Your next manager could be powered by AI, with Katherine von Jan

RANA EL KALIOUBY: Picture the best manager you’ve had to date. The one who gave you the perfect advice, challenged you to be better — maybe even helped you jumpstart your career. It’s a rare thing to have a manager who can also be a mentor and champion of your professional and personal growth.

On the other end of the spectrum, we’ve all had a bad manager. Those that micro-manage, parachute in-and-out, or lack communication skills.

And it’s no slight to the individual. Managing is often a thankless, difficult job. Which makes it a role ripe for improvement. AI could help.

KATHERINE VON JAN: I have seen some recent study that 41% of Gen Z trust AI more than their manager, so I think one misconception is that people don’t really want to use it or don’t trust it. I think actually they do, and they’re using it, whether you like it or not.

EL KALIOUBY: Katherine von Jan is watching this trend closely. She’s co-founder and CEO of Tough Day, a company she started after decades working in people-management at big tech companies, including Salesforce. She’s building AI managers to help professionals thrive. Her advice? Try new AI tools, right now, at every level.

JAN: People are in sort of analysis paralysis over what we should do with AI. And I would say start using AI — thinking through unintended consequences in advance as best you can and preparing for those is great. And you do that alongside the team that is actually trying to make the solution work.

EL KALIOUBY: Today, Katherine and I are digging into questions around the future of work and AI. We talk about the concept of un-bossing, approaches to DEI work at this critical moment in time, and how AI could be a solution to how we work moving forward.

I’m Rana el Kaliouby and this is Pioneers of AI – a podcast taking you behind-the-scenes of the AI revolution.

[THEME MUSIC]

Hi Catherine. Welcome to Pioneers of AI.

JAN: Hi Ronna. I’m excited to be here. Thank you for having me.

Copy LinkHow workplace pain points inspired an AI support system

EL KALIOUBY: Katherine has been in management at tech companies for decades. And in all of those years, she gained some valuable insight on how tech can improve our work lives. A couple of years ago, she saw an opportunity, which eventually led her to co-founding her company Tough Day.

So, I wanted to know more about her background and the “a-ha” moment that led her to founding her company.

JAN: I started my career at Lotus back in the nineties working on collaborative technology at a time when the internet was brand new for business, trying to figure out how that technology would empower people to work better and smarter and create value and all of those things.

So the last 10 years before starting Tough Day, I was at Salesforce and I was leading innovation projects there internally and externally with customers and partners.

EL KALIOUBY: The leadership team.

JAN: ELT leadership team. And along the way, one of the projects that we did internally was how to improve belonging and inclusion inside using tech.

And of course that work starts by deeply understanding and researching what the challenges are for folks internally, not only inside Salesforce but in the world. And that research was fascinating. We learned a lot about what happens when people struggle and where they turn. So there are two common paths in that. One is that they go to some kind of reporting system internally. And actually that’s quite rare. Only about half the people will ever report something that actually needs to be reported. And there’s a lot more that happens before then. So that solution typically means we’re going into investigation.

And that is scary for everybody. And then the other place that people go are colleagues and often friends and family at home. That’s what we hear the most even today. So when you go to friends and family at home, you are not necessarily getting great advice. You’re not necessarily talking to someone who knows HR or knows management best practices or employment law or any of those things.

So the big insight was that we need something before you actually talk to someone in HR about a problem, a place where people can go. And we created that there in something called the Warm Line.

EL KALIOUBY: A “warm line” instead of a hot line – a place for people to call when something felt off but hadn’t reached any kind of critical stage.

JAN: So when we built this warm line and staffed it with amazing humans who were really smart on the topic of DEI and lots of other things — HR and management — they were able to handle a small population. The initial cohort got the answers that they needed. They got guidance, they got help, they got empathy. They were able to go back to work very quickly. Their performance improved. It was a magical story. And unfortunately, that solution doesn’t scale.

EL KALIOUBY: It doesn’t scale because the Warm Line needed to be staffed by humans with the right skill set. Katherine and her team simply could not find and train enough people with the right mix of skills and experience to serve every person at the company.

JAN: So we know that everyone in the organization has these moments where they need help. And the only way to scale the solution so that everyone gets the help they need is to use AI.

