing you might actually be a little more productive with an afternoon boost.
Maybe the future of work isn't entirely new. Maybe it's the best of what we already knew about humans, augmented by everything we're learning about machines.
The future may run on AI, but the people building it still run on coffee.
Just try this.
Be kind to the humans.
Just Try This #4: One for the Strategy People
Just Try This #4 - one for the strategy people.
Over the past year, working as an AI transformation leader in banking, one thing has become very clear to me: we have a time-to-insight problem.
Quite a bit of AI strategic thinking is made public. McKinsey. EY. BCG. Deloitte have published research sitting in plain sight. What we lack isn't access - its the hours to find it, synthesize it, and connect it to the decision in front of us.
I use Claude to pull that public research, and I ground my thesis in work they have already published.
That has saved the enterprise thousands in consultant fees and it has changed how I am able to show up in the room with an already grounded thesis.
When I'm writing a position paper, I don't start from a blank page and a strong opinion. I start from what the major firms have already put in market, then I apply it to our context. I walk in with published research behind the point of view, not just my take that week.
Here's the move I have been using and share with teams:
Ask Claude to find the last 12 months of public reports from McKinsey, EY, BCG, or Deloitte on a specific strategy question. Not "research AI." A real question. Then ask for the title, date, source, and what would actually matter for this organization (which later lands in my appendix).
Then do the human part. Read the original on anything you'd stake a decision on. Write the position. Create the deck. Walk into the meeting with a fully prepared and grounded thesis.
This is what Human First looks like in practice. AI compresses the search. I own the judgment.
The expensive work is still the expensive work: our data, our culture, our trade-offs. The public thinking is already on the table. I use it before the meeting, not after.
Strategy people: what are you using AI for this week that used to take a longer cycles?
Just Try This #3: AI Troubleshoots the Boat
Just Try This #3: Let AI troubleshoot with you and save the day!
My husband and I were getting ready to take the boat out when it started making that sound. Relentless beeping that tells you something is wrong, while providing absolutely no useful information about what that something might be. The boat was loaded. Our day was at risk.
So I pulled out my phone. Not to Google the alarm code. Not to search boating forums. I opened ChatGPT. I recorded a video so it could hear the alarm, then started taking pictures of the gauges, the controls and the equipment it asked me to look at.
I didn't give it one brilliant prompt and magically get the answer. We worked the problem. It would tell me what the sound might indicate and what to look for next. I would send another picture. At one point I'm holding my phone over equipment I couldn't have named five minutes earlier, essentially saying, 'Okay, this thing?' Yep. That thing. Now check this.
Eventually we isolated what the alarm was actually telling us, dealt with the problem, and moved on with our day. I wasn't asking AI to give me an answer. I was giving it eyes and ears and letting it help me investigate something I didn't understand. That's a different skill.
The next time something in the physical world has you stumped, show it what you're seeing. Take the picture. Record the video. Let it hear the sound. Then tell it: 'Help me troubleshoot this one step at a time.'
Don't hand your judgment over to AI, particularly when safety is involved. But don't underestimate what happens when you stop treating AI like a search box and start letting it work the problem alongside you.
Sometimes the biggest leap in using AI isn't learning a more advanced prompt. It's realizing you don't have to know enough to ask the perfect question before you start.
Just Try This #2: Make Copilot Argue Against You
Today something for my advanced Copilot users: try this before your next important meeting.
Most people get pretty good at using Copilot to create — draft the email, summarize the meeting, tighten the deck. Creation is only one role.
Take a recommendation you're planning to put in front of leadership. Before you finalize it, give Copilot the material and tell it: 'Assume you disagree with my recommendation. Build the strongest case against it. Identify the assumptions I'm treating as facts, the risks I'm underweighting, and the three questions a skeptical executive should ask me.'
Then don't immediately defend your work. Read the response and ask: 'Which of these objections is strongest, and what evidence would I need to prove or disprove it?'
Now you're not asking Copilot to make your work prettier. You're using it to expose where your thinking may not be ready yet. Copilot isn't the judge. You are. But if it finds one assumption you haven't tested, one stakeholder you haven't considered, or one question you don't have a good answer for yet — that could change the meeting.
The more advanced the user, the less interesting 'write this for me' becomes. The better question is: where can AI make my thinking stronger before I ship it?
Just Try This #1: Pricing a Lake House From the Pontoon
JUST TRY THIS | #1 — For the person who keeps hearing about AI but still isn't quite sure what to do with it.
This weekend I was out on a lake in Minnesota, sitting on the pontoon, when we passed this house being built. I loved it. And my immediate thought was: Hmmm… I wonder what it would cost to build something like that?
So I took a picture. Instead of going home and disappearing down a Google rabbit hole, I just started asking AI questions. Right there. From the pontoon.
From one photo, AI estimated a modern lake-style home around 2,400–3,000 square feet, pointed out the glass, multiple rooflines, the walkout and the complexity of a waterfront site, and gave a very rough starting range of about $750K–$1.3M+ for the house itself. It was also clear about what it didn't know. That might actually have been the most useful part.
Would I call a builder tomorrow and tell them AI says my house should cost $900K? Absolutely not. But I now know what questions to ask. And I did all of this while still sitting on the pontoon enjoying the sunshine.
AI doesn't always need to begin with some complicated business problem. Sometimes it starts because you see something that makes you curious and think, 'I wonder…' Just try this.
