We Killed the Curriculum: How the Agentic Readiness Hub became an operating system

I thought I'd choose a bit of a very strong statement. And of course, you know, we kill the curriculum. I think what I'm trying to say is that we have changed how we actually go about learning. And it's not just selling AI or working with AI, it's also learning with AI. And that is incredibly powerful. So we've built something that makes sellers faster, more targeted, and more confident. So that they can sell more and be more relevant for what consumers or customers are actually looking for. And why do we need to do that? In my view, the old model is broken. You know, like when you build a comprehensive programs and they can be fantastic, you know, really good content, you know, the completion rates are good, you know, you're issuing the certifications. But what happens, you know, like doesn't mean that people therefore can apply it, can do it and will do it. And innovating, the time between learning and applying it have shortened dramatically. And that is, of course, you know, the forgetting, I don't know, it's a forgetting curve, you know, how quickly you forget what you've learned anyway. I think you need it, you know, as and when you need it. So in the past, contextual triggers weren't necessarily there. The content may be brilliant, but you really need what you need at the moment of, of need. So what have we built instead? It's actually an operating system in inverted commas. It's not in the technical sense, but more in a human operating system sense. We're calling it the Identic Readiness Hub. When I say we, it is actually in one of the regions of Salesforce. You know, we have quite a few operating units across the world. So I'm representing our central EMEA OU, and it's really good because everyone is doing something slightly different and we're trying to get the best of the best. Certainly our work here is a contender for that. What have we done? You could say we've actually created 20 modules, but that would be more suggesting it's still part of curriculum thinking. So I like really calling it 20 activators across the full sales cycle, across all the do my job and, you know, the sales skills you need. We have not created a separate platform. We have actually, we're running it where people work, in Slack for us, of course. And within 60 minutes of working with, um, Slackbot in the, with the Agentic Readiness Hub skill, um, you get an action plan. You know, you give it the context, you get a plan out, you don't need to search around. And people like Shaz sometimes look, why, why can't I just ask Slackbot directly? Why would I need to go through this detour? It's not a detour. If you know what you're looking for, then always ask the question direct. But you know that if you don't know what you're looking for because you don't know the whole topic, um, completely yet, um, it's better actually to access that curated content, that curated material of sources that are inspiration and these very specific points. So, um, it's just in time, right when you're working on the deal and not just in case. So you really wanna make it very specific. And I've put underneath the, the, the key points that I would, um, um, define for Agentic Learning. Just in time when you need it. Personalized. I mean, that's the holy grail for what we've been trying to do for a long time, but now it is actually what do I need at this moment in time given my deal, and it adapts to your circumstances. And it's in the flow of work where you're already working, and it's action-oriented. It's not going to give you a long spiel about, in theory, this is what you should do, and this and that and that. But it's really literally, hey, in your specific case, this is what we are suggesting based on the knowledge we have available. And that's pretty powerful. So the way it works is actually we've built this a Slack skill, like a custom-built instructions. It's a complex prompt. And it actually asks you, you know, where am I in a deal? Who is involved? And what do I need to do right now? And then it delivers to you and summarizes what you need. So that as an account executive, you walk into the meeting, prepared, you know, not trained. That would probably take a lot more time and a lot more context and a little, you know, but time you don't have, but prepared for that specific situation. And that is ultimately, um, what learning with AI is about. And in order to improve it even further, um, we've also started to look at it as an edtech product. I mean, just for inspiration. I mean, of course it's an internal tool, you know, we don't need to sell it. But of course Slackbot can be bought, but you know, like this is our, our knowledge, our, our work. But looked at, you know, how can we improve on it? So first of all, you know, like, yes, starting with the account executive in the situation, gap, tick, that we've already done. Building a habit loop, that was a lot more and is still an open question is like, how do you nudge? How do you make sure that people see it at exactly the right moment in time? And so we've actually made certain practical experiments on this already, but that is really one of the holy grail. How do you make sure people use it at the right time and on an ongoing basis? Then proving the ROI. I think you see, I'll actually get to some of the challenges because I wanted it to be both and like what's working and also where are the challenges. In terms of hub users versus non-users, that's still a bit of a challenge. That's to do with how information data is accessible, not what's technically