EL KALIOUBY: And here was the opportunity. Imagine an AI people-manager that could help employees express their personal challenges and concerns, and jump in on problems so they didn’t need to escalate. That’s what Katherine set out to build.

JAN: It was just at the time when ChatGPT was born and I knew that that technology could be used to do things differently and support workers.

And I did a couple of experimental projects on my own, got a couple of teams together. And what we learned in that is that actually we could get great content from HR management, employment law, and create this experience with AI and be able to serve everyone even more safely and consistently and expertly than most humans can.

Copy LinkHow Tuffy turns company knowledge into employee guidance

EL KALIOUBY: So let’s get into exactly what Tough Day does. Walk us through how an experience would look like for a person using Tough Day.

JAN: Yeah, so first, we have created both generative and agent AI to support workers, and it is a B2B sale. So we’re selling to organizations and essentially the core product has already been fine tuned on great management, HR, employment law, and other kinds of content. And then in addition to that, we’re ingesting the company’s information. So their employee handbooks, their strategy, their values, their organizational wisdom, learning and development. Anything that the organization would use to onboard an employee, we would say use that to onboard our AI.

And over the course of a week that AI gets really smart about that organization. The point is that once it’s really honed to the organization, then they invite their people to engage with our AI, which we call Tuffy. And that was named by a user.

So an employee will get an email that says, hey, we’d like to offer you the opportunity to talk with Tuffy. Click to go, set up an account. They sign up though with their personal email and their personal tech, their personal phone number, their personal device, and we tell them, and the organization tells them.

Please do this so that you have complete control over all of your own data. This is a safe place where you can go have a conversation about anything and not worry about saying things the wrong way or asking a dumb question. It’s really very much a place where they can feel comfortable with whatever is going on.

EL KALIOUBY: I think this is really important because trust is at the center of this — we’re gonna come back to trust a little later on. But if I have any inkling that the company might get access to my conversation with Tuffy, I’m just not gonna go to it. So I think it’s really smart that you did that outside of the company, that you could do this just kind of on your own personal device.

JAN: Yeah, absolutely. And we’re really providing the trust to both the employer and the employee. So first of all, a lot of people in the organization would never say anything. We know that people don’t necessarily tell the truth on employee surveys, and they’re afraid to say what’s really going on.

And if they did any of this on the company’s devices, they would have access to that data, whether they were looking at it or not, which just creates an air of concern. And for the organization, they want to be able to understand what’s going on, but they don’t want to have the whole mess exposed to them either.

That’s a danger to the organization. It’s a risk. So for them, having a bit of a barrier and saying, go here and have the conversations you need to have, is actually very helpful for the organization as well.

Copy LinkWhat employees actually need help with at work

EL KALIOUBY: Can you share some of your key learnings? Like, how are people using Tuffy? What are they going for?

JAN: So first I will say about a third of the challenges are really policy related — explain how FMLA works, or I don’t understand how to get my corporate card, all the kind of operational things. Do I get a lactation break? Things like that. It can get more complicated — what are the gender rules now, especially with a new administration? What are the bathroom rules? So there’s a lot of interesting change happening where each organization can define how they want to address those issues.

And the one example I use often is nepotism. So in a family owned business, nepotism is expected and positive. And that kind of organization would want to potentially explain that. And in other organizations it’s frowned upon or against policy. So there’s a customization there and I think a lot of employees have questions about things — what is ethical, what is allowed — all of that we can answer very well and very quickly.

The other two thirds of the issues are the really sticky and interesting things. So you could think about a manager relationship and all the different types of managers. Maybe a micromanager, maybe there’s some toxic behavior, maybe there’s favoritism, maybe someone just has poor communication style and choices. And the alternative there often is an employee calling a manager and saying, hey, can I have five minutes of your time? Can we have a quick coffee? I’m confused. Of course that would normally happen, but think about the hours and hours of time that it ultimately takes for managers to have all of those conversations. We’re just saying Tuffy plays the role of being the first point of contact. When you’re confused or you’re struggling, come to Tuffy first so that when you go talk to your manager or you go talk to someone else in the organization, you’re fully prepared and you’re focused on the right things.

EL KALIOUBY: How does Tough Day deal with issues that have legal repercussions? So, for example, a sexual assault — like what would it say?

