Leveraging Leadership
Leadership is messy. Most advice isn't built for the reality of competing priorities, difficult stakeholders, limited time, and imperfect information.
Leveraging Leadership is a practical leadership podcast for Chiefs of Staff, executives, founders, and senior operators who want to lead more effectively and navigate complexity with confidence.
Hosted by Emily Sander, former Chief of Staff and executive advisor, each episode delivers real-world lessons, practical frameworks, and candid conversations with leaders across business and beyond.
Topics include executive communication, leadership presence, decision-making, delegation, organizational influence, operating rhythms, team effectiveness, and the often-unspoken challenges leaders face behind the scenes.
If you're looking for thoughtful conversations, practical takeaways, and leadership advice you can actually use on Monday morning, you're in the right place.
Leveraging Leadership
Q&A: Balancing AI Adoption and Quality Leadership Strategies for a Maturing Tech Team
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Emily Sander answers a listener’s question about moving a tech company to AI-first, highlighting challenges when people use AI excessively or without checks. She shares practical ideas like approved use cases, tiered approval for AI spend, and the need for quality controls. Emily Sander also points out the importance of balancing AI experimentation with making sure standards and outcomes stay high.
Get in touch with Emily:
Want help leading your organization through AI adoption? Book a clarity call w/ Emily here.
Who Am I?
If we haven’t met before - Hi👋 I’m Emily, Chief of Staff turned Executive Leadership Coach. After a thrilling ride up the corporate ladder, I’m focusing on what I love - working with people to realize their professional and personal goals. Through my videos here on this channel, books, podcast guest spots, and newsletter, I share new ideas and practical and tactical tools to help you be more productive and build the career and life you want.
Time Stamps:
00:25 Listener Question Setup
01:20 AI First Culture Shift
02:11 Experimenting vs Standardizing
03:34 Trust but Verify
05:05 Token Spend Guardrails
06:45 Quality Checks and Controls
07:39 Center the Pendulum
Welcome back to Leveraging Leadership, where we unpack the art of business leadership. I'm your host, Emily Sander, chief of staff to an executive leadership coach. This show is all about finding your points of greatest influence and leveraging them to better serve those around you.
Listener Question Setup
All right, listener question time. So Robert, ooh, I'm so sorry in advance, Robert. Taglonionly, Tagon, Taglioni, uh, from Atlanta says, "I work at a mid-size technology company, and over the past year we've undergone significant leadership and culture change. Several long-tenured executives have departed, and many, many members of our leadership team, including our CTO, have joined within the last six to 12 months. Our new leadership team is extremely bullish on AI. Recently, the CTO had our engineering team spend two weeks immersed in Claude and other AI tools, and leadership team has made it clear that AI adoption is a strategic priority. The CEO is fully supportive of becoming an AI-first organization, although he tends to focus more on the vision and business outcomes than the day-to-day realities of implementation.
AI First Culture Shift
Here's the challenge. While I absolutely agree we need to embrace AI, it feels like we've swung from one extreme to the other. Some employees are using AI thoughtfully and productively. Others seem to be using it for everything, including tasks where it doesn't make sense. We're seeing hallucinations, poor quality outputs, people skipping validation steps, and work products that sometimes feel less reliable than before. How should leaders think about managing this transition? How do you encourage experimentation and adoption without creating a culture where AI is used indiscriminately? And how do you help a leadership team move from AI enthusiasm to a more mature, balanced operating model?" Excellent question. Many chiefs of staff, many executives are dealing with this exact same scenario or something very similar. Okay.
Experimenting vs Standardizing
So first of all, I think there are important distinctions between experimenting with AI and standardizing AI. So between experimentation and standardization, put it that way. So you can encourage a team and a company to explore and become more knowledgeable and all these things and not have that fully baked in your required processes. So just-- I'm not saying one is better than the other or which one to do in your particular scenario. I just wanna throw out that distinction. Another important distinction, in reading your question here is How do I want to put this? AI, like, knowledge, like literacy, like how conversant you are, how trained you are in AI, and being dependent on it. So AI literacy and AI dependency. If you're fully dependent on AI, meaning we can't do anything without AI, where we've lost our ability to think critically ourselves, that's, that-- to me, that's a problem. So just have those distinctions in mind as we're talking about this. I think a lot of people are just like, "AI!" And they push it, and they're like, "Use AI, and use it a lot, and use tokens." And it's like, okay, but I can do that, but the outcomes and results are totally different than how you're pushing me to use this.