possible, but also what's in an agreement with worker councils, etc., in terms of which data we would access. Last but not least, expanding the TAM. Of course, we are going beyond the account executives. You can do it also with solution engineers. Managers and new hires. So really that personalized learning path, so a learning path and it being personalized is all possible now with AI and using these sorts of approaches. But I wanted to leave you here on that slide with the thought that building a habit is really the key in terms of learning with AI. And I also know how difficult that is because there's things changing all the time and so fast. But you probably wonder what, what was actually driving adoption? What are some numbers around it? So 220 users, what does that mean? It's, it's the majority, it's nearly everyone in our Sales 3.0 universe, in our OU. You can see that they're using it quite often and they're sharing it with other people. The key people actually that are driving that adoption though are those people who actually tested it. So we are actually always saying is I'm, I'm not going to, to launch it before someone, uh, before the people who are, are going to use it have tested it, have given input, and we have been able to improve on it. But, and whoever has, uh, been able to see the panel yesterday, don't wait until it's perfect because what is possible and what is needed is changing so fast that if you were to wait or let it mature until it's perfect, the expectations will have changed, what is possible will have changed, and you know, you'd never, never get there. So I've actually very specifically invited our account executives to co-create what we're doing, the learning that they need, so that they can also not just test it— that's just a small number— but also give real-time feedback in a channel and iterate together with us. So that is actually a very powerful thing. It's not, you know, the learning that comes from someone you don't know, but it's also from your colleagues and you know who's contributed to that. And I personally find that really exciting because that's a lot more like community and we're helping each other. Last but not least, of course, what's driving adoption is also a cultural mindset that we are trying our own products first. And building those agentic capabilities they're selling. So I feel very much for anybody who is saying, oh my God, this is not such an easy thing. I know. And we're all learning in that journey together. One of the agents I had actually built— when I say I had built it, you never do any of that alone. I mean, I hope I've made that clear. Oh, there is a mosquito. Something. Going to attack me. Anyway, I will not bother you with this animal that's buzzing around me. So you're never doing it alone. Oh, that was what actually attracted the mosquito. You're doing it in a team. You need to contribute to each other. You know, you need to make sure who is doing what. And I think I mentioned yesterday the case about, you know, needing to build a, building a prototype within an hour. Yes, on my own. But then needing a whole week to establish the governance, who's going to provide what and who's making sure that the, the content canvas is correct and it's updated. And who is actually going to approve it in the end? Is it, is it ready for primetime? Is the, is that working? But here I want to leave you with a thought. Test what you're doing, live with it not being perfect because it's ultimately those testers who show some pride in, hey, I know how this is. This is practical, that will actually help you with the adoption of what you're doing. For the business impact, there's also fun discussion yesterday because I, I'd also brought in that I think speed in itself is a, is a business impact, specifically in times where everything needs to be so much faster and we need to learn faster. So here are 3 examples. I don't want to go through them in detail, but they're all Salesforce customers and publicly available information. If you were to come to a world tour, you could hear people speaking about those successes live. But I, I wanted to make the case here that we should always measure our success in the way of how it is driving success for our customers with their customers. And here you see some of the examples of that. Now the part no one likes talking about. But I deliberately wanted to make sure to say what is working and where do we still have room to grow, because that's what learning is about. You're never satisfied with what you've already achieved. You always want to learn something new. You're curious, you know, and you're trying to improve. So one thing is actually what I've observed, you see, I call it the hidden leak. And that might be a culture thing more in my culture. Not sure to which extent it would apply to everyone in the room here. But often AI starts working for one person at a time. And if they're actually not sharing it because it's very personal to them, you know, it's their success, it's how they are working, then actually it doesn't scale across the whole team as it should. And I know a lot of you are leaders of a team. I think you'd want to make sure that you are unearthing these sorts of successes, not to be a big brother is watching you, but in terms of making sure it scales so that actually everyone can save some time and readjust, you know, where you spend time, get rid of some of the drudging work, and you can actually do a lot more things that you haven't been able to do in the past. So, um, that is one thing. The next one is actually, uh, the dual operation, as