WorkLab: Culture, not tech, softens AI’s impact

0:00Hey, folks, Jeff Berman here, co-host of Masters of Scale. I am thrilled to share some of the new names who will be joining us at this year's Masters of Scale Summit. They are the leaders driving the most pressing conversations in AI. Replit founder and CEO Amjad Massad, Cloudflare's Matthew Prince, Signal president Meredith Whitaker, and many, many more who will take the stage this October 20th through 22nd in San Francisco. We want you there with us, too. Join us at mastersofscale.com slash pioneers. That's mastersofscale.com slash pioneers.

1:00and the bottom line, how they navigate uncertainty, pressure, and high-stakes moments. I'm Bob Safian, former editor-in-chief of Fast Company, and I'll be your host as each episode breaks down what you need to know right now. You can find Rapid Response wherever you get your podcasts.

1:27One of the biggest challenges of AI today, it's so easy to start experimenting. It's a lot harder to integrate whatever you build into your everyday workflow. And that to me kind of underscores that it's not just about like getting the work done. It's actually how do you dissect an everyday workflow and inject AI in it in a way that's sustainable, that's repeatable, that's trustworthy.

1:58Welcome to WorkLab, the podcast from Microsoft. I'm your host, Molly Wood. On WorkLab, we talk to experts about AI and the future of work. Today, we're joined by Rana Al-Khalyubi, AI scientist, entrepreneur, venture capitalist, and the host of the Pioneers of AI podcast. Rana is also the co-founder and former CEO of Affectiva, one of the first companies to bring human emotion and intelligence into AI systems. Rana, welcome to WorkLab. Thank you for having me. I'm excited for our conversation. Okay, so you are a longtime practitioner here. You have spent years building and deploying AI into organizations.

2:33I want to ask you about the human factor of that deployment. What do leaders misunderstand about human readiness that can cause them to stall out? So I've spent a good part of my career building emotional intelligence into machines. And that led me to this realization that, you know, when we think about human intelligence, your IQ matters, your cognitive intelligence matters. But actually what matters more is your emotional and social intelligence. And that's the part that we're missing in this whole AI revolution. Even when we're thinking about how to deploy AI, we're not really thinking about the human aspects of it.

3:12And what I found in a lot of cases is when organizations are trying to implement AI, they're forgetting that you have to kind of bring people along the journey with you. And so there's just there often tends to be a lot of resistance to some of these new technologies. And it's a cultural challenge, not just a technological challenge. I want to go back and ask about building emotional intelligence into AI in the first place, right? Here you're building it into AI and into systems, and yet the humans still could use work there too. Right. How do you model what we don't always see in leadership? Yeah, well, that's the thing.

3:52It turns out over 90 % of how humans communicate is in your facial expressions, your body language, your hand gestures, your vocal intonation. This is why we're doing this in person, right? The problem is, though, when you think about AI, AI is completely oblivious to these nonverbal signals. It's just focused on the actual words you're using. So we try to build using machine learning and computer vision and sensors, AI that can read and recognize these social and emotional cues. And by doing so, it's not just improving human-machine connection and communication. It's actually the idea is to improve human to human connection, especially in a world where AI is just so ingrained in our everyday lives and our workplaces and whatnot.

4:38Right. So is there a world where leaders could be asking AI, how do I cultivate the human skills that I need to bring my employees along in this journey? You should have an AI thought partner that can help you be a better manager, that can help you deal with challenging situations at work. We're an investor in a company called Tough Day, and their conversational AI agent is called Tuffy. And it's designed to tackle workplace challenges. So it helps you kind of deal with challenging people at work or situations where maybe it's hard to get support. Fascinating. So even your investment approach takes this human-to-human kind of philosophy at its core.

5:18Yes. It sounds like. Yes. Our investment thesis is human-centric AI. So we believe AI ought to augment human potential and amplify our abilities, not replace us. You're getting to the heart of some of the cultural resistance, right, which is that people fear the replacement instead of the augmentation. How do we lay the table and deploy tools in a way that reassures people that that's not the goal? Yeah, I get why a lot of people are concerned about AI because obviously it is changing every job. I think the call to action here is that organizations should really lean into helping their teams, like, lean into AI and reskill and also experiment.

6:03Like, I tell people, like, take a very playfulness approach to testing AI and seeing where AI fits. We just trained a chief of staff AI agent for my fund, Blue Tulip Ventures, and we call her Blue. And Blue's, you know, she still struggles in some cases, but it's helping us become more productive. I'm iterating with Blue at night and I'm like, okay, Blue, like overnight, here are the set of tasks that you should be working on while we're all asleep. So I think there are ways to take a very, yeah, experimental approach and iterate. It sounds like it's also pretty clutch for them to have names. Tuffy, Blue.