Trust but Verify
So I would, I would... There's that old saying, uh, from, like, the Cold War days, "Trust but verify." I think that might be a helpful tack to take with AI where it is right now. AI could get to the point in, you know, my lifetime and your lifetime where it's like, holy cow, okay, just let that thing loose on things like this, and it's good. Like, you don't even have to... You can just put your hands up and, like, that is doing the thing better than any human could and whatever. Right now, we're still in the trust but verify mode for a lot of these things we're talking about. Okay, um- I've also heard chiefs of staff in this, this situation, this similar situation, define approved use cases. So this can happen in a couple different ways. One, I've seen, um, like the, the managers define, "Here's the approved use cases for my team," and they've quality checked all the workflows and done all the QA for halu-hallucinations and things like this. And they're like, "These are the ways that AI can legitimately help our team, and so I'm going to roll this out, train our team, walk them through the steps, and say, 'Here's your five to seven approved workflows.' And the team members can use that." And then the team members can and should be bringing back feedback to that manager, and the manager and team iterate off of there. So maybe it's a tweak to one of the existing five to seven workflows, and maybe it's an addition of, of an eighth workflow or something like that.
Token Spend Guardrails
The other way I've seen this go is to a certain token threshold, like dollar amount spent on tokens, people can do whatever they want. After this threshold, they have to get approval from this person. And after this threshold, like a higher spend, they have to get approval from this, this person. So there's a little bit of, a safety step in there or a QA step in there. A lot of companies I'm seeing are like, "We want AI adoption and usage and spend, and so we're using the amount spent on tokens as a proxy for how AI forward we are." And to me, that's-- I mean, it's a proxy, right? If you have nothing else, it's a pretty quick proxy. But it's-- there's a lot of room for like, "Whoopsie daisies, we just spent a whole boatload of money, and we came no closer to our stated objectives. In fact, we're just burning cash and burning tokens and not getting close to what we actually want or close to the actual work product and high standards and excellence we want or we're known for." So I would be real careful with that distinction. there's also like nuances with, okay, uh, certain, certain kind of AI tools use less spend but more time and all these different distinctions. So I get real-- I get a little, a little cautious when I just hear board members and CEOs lightning bolting down like, "Just use AI, and I wanna see AI spend go up quarter over quarter." It's like, hmm, okay. I mean, I can spend your money if you want, but that's not actually getting us closer to what we want. So I would just be aware of all of those nuances
Quality Checks and Controls
going in. I would have some sort of- Hmm. Some sort of quality checks or quality controls. So that could be ideally on the front end where you're not releasing anything or using anything that, uh, is, is not the quality or standard or actually doing what you want it to do. I would have that checked beforehand if you can. At the very least, have a quality check after the fact. So if someone uses AI and releases it out to the teams or uses it with the customers or what have you, and it's out there, at least quality check it once it's out there. Have some level of check, even if it's just, a manager doing a random, sample set of, "Let me check these AI pieces that are being deployed across my team and see if they're actually working as intended or if there's any risk involved," or anything like that. So to have some level of that built in somewhere.
