When Scarcity-Built Systems Meet Abundance - The Judgment Gap in AI-Powered Sales Leadership

I'm very interested in hearing your thoughts. I think your book is very illuminating. And my first question is really, tell us, what did you deploy and what problem were you trying to solve?

Yeah, first of all, thank you, Gerhard, for having me. There's been some really great AI in sales conversations the past 2 days, and I'm really excited to share my thoughts in today's session. Back to your question, we actually developed and deployed 2 AI tools, and they really taught me something that I wasn't expecting. The first one was a show me you know me outreach tool. It pulled together contacts on the company, the person's role, their LinkedIn profile, buying signals. And instead of sending a generic template, reps could really create outreach that felt researched and relevant to the potential customer. And it actually worked. We could produce thoughtful outreach at a scale that would have been impossible a year earlier. The second tool was OPA, an opportunity progression AI assistant, and it helped reps prepare for meetings and inspect deals using MedPic, challenge selling, mutual action plans, and our pipeline stages. Most importantly, it really separated what the buyer had confirmed from what the rep was assuming, which is a big problem. If there wasn't an economic buyer, if there was an unknown paper process, if there was no buyer-owned next steps, it really surfaced all of that so that the manager could see what was going on. The first tool really created activity. The second became necessary once we had abundance because the activity was no longer the constraint. Judgment really became the constraint.

So what stopped working once AI made the activity abundant?

You know, the conversation model was now broken due to the abundance of activity. So the coaching instinct to simply say, do more, had to change. We gave our team a tool that made research-backed personalization almost free. Every message could reference the right person, role, company context, and on paper it sounded like a dream, right? But here's the catch. When personalization costs almost nothing, It stops being a differentiator after a while. Everyone starts using the same capabilities and the buyers start getting a full inbox of messages that all look personally directed at them. So the manager's questions really have to change. Not can you produce more of this, but who do you choose? Why them? Why now? And what did the buyer actually do? What I learned is that AI is very good at scaling good judgment, but it's even better at scaling bad judgment, and it does it faster.

So is this only a prospecting problem?

Oh, not at all. I just feel that prospecting is where you notice it first. Take account planning, for example. We were never short on documents. No one ever needed one more file to sit in a folder. The value of account planning was that it forced someone to really think about the questions like, who matters? What is happening in this business right now? Where is the risk they have? And what don't we know yet? Now, AI can now create business plan accounts in seconds. So the question becomes, does your understanding of the account get better, or did the document just have a bunch of great information? And you can actually ask any of this with CRM updates, call summaries, follow-up emails, forecast notes, all of it, right? So the message is that the information on the document was never the point. It was the thinking behind it that was important.

So obviously, you need to shift focus. And so what do you inspect instead?

So AI made show me you know me content very easy. Everyone has personalization tools now. So the question isn't whether a rep can generate a polished message, it's whether they turned what they learned into something meaningful. I wouldn't stop at the rep's reasoning. I'd inspect what the reasoning converted to. Did they get a referral? Did they activate a mutual contact? Did they ask for introductions? Did they actually get a real meeting? What did they learn that changed the view of the account? What surprised them? Did the buyer actually respond? And if not, what did the— what did we miss? Right? The thought is now that a rep can sound convincing in seconds. But what I learned is that they can't fake a warm introduction or a meeting that actually happened.

So when somebody rolls out AI, And activity spikes. What should the manager inspect?

They need to find the point where a seller activity becomes buyer action. Don't start with what the rep sent. Start with what the buyer did. Did they respond, accept the meeting? Did they agree to next steps? Did it turn into a qualified opportunity? Seller activity is close to free now. Buyer attention is really what got expensive. Because everyone else got— everyone else's activity got cheaper at the same time. And that was really the premise of OPA. Forecasting confidence should come from a buyer behavior, not seller activity, or worse, seller assumptions. So if there's no economic buyer, no known paper process, no known buyer next steps, the deal shouldn't be qualified as high confidence. And mainly because if nothing on the deal is buyer confirmed, you don't have a deal. You have a well-documented hope, and it's important for the frontline leader to be able to identify this.

So in all of this, what surprised you and how did that change your belief system?

Yeah, so I had to let go of the belief that more activity is automatically better. Uh, what surprised me was what happened with OPA. We built it as a training wheel to help people develop deal movement more efficiently and aligned with the customer, not to replace the judgment. Um, I'll give you an example. We had a new rep that did not have med pick knowledge, but they ramped up much faster than I expected when we introduced the tool, and that was really exciting. We were facilitating people to develop themselves faster. But then I noticed something after some time. Some reps weren't learning the way of thinking underneath the tool. They were getting good because of the tool. And that really led to an aha moment for me. So the question shifted: were we creating better sellers faster, or were we just creating people who were simply using the tool And learning how to use AI. So now activity has become abundant. Judgment became the constraint.