6:45I know. Like, what other agents do we know? who else is in your agent social circle. Right, exactly. In many of these interviews, people I have spoken to have said to organizations, it's important to focus on the outcome that you want from your AI. And that is true. Right. But when can focusing on outcomes cause you to lose track of what the humans are still better at? I think it's important to focus on outcomes because we also see a lot of situations where companies are either building AIs that aren't really solving a real problem or they are experimenting with AI, but they're kind of just trying to fix or they're trying to use AI in an already broken workflow, right?

7:26So if your workflow doesn't work, using AI to optimize it is not going to work. And I think that's kind of sometimes missed. Like there's groundwork that you have to lay. You can focus on outcomes, but you have to take other steps to get there. You can't just sort of say like, we put some AI on it, we're there. Right, absolutely, yeah. And then I think human oversight is really key. So we are big advocates for that. I think it's important to not take the humans out of the loop, but be smart about where is that human oversight necessary? Where is it optimal? Speaking of agents with names, there are a lot of companies who are starting to add in effectively agentic employees.

8:08Yes. So when you think about that human connection and when you think about when you look at companies that might be trying to address this, what does it mean to have humans working alongside agents as co-workers? It's inevitable. And one of our investment theses is actually one of the ways AI is creating a value and is shifting how value is being created is through this idea of an AI co-worker. And I actually think we are so focused on what can this coworker do? We're not thinking about, okay, how is it going to coexist in a team of other human beings? And what does that look like? So, for example, how do you enforce culture in a company where it's hybrid human and AI team members, right?

8:51How do you enforce accountability? What does that look like? When we are creating these AI agents, there's all sorts of biases that come into play. In the same way, when you're asking these questions as an organization around who to hire, what skills are you looking for and other kind of like implicit biases that come into place when we're hiring. I think the same thing is happening when we're creating these AI coworkers. So we have to think about that, too. Yeah. And again, I don't think we're paying enough attention to all of that right now. Well, and it feels like leaders at businesses are going to be in charge in some cases of creating these agents or directing their creation and then managing them.

9:33Like, what do you think starts to look like a framework for that? Yeah, that's challenging, too, because the underlying AI models aren't there yet. So an example, a lot of these models have memory today, but the memory is not optimal. The AIs can sometimes forget things. It doesn't have a good sense of time. So it will say, okay, I'll get you this on Monday morning. And then it doesn't really know what Monday morning is, right? So these are things that, you know, they'll get better over time. We'll fix it. But for now, it's kind of hard to incorporate these agentic AIs into everyday workflows without having to iron out all these kinks.

10:15So it's non-trivial, but I think it's powerful when we get there. You are, like we said, a believer and a practitioner. Like you started a company around that built emotional intelligence into AI systems. Talk a little bit about this journey and how you saw that so early and how important you think that is to overcoming some of the potential barriers to resistance that we see right now. Yeah, so I started the company out of MIT in 2009, and we were so early. We were basically pre-smartphones and trying to bring these like computer vision-based algorithms to the world and like talking about emotions when nobody cared about kind of social and emotional intelligences.

10:56So we were very early to the market. We ended up commercializing the technology in a number of industries. So half of the Fortune 500 companies use us to test how people respond to content and products and services. We also ended up selling the company to a Swedish publicly traded company in the automotive space. So that's another use case where you want to understand like what is the kind of how are drivers engaged and driver attention and drowsiness, etc. But I actually think it's way more timely now because as you think about generative AI and how we are integrating AI in every aspect of our lives, for this AI to be truly intelligent, it really needs to understand, like, how are you feeling?

11:45Are you stressed? Are you in a rush? Or do you have plenty of time, right? It needs to understand the general context of this interaction. And so we're missing all of that. I think we'll get there. And I think it's important both for kind of the AI co-workers or thought partners, but especially so for physical AI. Even in an agentic co-worker, I was just thinking like the best thing about having a co-worker is chit-chat. Right. You know, is throwing ideas around. Like I could see it being very important for an agentic co-worker to also maybe understand like you're a human and you're tired today.