Center the Pendulum
Yeah. The other thing that I'm kind of sensing from this email as I scan through it is it sounds like the team has gone through a lot of change. Like if you're getting new leadership team members and several long-tenured executives ha-are, have been leaving, and maybe it sounds like the culture is changing and it sounds like, "Hey, we were kind of AI neutral, and now we're swinging full board into AI," and the engineering team is taking two weeks to do a fully immersive training off-site or on-site or whatever you guys did. Um, which on one hand is fantastic, 'cause if you're trying to be an AI-forward company and then you're not equipping your people with that and asking them to do that with, with no training and no knowledge, that's, that's tough. But it sounds like you're actually putting your money where your mouth is or your team's time where your mouth is and saying, "Okay, we're gonna take two full weeks to do, um, immersion training," which is awesome. But, but I think with that comes... I, I kinda feel like the company and team was on one side of a spectrum, and then it got swung all the way over to the other side. So the pendulum was on this end, and then it swung all the way over to this end. And so when you say things like, you know, I, I, um, you know, overcorrecting old habits, uh, AI everywhere versus, more mature, balanced operating model, to me, that is letting the team center itself. We've swung from extremes, and now we're right-sizing and centering ourselves. If you can provide any clarity about how to center themselves, I think that could be helpful. Meaning It seems like the teams are picking up the signal like AI, AI now, AI go, and so they're doing that. And then you're seeing the, the unhelpful pieces of that. So it might be a little bit of clarification, a little bit of nuance. Yes, AI, and we want these checks. Yes, AI, yes, creativity, yes, innovation, yes, get rid of redundancies, and we want these kinds of outcomes. We want these kinds of results. Or maybe if it's, you know, even easier, like we don't, we don't want these kind of results. Don't do this, don't do that, and do something more like this over here. Okay, so maybe a little bit of nuance, a little bit of clarification of expectations, 'cause it sounds like people are responding to the expectation that was given to them, and that's great, they're responsive, but now just a little bit of, a little bit of centering or right-sizing there I think is, is, maybe warranted. Okay. And then I think, um, it sounds like your, your CEO is pretty good on this, but, again, that adoption is not the same thing as effectiveness. Is AI being used effectively in your organization is a different question than are we using AI? Is adoption r- are adoption rates high? And then I think you said something else about the CEO up here. CEO is fully supportive of an AI-first organization, although tends to focus more on the vision and business outcomes than the day-to-day realities of implementation. Okay. So my initial thought on that would be, um, connecting the dots between the implementation steps and the outcomes that the CEO wants, and turning that vision th-the CEO has into a reality. So sometimes they're like, "Just make it happen," like, "Just make it happen," and like, "Go more, AI, AI more," like m- "Tell them more and make it faster." And sometimes it's like the fastest way to get there is to set the foundation, set a strong foundation, get them fully immersed, get them trained, get them excited about it, get them organically exploring these things, and then we'll, we'll guide that process and channel that pr-process, put some guideposts so we lead them to the business outcomes we want. Um, I think sometimes CEOs get very enamored with like, "Oh, I wanna be able to say this, this, and this thing to my board and to, the public and all these different things," which is admirable and understandable certainly. But connecting the dots of like, we are doing those things, and here are the three steps that have to happen first for that to go well, for you to be able to tell a good story, for you to have the visible, tangible, reportable outcomes that you want to have. So I would just make sure that sometimes those dots aren't connected in some people's brains. It might seem common sense to a lot of us, but, um, sometimes you gotta connect those dots. So anyway, those are the pieces of guidance I would give you. I'm just gonna scan your question one more time for anything else. Um Yeah, and you might e- like, depending on your relationships with the executive team, you said it right here. Like, I absolutely agree we need to embrace AI. It feels like we've swung from one extreme to the other. You might say that if that's an appropriate thing to say in your, in your leadership team meeting. You might just say that like, "Hey, like I'm, I'm sensing that this is happening and we're swinging from this end to this end. I wanna make sure people are clear and feel supported and we're actually providing, um, the, the guide rails, the guard rails we want, towards the outcomes we want." So yes, that's what I would say. Robert, hopefully that's helpful. I think this is a very common question that's coming up for a lot of folks right now, and I think the fact that you're thinking through this the way you are and you're being thoughtful about the AI strategy and approach is, is really good and a huge advantage for your team. So thank you. thank you in general for being that thoughtful, uh, about your teams. I appreciate it and I'm sure your teams do too. But if anyone else has an AI related question or anything else, then feel free to send that on in and we will share that on a future episode. Other than that, we will call this a wrap and I will catch you next week on Leveraging Leadership. This episode is brought to you by Next Level Coaching. If you or anyone you know would like to learn more about executive leadership coaching, please visit www.next level Coach.