I, as I called it, is, um, often you have two systems. What I've observed, two systems working in parallel. Let me say it, the old system and the new system. The new system being, um, the AI part. It's not so new anymore, but you know, it still feels new to a, to a lot of people. And I think somehow you need to make get the balance right in how you're moving from one to the other. I know you can't just switch off one immediately and move to the other, but you do need to have a view in your mind and discuss it with your teams on how you move through that journey. And I know how terrifying that can be because things are changing so fast, you know, are we moving fast enough? You need to get a, a picture in your mind and in your team's minds of of, of how you want to move. And then the last piece is more of a, um, an operational point, probably, you know, like what I've shown to you is really a routing mechanism. You ask the system a question, then the AI helps getting the right curated, um, content that you can then apply to your— and applies it to your specific situation and is offering you some opportunities on how to solve for the challenge or for the opportunity you're facing. But it's still a routing. It's not something where you are sensing, you know, it's not automatically doing it. It's not really proactively suggesting something to you. You're still human in the lead. You need to ask the question. So really, routing assumes someone asks. It assumes people are curious, they're really staying on top of things, and they're following that line of thinking. Wouldn't it be great though if it was starting to sense and could serve that proactively? That is open. So it can sense, and at the moment, the, the, the measurement gap is, is a real one. I mean, I have suggested that speed in itself might be worth a measurement, but some of the ROI figures that we would be looking for are still, um, a bit elusive. And I think we're in a time period where we are actually defining those measurements. Pure token usage is interesting because you probably unearth the people that are using it more, but in itself and over a long time period, that may not be sufficient for collecting the business impact that it has on customers. We have a lot of leading indicators, and the ROI, I think, will be something we'll all measure over time. I'd also want to state, um, make a statement on, on governance. Um, I'm very active in the board, um, scene in Germany, and, um, SME boards are actually often asking questions that are also quite challenging. You know, what about vendor dependency? What if someone's you know, it moves the switch. What do we do then? You know, what is our backup plan? What is our contingency planning? How are the overall supply chains working through? How do we actually make sure we, you know, we get that transition working right? So ultimately, no agent can answer that question. That is definitely still a human conversation. And requires human thinking, human ingenuity to, to solve for that. And your sellers need to be ready to have those discussions in real life. A very quick story on something that, that I did with the Agility Readiness Hub, because the night before our regional kickoff, the content shifted. Well, things normally don't shift quite that short notice or, you know, at these sort of critical junctions. But sometimes they do, probably just before a customer meeting. You may need to include new information, and it's important that your architecture moves with those changes. And that's the point, you know, like speed, how you can flexibly, um, account for, for, for these is important. And I definitely did that with the Sagentik Readiness Hub. Um, overnight, really the night before, before the meeting. That was fun. So the playbook I want to leave you with is actually that the context beats content. Knowing the questions and in which content the question comes is much more better than providing a lot of content that someone needs to, I was going to say drudge through, but that's probably not the right word, you know, work through, taking a lot of time for it. The habit, the habit loop is key, you know, like making learning with AI in whichever way you do it a habit that actually sticks and, and building it in the context, building it inside the workflow with the tools you're using already anyway. And what we are currently doing with the routing is I think the first step of the first chapter. And we are working as a company also on how you are actually getting to the sensing part next. And I think that's really the total game changer. And it also shows that learning and helping your teams to learn will move much closer to the doing because it'll be much more targeted, much more personalized, much more in the specific situation. So I quite like the, the thought about your internal series. A pitch might be a bit of a big word, but really also thinking about if this is an edtech product for you, or if— what about your edtech things you're building or using? What will be the pitch for that? I leave you with that question and want to summarize my presentation before we go into the question rounds. With actually a thought for the future. Because if you can prepare your account executives for the conversations that agents cannot have, I think that's a, a winning recipe. That actually means they need to spend less time on some of the drudgery work, and you need to help them, you know, get to these answers faster than the context so that they can be prepared for the conversations agents can't have. Now, I hope you have some interesting questions for me.