12:22Right. Or now is not the right time to ask you about this because you're clearly like in a rush or you're clearly like stressed out. Or even a little bit of empathy. Empathy is actually one of the main ways humans build trust with one another. Because I'll see that you're a little down today and I'll say, Molly, like, what's wrong? Like, what happened? Like, did anything happen at home? And that kind of empathy is really key. And so if these AI agents don't have any of that, that's going to affect the level of trust we have with these technologies. And then, of course, now as an investor, you're choosing companies that you want to succeed, which gets us to this question of adoption.

13:05A lot of companies are doing pilots. They're doing some experimentation, but it might be siloed and it doesn't grow into the organization. What are you saying to your portfolio companies about adoption and what are you hoping to see companies start to think about so that there's a road to success? Yes, I think that's one of the biggest challenges of AI today. It's so easy to start experimenting. It's a lot harder to integrate whatever you build into your everyday workflow. And that to me kind of underscores that it's not just about like getting the work done. It's actually how do you how do you like dissect an everyday workflow and inject AI in it in a way that's sustainable, that's repeatable, that's trustworthy, right?

13:49Like, I feel like a lot of these AI coding platforms and co-working platforms, that gets you maybe to 60, 70 percent of the way. But the remaining 30 percent is actually a lot of work to ensure that it's accurate. It's not going to hallucinate. It's not going to break an actual workflow. What do you advise the companies that you invest in in terms of thinking about that adoption? Like that seems to be, you know, it's, there's a lot of, I have seen this many times in the tech world. There are a lot of companies who say, I have built a great technology. And if it does not fit the workflow, if it doesn't integrate with the organization, if it doesn't fit culturally, it's not going to take off.

14:33Right, exactly. That's actually one of our criteria when we're looking for investments. One of our theses is kind of the application of AI in vertical, like antiquated industries where you can come in and reimagine entire workflows with AI. But we tell our companies, if you're like building the solution that sits on the side and is not at all already integrated, like there's so much friction. They might do a proof of concept. They might try it, but they're not going to end up using it on a day to day basis. So we look for companies that are deeply integrated. And then I also think, I used to say that at Affectiva, my company all the time, at the end of the day, you're selling to a human, right?

15:15And so humanize the selling process. Like think about this champion at the other end and kind of think about like how can you help them make the case for your solution? So, yeah, just kind of humanizing the whole thing I think is so important. I like this thing you said, too, about reimagining your entire workflow. We talk a lot about frontier firms. I imagine that the companies, certainly the early stage startups that you're talking to are right on the cutting edge. They are AI natives themselves, I would imagine. Just give us a sense of the difference in mindset. Like, if you're just a normal person in a business, it's almost like you can't understand the universe that a true AI native is operating in.

15:57I absolutely think it's a mindset. And I think the mindset is you got to have like an experimentation mindset, an innovation mindset, like an AI forward mindset. I actually joke that in my household, I have two kids. My daughter's 22 and my son is 17. And they sit on the opposite ends of the AI spectrum. So my son is super AI forward. He's always trying the latest AI tools. My daughter, on the other hand, like refuses to use any AI tool. And her whole thing is like, we need to double down on human connection, which I love too. So this is ironic because this is what you have spent this big chunk of your career working on, right, is combining these tools together.

16:40And it really does represent where a lot of companies are, where a lot of employees might be. You've got some who are all in. You've got some who are on the complete other side of the spectrum. Is there, do you think, a bridge? And based on your experience doing exactly this thing, what might that be? Yeah, it's a great point because I do think it's very representative of what's happening in the world today, right? Like, yes, there are people who are so leaned in and there are people who are very skeptical and they're kind of leaned out. And I think the right answer is in the middle. And it's under this umbrella of human-centric AI.

17:17Like, how can we lean into AI while keeping humans at the center? And so I love that my daughter is all about human connection. I think we should not lose that and we should not let AI get in the way of our human relationships. But at the same time, I love that my son is at the forefront of all of this because he will get to shape it. And yeah, he will get to have a strong say in how this turns out to be. And I think so. I think the answer is somewhere in the middle. You can't opt out. And I think when you opt in, you kind of have to opt in with keeping the humans at the center. We have to be honest and say that there are leaders in corporations who are hyper-focused on outcomes who are willing to cut humans out of the loop.

18:03You have staked your, you know, investment thesis on the idea that cutting humans out of the loop is actually bad business. We will need humans in the loop, and it's important to have human oversight. Now, that doesn't mean that some jobs won't go away. Like, I'm a pragmatist, too. I recognize that the whole job landscape is changing. But I think there are ways to do it with a lot of intentionality, with a lot of thoughtfulness. When you think about and evaluate new companies, some of these tools, some of the speed and efficiency benefits are likely to become commoditized. So what do you see? Maybe it's human-centric.