JAN: Yeah, so we are a communication platform and not a reporting platform. So I think this also provides the workforce a safe place to go to talk about what’s going on. And if people do come in and there’s something egregious — someone was assaulted — the AI will be very, very curious and ask for more context, collect more information, and ultimately diagnose the situation as something that should be reported.

But we are providing the agency to the individual employee, so we want to build their confidence and help them understand why it should be reported, why the company really wants them to report it, how they should report it, and what to expect in the process. So, as an example, a lot of people will worry, if I report this, will the other person be alerted?

Well, the reality is they’re gonna be part of an investigation, but there are a lot of ways that you can engage with the employee and support them in the process so that they feel comfortable with that, and ultimately they’ll make that decision.

Copy LinkWhy companies benefit from AI support beyond cost savings

EL KALIOUBY: So we can see the benefit for employees. Why would a company be incentivized to do this? Is it mostly cost effectiveness and cutting costs, or when you sell Tough Day, what is the value proposition to an organization?

JAN: Well, there is a lot of efficiency. So going back to that use case of the manager spending a lot of time having a lot of conversations — you could probably spend 150% of your time answering questions or taking coffees and having coaching sessions, mentoring sessions. A lot of what is discussed can probably be handled by the AI.

So if you can cut back 80% of those questions, that’s a big savings. And that’s really the value prop. The other value prop is retention. So if you have these challenges sort of festering and people are not getting their problems resolved, even if they’re small, they tend to quiet quit or take PTO or call in.

EL KALIOUBY: Quiet quit, by the way?

JAN: Quiet quitting is when you are doing the very minimum in your job so that you don’t get fired, but you’re not very productive.

Exactly. You’re not engaged. You’re just doing the minimum. And you know, that’s where if you can get that work time back, great, but it also results in regrettable attrition.

And often organizations don’t know until it’s too late — often in the exit interview — that something was going on. So this is an opportunity to serve people better, listen and learn from what’s going on. We provide anonymized data to the organization so that they use that like a weather vane to know where the problems in their organization are so they can go address them.

And as they do, they improve the employee experience, ultimately improve retention. And the cost of replacing someone who attrits is two times their salary. So you can think of losing one good employee that might be making a hundred thousand dollars a year as costing $200,000 a year.

EL KALIOUBY: After a short break, we talk about the trend towards conscious un-bossing popular with younger professionals. And this new concept might just change the workplace as we know it. Stay with us.

[AD BREAK]

Copy LinkWhy younger workers are rejecting traditional management

EL KALIOUBY: So I wanna zoom out. Tough Day is coming out at this really interesting moment when it comes to the future of work. So I wanna kind of dissect some of the current trends that you’re seeing around the future of work and how, I guess, Tough Day is helping address some of these concerns or some of these opportunities that these trends are creating.

So the first is this whole idea of conscious un-bossing, which I only recently heard about. And it’s basically a trend where young professionals, like my daughter — she’s about to graduate from college — essentially don’t want traditional managerial roles, and they don’t also want to have managers. So tell us more about that. What is that?

JAN: Yes. 72% of Gen Z says that they are consciously un-bossing. And to your point, they do not want to have a manager. And they do not want to be a manager. And the reason is, if you dissect what has been modeled for them by their managers, it’s burnout, stress, a not very enjoyable job, number one. But also, no matter how hard they try, a lot of them still get fired. Their employee reviews are bad. They’re just like, why would I even want to do that? And the truth of the matter is, most people who become people managers didn’t do it because they wanted to become people managers. They did it because it was the next step.

EL KALIOUBY: The only way to, like, yeah.

JAN: Like what you have to do to get promoted or what you have to do to make more money. So 82% of people are what they call accidental managers, and they just don’t have the capabilities, or frankly, many of them, the interest. So Gen Z is saying, I’m happy to manage myself. I can get the information that I need.

I know how to network. I am very open to feedback and I can get feedback from anyone. Why does there have to be this formal structure? And I have to say, as we make our way in the world with Tough Day in one-on-one conversations and behind closed doors, everyone agrees the role of the manager is kind of outdated. And great managers do exist, and the problem with the great managers is they don’t scale. And so if we could give them a way to augment themselves and create their sort of digital replica so that they have a partner — they can spend their quality time having those coaching conversations and great conversations that they should be having with their team, and then offset the rest of it with an AI that will help improve the experience for everyone.