Yes, let's dive right in. You wrote a paper that was published in German, the KI Governance Leitfaden 2026, and you talk about accountability. And I want to translate your quote, which is, you don't govern AI, you govern the decisions AI informs. So I thought it was a very profound insight, and I'd like to know, how should revenue leaders apply that principle to ensure that their AI agents maintain conversation integrity and don't erode trust with bias?

First of all, maybe I want to say what they should not do. Because if you consider this a technical question and you leave it to your IT department, then I think you're not fulfilling your leadership role. It's first and foremost a leadership question. You know, who is taking which decisions and understanding that even if an agent takes a decision you have delegated to them, you're still accountable and responsible for, and you need to make sure that these decisions are an expression of your values and you're still understanding those. So in that sense, you're not governing technology, AI, but you really govern what are the decisions. You decide who is taking which decisions. And if someone else is taking a decision on your behalf, that it is within the guardrails that you have defined. You would do that with people. You know, you wouldn't just let them go off and just do whatever they want. You, you would actually set a specific brief, a job profile, And, um, and, and I think you need to do the same with, uh, with AI. So it's a, um, a leadership question, I think, not a technology question. It has technology consequences, but it's still a leadership question.

So AI needs adult supervision.

Excuse me.

Sorry.

I think the easiest way to, to, to, to look at it is actually, um, I mean, we know AI is not a person. We do all know that it's a probability engine. But if you still, um, give it a briefing like you would to a, to a, to a person, and if you also still make, you know, that is less, um, that is in your team, if you treat it like a, um, a team member, I think a lot of those questions, a lot of those answers flow quite naturally if you don't see it as something like dangerous that is out there, or something like magical godlike thing. It's not. I mean, it's, it's doing what we as people tell it to do. Yeah, that may be a bit of a simplistic way, and I know that, that we, we could have a big discussion around that from a technical perspective, and, and there are lots more possibilities, but Let me simplify. Bear with me and let me simplify it.

Tripp is asking, it looks like a fantastic program you designed. How did you get internal users to adopt this?

That's one of the things where I've highlighted the role of the testers. So first of all, I got a senior person to sponsor the module. I've also, you know, did the customer discovery in terms of what is important for you, what do you want to see in there? And I've also asked this person to give me names of people who would test it. I usually found some other people as well, and I made sure that they are from different areas. So it's not just from a narrow area of, of a field. And then the easiest way was actually to just let people do it with their case. I said, look, this is how it works, you know, go ask your questions. And does it give you something that's useful to you or not? And if not, why not? What, what would need to be different? I mean, I've learned some really wild things. I mean, for example, we had a competitive, um, um, module and I literally did, did put in, um, work with our competitive team. We put in all the information in and when they tested it and for a specific customer situation, the biggest competitor of course is doing nothing. Staying with where you are. And I totally forgot that. I mean, because, you know, it wasn't in my frame of thinking. So that was a really good one. They, they tested advice that was there and I said, oh my God, no, this, you'll not get anything on this one. Please test with something else. And then of course I, I, I put it back in and, and it was fast, you know, like the next iteration I sent a couple of hours later, had that information in. Um, So I think that to answer it very briefly is, is how do you get people to adopt it? You work with those people who like doing this sort of work, who are curious, who are pulling. They in turn will share it with others. Um, and, and it really grows, um, grows from there. It's not something I would want to mandate to anyone because, yeah, I think learning that is should be pulled and should be, um, some, some things you need to actually, uh, push because you need people, you need to speak a joint language and there's certain things that are a minimum acceptable level of what, what everyone needs. But ideally it's something you've hired curious people, you've hired people who wanna know about this and then they, they pull the information.