ChatGPT and Claude Argued Over My Basement
I've been working on finishing roughly 2,000 square feet of our basement, and what started as a home project has become one of the clearest examples I've had of how AI can elevate the human in the work.
I started with an idea and very little expertise, so before I hired a designer I used AI to help me understand what was possible. Then I hired a designer, but I kept AI in the process. I would take a design into ChatGPT, ask it to evaluate the layout, then move the same problem to Claude. Sometimes I would have one model critique the other. Basically, I had ChatGPT and Claude arguing over my basement while I sat in the middle deciding who made the better case.
And there were some spectacular misses. One layout blocked access to the shower with the toilet. Another ignored a window I had already said could not move. Then came a pocket door apparently opening directly to the outside. My response: 'Ugh. THAT'S AN EXTERIOR DOOR.'
But every bad iteration forced me to get more precise. By the time we landed on a design that worked, I understood the tradeoffs and could engage the professionals from a much more informed position.
AI didn't make me a designer. It made me considerably better at working with one. That is Human First AI. Transform the work. Elevate the human.
It Starts With a Cabin in the Woods
I had one of those little AI moments this weekend that reminded me why I care so much about AI enablement.
My friend's husband is about as Minnesota as it gets — hunter, fisherman, outdoorsman. He doesn't need AI because someone tells him it is going to transform the world. But give him a tool that helps him visualize what his cabin could look like before he starts the next project? Now we have his attention.
I'm becoming more convinced that the first job of AI enablement isn't education. It's helping someone find their reason to care.
Because enterprise AI adoption will not be won by the people who are already excited about AI. They are already coming. The harder — and more important — work is reaching the person who doesn't think AI has anything to do with them yet.
That is how you bring a workforce along. Not one technology deployment at a time. One person at a time. And sometimes, it starts with a cabin in the woods.
A Different Kind of Independence
Most organizations treat AI Enablement as a function. Training. Communications. Governance. Change management… But what if we're looking at it the wrong way? What if AI Enablement is actually an enterprise product?
Product teams obsess over understanding users, reducing friction, driving adoption, and continuously improving the experience. AI Enablement is beginning to look remarkably similar. There's one important distinction, though. The product isn't AI. The product is human capability.
Maybe our customers aren't business units. Maybe they're employees. Maybe our roadmap isn't about deploying more AI. Maybe it's about empowering more people.
Accelerating the Dream-Build Loop
Once you start doing this you will be hooked!
Start small — see what you can dream-build with AI in a 45 minute walk/chat. Post what you're working on… no ideas are too small!
Leadership in a Year of Unprecedented Change
One of the most valuable reminders from Mastering Leadership Dynamics is that leadership is about influence, trust, adaptability, and the ability to bring people together around a shared vision.
As we navigate unprecedented change driven by AI, digital transformation, and evolving workforce expectations, the human side of leadership becomes even more important. The ability to listen, build alignment, lead through ambiguity, and create environments where people can thrive will continue to separate good leaders from great ones.
The Gap Is No Longer Access
It's clear now… the gap is no longer access — it's capability. And capability doesn't build evenly.
It comes from people and from experience, not training. From iteration, not instruction. Once people hit that “unlock,” in their heads things start to shift: work moves faster, new questions get asked, opportunities to rethink processes become more visible.
That's where meaningful value begins to show up. So do you let that capability evolve unevenly — or do you help scale it more intentionally?
And at the center of it all — it's people. People will unlock the change. People will orchestrate the value. People will define the boundaries. People will maintain the trust.
AI Isn't Failing Because of the Technology
Over the past year, working in AI business adoption, one thing has become very clear to me: AI isn't failing in enterprises because of the technology. It's failing because we haven't prepared our people to use it.
AI Business and Workforce Adoption doesn't come from big launches. It comes from small moments: someone trying their first prompt. A team sharing a real use case — the small ones resonate quickly. Curiosity replacing hesitation and building confidence.
That's when it starts to click. AI transformation isn't a technology rollout. It's a workforce shift — and most companies are still early. But not for long.
Two Prompt-a-Thons, 70 Associates
Today we ran two high-energy prompt-a-thons across 70 associates — and the results were outstanding.
What stood out wasn't just the creativity, but the speed at which teams moved from idea to working prompts that can actually drive impact. This is what real AI adoption looks like: hands-on, practical, and driven by the people closest to the work.
This wasn't just a hackathon. It was a signal of how quickly capability is building across the organization.
What the Atlanta Prompt-a-Thon Proved
Delivered our Atlanta Prompt-a-thon this past week — and the signal is clear: when you put the right tools in people's hands, capability scales fast.
Over a focused session, teams worked directly in Microsoft Copilot across M365 — moving quickly from basic prompts to real business use cases. It was practical, immediate, and measurable. Real workflows improved in hours, not weeks.
This is how adoption happens — hands-on, business-led, and outcome-focused.
One Year In
From leading a team advancing Strategic Remediation in CT&O to now driving AI enablement across the enterprise, this journey has been centered on transformation, scale, and impact.
Equally important has been establishing a leadership voice around AI — bridging strategy and execution, simplifying the complex, and helping teams navigate what's next.
Women in Tech, Charlotte
Enjoyed my first visit to NC Tech. As a new resident of the Charlotte area, it's a wonderful perk to connect with other women leaders just a few minutes from the office. I teamed up with a few local Truist leaders and supported Jennifer Hewit on a panel discussing tech, leadership, AI, and navigating career pivots.
One topic that particularly resonated with me was ruthless prioritization.