The other side of the coin is bad managers and mediocre managers who are actually costing the company time, talent, and money.

EL KALIOUBY: Yeah, we’re just gonna use Tuffy instead. So the second kind of interesting trend that is also very related to what Tough Day is doing is this idea of the great flattening. So tell us what that is.

JAN: Yes. So organizations are flattening out the middle layers, and Gartner actually has some research recently that says 20% of companies are getting rid of 50% of middle management in the next two years.

And I think part of that is because we’ve spent so much money trying to train managers to be good managers, and we’re not showing a return on that learning. And organizations are looking for other answers. So I think it might be a little bit challenging right now, but actually if you think about the real role of the manager, a lot of those things can be outsourced to AI.

Copy LinkHow shifting HR and DEI pressures are reshaping work

EL KALIOUBY: So there’s another shift happening specifically in the HR landscape. A LinkedIn survey found that HR had the highest turnover rate out of jobs that they tracked. Why do you think that is the case and does that kind of solidify the opportunity for Tough Day?

JAN: I will say in terms of HR, we’re seeing a few different things. One is, everyone who goes into HR goes in thinking they’re people-focused, right? Initially you think, I’m gonna do a good thing. I’m going into this field. And as you get into it, it may or may not be your cup of tea. It is largely about protecting the organization, and that’s great. We need people to design employee experiences and deliver services and measure all of that. It may or may not have been why you got into HR in the first place.

So I think there’s some of that. I think there’s so much stress and pressure on HR. The typical first point of contact for an employee is their manager, and if that piece is broken, one of the next calls is HR. So they’re getting bombarded with all of these challenges that are really hard and emotionally draining. So I think there’s some burnout there. And then in terms of all the solutions, how do you solve for all of this? It’s just hard work. HR business partners are the heroes of an organization.

They’re doing lots of great work. But I think, depending on the state of the organization, what resources they have available, what their benefits are, every organization has to figure out what’s the right recipe for their culture and organization. And that might be part of it.

EL KALIOUBY: Now we’re also kind of in this moment in time where we’re seeing top-down pushback on diversity, equity, and inclusion efforts. We’re seeing this at the federal level, of course — President Trump is seeking to end government support for programs promoting DEI. We’ve already seen a lot of references to DEI taken offline on federal websites. And this is all being contested in the courts. Of course, we’ll see what happens, but we have seen a ripple effect already at companies like Meta and Google, where they’ve rolled back some of these efforts. I would imagine that a lot of the concerns that show up in the workplace are related to equity at work and inclusion. What do you think of all of that? And again, how can Tough Day help?

JAN: Yeah. On the topic of DEI, when an organization works with Tough Day, we are tracking what is happening in the federal government and we’re tracking those trends. And this is an area where every organization has to do a bit of their own way-finding. At the federal level there are things going on, but then also at the state level.

So state by state, there’s new AI law, there’s also new DEI-related regulations. So right now we’re at a point where organizations have to navigate a lot of that on their own. And you see companies like Costco and Apple and Microsoft doubling down on DEI and saying, you know what, this is super valuable for our organization.

It’s just part of our values. And we have to invite diversity into the organization to be a strong, resilient organization and to serve our customers best. And then on the other end of the spectrum, you have organizations saying, I just don’t want to fight this machine. Some are saying we don’t have to call it DEI — like, we love the diversity and what have you, but maybe we’ve moved beyond this language. So I think there’s some of that where they’re kind of in the middle, and then you have organizations just saying it’s all out the window. So when we work with companies, we ask them, how do you want to handle things like bathroom issues? And there we’re customizing the experience and the answer for each organization. Even though a lot of the inspiration for Tough Day came from looking at underrepresented groups and their challenges, we continue to see that none of these issues are just about one population. And it just doesn’t matter who the person is. I think it matters how you handle the situation and ultimately we want to help people have better communications and interpersonal relationships and working relationships. And that’s the point.

EL KALIOUBY: But to do this, Tough Day needs to train Tuffy with each company’s values and cultural norms. So how do they do that and how are they any different from, say, Microsoft’s Co-pilot? We get to that after a short break.