So with, um, Agentic AI, with, uh, digital twins, with, um, all this technology assisting salespeople, uh, what is the number one human skill that, um, you believe will be the ultimate differentiator in, uh, creating new customers?

Curiosity. Because if you're curious and want to know about the business of your customer You will listen more, you know, everything else flows from there. You will listen more, you will try and find out how they can grow, um, you will find out, you know, what I can do with AI, what do I— what I cannot do with AI, um, and, and the next question is probably, can, can you learn it or can you not learn it? Now I'm asking myself the difficult questions.

Um, well, I like where you're going. I think that, um, I'm reminded of Werner Heisenberg, uh, who said that nature does not reveal its secrets, it only responds to our method of questioning. And maybe the, uh, your answer, curiosity, really lies in the ability to ask not only better questions or more relevant questions, but more existential questions. Why are we here? And what are we trying to achieve? And how do we optimize things in, in, in the technology environment where everything becomes more messy, and, uh, we, we want to restore order and equilibrium.

But you know, when, when you, when you go in that direction, I'd probably say the, the, the, the most helpful, um, trait, um, actually I found in my own journey is actually— and it's actually one of our senior managers who first said it at our, uh, uh, global kickoff but I really liked it, and is, um, to start feeling comfortable with being uncomfortable. Um, and I still remember that because so often you need to do something earlier than you'd ever done it. You know, normally you get—

and I think I find it because Mark Vanyov always says, uh, you need to approach things with a beginner's mind. And, uh, and, uh, it is You know, when you, when you look at trust in, in sales, I remember Harvard Business School professor talking about trust as a, an exercise where you put yourself in a position of vulnerability. And just by admitting that we don't know it all, we're making a huge, huge amount of progress over people who pretend they know things that are not true.

I think you just hit the nail on the head because I think certainly in our culture, I think people don't want to say, oh, I don't know, because it seems like, you know, not a good thing to say. But actually, I think this is the most powerful answer in any AI discussion because often we as humankind, we don't know yet. But I think the most powerful word behind it is the yet. We don't know it yet, but we'll find out and we create it most likely together. So I think rather than waiting for it to be perfect or rather than staying out of the discussion because you don't know yet, I think it's incredibly powerful to just say, I don't know yet. And that's actually, I think, also an invitation for learning. So if I don't know it yet, well, I'm going to learn it or I'll figure it out. This, I think, is also what you described as the beginner's mindset that I really like from Mark as well.

Yeah, I think it was a good idea for Mark to go to the Far East and study a different set of philosophies other than the Western philosophy. And I think it is in the divide between two cultures that new ideas are born. And, and I, I believe in the power of ideas. And we are only as strong and as powerful as the ideas that we adopt and pursue and turn into reality.

Yep, I like that.

Well, it was such a pleasure, Dagmar. We have to do this again soon. We have run out of time. We are in overtime now. But I can't thank you enough to— thank you so much for joining us today and sharing your insights and your wisdom and showing your ability to think on your feet in the moment and even battle a mosquito doing a presentation that had a t-shirt that had the Microsoft logo on it.

Did you know it was— oh, did you see it? Yeah, there was somebody, the person who owns this, he came in and he needed to take something out of the fridge. I don't know what mosquito, it was quite a big bug anyway. Thank you for inviting me, I so much enjoyed it.

I truly appreciate your time. Was there any Bremse?

Ah yeah, that, that can't be. It was bigger than a normal mosquito. It was, it was, was more like a— I'm glad it's gone.

Thank you.

Thank you. Ciao.