18:46But what do you see as differentiators for these future tools and companies? That is a great question because it's important to look for defensibility. So for us, it's the kind of underlying IP. That's a moat. Data is a moat, right? Do you have access to proprietary data that off-the-shelf models don't have access to? That data could be kind of the user's data, like personal data, like our health data, for example, is a great example of a potential moat. But I think because AI is moving so fast, defensibility isn't like at this moment in time. You have to look at the longevity of the defensibility.

19:27So I tell companies, if you are worried that the next version of these AI models is going to render your product obsolete, that's not a defensible business. And I certainly don't want to be investing in it. So we try to really think about where is this all headed? And do they have a defensibility that is, you know, that has longevity? to it. Let's talk about what humans should do to be better co-workers, to keep themselves in the loop. And I don't just mean learning AI tools. Like, what are the human skills that we all need to keep working on? Yeah. Well, I think there's a few skills that are going to continue to be really important and kind of, yeah, uniquely human.

20:08So one is communication. We're going to continue to need to be great communicators, whether in writing or in like in real life. So communication is an important one. Critical thinking. I think we're going to continue to have to think critically about things. Creativity. I think that's an important one. And collaboration. I do think that's an underrated skill. Whether you're collaborating with humans or AI, you need to be an excellent collaborator. I mean, all things that make you a good employee anyway. Right. Exactly. And empathy. I'd add empathy too. Yeah. When you think about adoption of tools and processes, what's going to be the difference between companies that really succeed in the AI era and the ones that just kind of stall?

20:50One is commitment from the leadership that sends a strong signal. Two is this experimentation mindset. And three is recognizing that AI is moving so fast, you always have to be experimenting with the latest technologies and tools out there. Which goes right back to commitment. You have to like stay in the game and be willing to be flexible. Totally. Yep. So, you know, we've been talking about things that maybe aren't there yet or the mindset and cultural changes that need to happen. But is it still your perspective that you got to do this? You got to do this. You got to. My call to action for anyone listening to this conversation is go download whatever AI tool that you potentially have access to and just give it a try and see if you can.

21:37There's one aspect of your work that you could try doing with AI instead. What would you say to companies? We've talked about where we are today and how we might get there, but what would you say to both the companies and the employees who are sitting in this moment and maybe thinking about setting it out? So for any organization, whether it's the leaders of the organization or the employees, it is imperative that they be using AI or they risk becoming obsolete. Just get working, get playing, get experimenting. So you're an early stage investor, which means you are seeing companies and ideas and workflows that no one even knows exists.

22:13What's some of the coolest stuff you're seeing? I'm actually most excited about the application of AI in health and wellness. I call it the trifecta of sensors, data, and AI. Sensors are becoming more and more mainstream, wearables, right? Think wearables. And then combine that with multimodal data and then both predictive and generative AI. And that basically means we are at the cusp of a health span revolution. So I'm excited about that. Just going to make us live longer, no big deal. Great. Yeah. I want to do a quick lightning round with you, if I can. So we'll just go through a couple questions really quickly.

22:50For leaders, what is one mistake that you would warn a leadership team against when they're scaling AI? I love that you can focus on low-hanging opportunities in using AI, but don't fear away from big strategic initiatives as well that can really reimagine what your business is doing. Think bigger. Yeah, think bigger. Outstanding. For individuals, what is one capability that employees need to build if they want to stay relevant and also not panic as AI becomes embedded in everyday work? I'm going to say playfulness. Yes, like be playful and experimental. I love that answer. Rana, thank you so much for the time today.

23:32Thank you for having me. It was so fun. If you've got a question or a comment, drop us an email at worklab at microsoft.com and check out Microsoft's Work Trend Indexes and the WorkLab Digital Publication. You'll find all of our episodes there along with thoughtful stories that explore how business leaders are thriving in today's AI era. You can find all of that at microsoft.com slash worklab. As for this podcast, please rate us, review us, and follow us wherever you listen. It helps us out a ton.

24:06On Masters of Scale, iconic leaders reveal how they've beaten the odds. Asking really strong questions is a superpower. You want to show up with something radically different and how they've grown companies to incredible heights. The greatest rewards always come from the greatest risks. That's hit the gas. Airbnb, Zillow, Microsoft, Liquid Death, and more. Hear from the founders who've changed the game. It's anything but business as usual. Find Masters of Scale on Apple Podcasts, Spotify, YouTube, or wherever else you get podcasts.