Wow.

And judgment is a commodity that they don't sell at Walmart.

I'm sorry, what was that?

Judgment is not a commodity they sell at Walmart.

No, not at all. Judgment is very highly needed in today's environment.

So walk me back when you first realized that something went wrong.

You know, the top of the dashboard looked better, but downstream the results were not improving at the same rate. We had more touches, more accounts being worked, more relevant outreach. If you looked at the top-of-the-line dashboard numbers, you actually felt good and you would probably relate that to the board and CEO. But when you started looking down the funnel, the questions became, were we getting any better responses? Were they real conversations? Did they actually convert to opportunities or even revenue? And that really changed the question for me and my frontline leaders. How much more activity can AI create? And was— what is all of this activity actually producing? And was it worth it? So the dashboards weren't lying when we were looking at the top of the dashboard. We were just asking the wrong questions.

So what you're saying is that we need to build the operating model around the tool and not expecting technology to solve that problem?

It's a great question. I would actually put it a little bit different. Building the operating model before choosing the tool Buying the technology first is backwards. The tool can't decide who your ideal customer is. It can't decide good messaging for your business, and it can't decide what a coach— a manager should be coaching on. Those are all business decisions. So before we scale anything, I would want to ask these questions: What problems are we solving? What decisions should get better? What behavior should change and what metrics should move if this works? Otherwise, we automate what the team was already doing, including the parts that weren't working before. And AI can hide a broken process. I'm sorry, AI can hide a broken process. It can't fix one.

Absolutely. So another question about coaching. How did managers react when the coaching model had to change?

Well, you know, I had to come in and think about what I was going to say. The wrong way to have a conversation is to tell experienced managers that they've been doing it wrong. They haven't. It's just the world has changed around them. I put it this way. I coached you for years to manage something scarce, which was activity. Now activity is abundant, so what we optimize for has to shift. Um, and I'm not taking away what made you good. I'm asking you to move your attention further downstream. Um, so spend less time asking, is there enough activity? Spend more time asking, is the right activity? And is this the right buyer moving? So That takes more coaching, not less. And that's why frontline leaders are probably more important than ever.

So final question, if you were rebuilding this today with all that insight, what would you do differently? And what would your advice be to the sales leaders in the audience?

Yeah, you know, we're always moving fast in sales. Things are always changing. We want to act What I would do now is I would slow down before I sped up. Before deploying anything broadly, I baseline what we do today, who we target, current response rates, meetings, qualified pipeline, and where people fall out of the funnel. Then I define 3 things before scaling. What does AI own? What does the seller still own? What does the manager inspect? And I decide upfront where the time saved by AI is supposed to go. The technology creates capacity, but leadership decides where that capacity needs to go. So I would build the inspection model before scaling activity and then train the frontline managers to use it.

Michael, I see you taking notes.

Yeah.

And I know you have a burning question.

How do we make—

yeah, I'm a furious note-taker, and I have been on the past 2 days here for these sessions. So, hey, so Joshua, how do we make judgment abundant?

Oh, great question.

So making judgment abundant is really about enhancing the skill set of the frontline leaders who are Taking the time to understand what is happening. Um, like any other tool, we have to make sure that it's used for growth and advancing the skill sets we already have. If we use it to replace what we're currently doing and the, the person using that tool is not scaling up and growing, then—

Mm-hmm.

The tool itself becomes something that hampers growth, and frontline leaders have to be able to adapt, understand when the tool is hampering growth or it's actually being used for growth, and coach to that on a continuous basis.

So frontline managers really need to look at outcomes, and outcomes being the experiential value of learning judgment or becoming skilled in the areas of judgment. And then they can look at reporting, they can look at dashboards, they can look at the entire metrics across the team and what they're doing to see what that outcome is or what the collective— what those outcomes are. So are you getting good visibility there and moving in the direction of focusing on measuring the outcomes as making the judgment abundant?

Yeah, sure. I mean, and, you know, the dashboards always are good to look at trends, to look at numbers, But at the end of the day, you have to inspect. You have to really sit down with your reps to understand how are they using this information to understand the client. What are they doing that is different from what the AI tool is doing so that that client is going to be willing to, one, talk to them and listen to what we're actually trying to solve for the client?

Yeah. Well, we need to stop. Um, I think that, uh, Josh, uh, gave us a very, very good lesson here. And my takeaway is that we need to ask better questions.

I agree. The, the questions have to be deeper, right? It's not just, uh, surface-level questions. They have to be second, third-level questions. to really understand and help coach. And AI is just progressing that more and more.

And that reminds me of the Werner Heisenberg quote, the physicist who said, nature does not reveal its secrets. It only responds to our method of questioning.

Yeah, 100%. I couldn't agree with that more.

Thank you, Josh.

Thank you, everyone. Have a great day.