[AD BREAK]

Copy LinkWhat makes Tuffy different from generic workplace AI

EL KALIOUBY: Let’s go behind the scenes and kind of geek out for a bit about how you’ve built Tuffy. So the first question I have is why can’t a company just take, say, a Microsoft Co-pilot and feed it its own policies and employment law and basically replicate what you guys have done. So just talk to us a little bit about the technology behind the scenes and what is your competitive moat and differentiator against, say, a Microsoft Copilot.

JAN: Yeah. So number one, coming back to the issue of trust, it’s really hard to have any of these conversations internally and know that your organization can see the data. You just can’t have that level of trust with an employee on the stickiest, trickiest kinds of challenges.

So I don’t think a Microsoft Copilot is going to be able to have that level of relationship and trust.

So that’s one. Number two, we have partnerships with a lot of content partners that have gated content. They will not allow the big LLMs, including OpenAI or Perplexity or with Copilot. They just don’t have the content.

EL KALIOUBY: What are examples of that?

JAN: So one of my favorite examples is Charter. Charterworks.com is kind of an up-and-coming thought leader. There are a bunch of journalists from Wired Magazine and Wall Street Journal, New York Times, the Atlantic. These are serious professionals that all they do is research and report on work. And they have amazing content and we have a partnership with them.

So we have all of their content that is subscription based. So that’s an example of how we make our platform smarter.

EL KALIOUBY: Yeah. Does Tuffy have a personality? Is it culturally sensitive or culturally specific?

JAN: Yes. Yes. So first, Tuffy has values and Tuffy is very curious. So our first phase of development, we did what’s called Wizard of Oz testing.

EL KALIOUBY: Explain what that is for people who are not familiar with it? That’s cool.

JAN: So sitting around a table at any given time during this process, we would have five or six experts — always a lawyer, an HR person, a manager, and a therapist. And we asked early testers to interact with the platform so that they would come in and ask questions. We told them it’s not really an AI, you’ll interact with it like AI, but we actually have these human experts behind the scenes.

EL KALIOUBY: That’s the Wizard of Oz piece, right?

JAN: Yeah, the Wizard of Oz. And the really cool thing is most people forgot they were actually talking to humans. What happened for all of us sitting around the table is that we would decide very quickly who was the best person to answer this question or tackle it.

And what that meant was usually that person was asking five or six different questions to really understand. So just like if you go to a lawyer and you ask a question, they’re not giving you advice at first — they’re going to ask you more questions. That is really how we developed Tuffy to be curious and get more context and then diagnose what the situation is.

Because so many people will ask one thing and they’re asking the wrong question, or they don’t really know what the issue is. So Tuffy is generically very curious and wants to understand a situation and bring empathy to the dialogue. But in terms of your question about cultural customization — when working with the Hawaii Employers Council, which is a network of 700 companies all based in Hawaii representing 170,000 workers.

And HEC is using Tough Day to join its HR consulting team and provide advice to those employees, but also as a go-to-market partner. It’s interesting. We started down the path and we got some feedback that maybe Tuffy sounded a little too New York.

So they asked could we make Tuffy seem more Hawaiian. And that was not language — that was really culture and value. So they asked us if we could apply the “aloha spirit” to our AI.

EL KALIOUBY: How did you do that?

JAN: So first they did give us redacted transcripts of conversations between really talented and successful HR leaders, managers, and others so we could understand the cultural tone of the dialogue. After we customized the experience, users noted it right away. We got great feedback in the platform, and our helpfulness rating went from 88% to 99.6%.

Copy LinkHow trust and safety are built into high stakes AI

EL KALIOUBY: And actually, coming back to this — because I’m always thinking, okay, what is a company’s competitive moat? That kind of training and that kind of data is extremely unique. Like, you can’t scrape the internet for this kind of data, and that’s very powerful. So one of the things that we’re very passionate about on this podcast is how do we build responsible and trustworthy AI? And you are in a very tricky space, right? Like people’s careers are at stake, people’s livelihoods are at stake. And a recent Pew study found that 52% of employees say they’re worried about the future impact of AI use in the workplace. So how are you building trust into Tuffy?

JAN: Yeah. Well, first I believe — we at Tough Day believe — that AI is a more human-positive solution often than a human. And I will say I believe in humans. Don’t get me wrong. I’m a human-centered person. But I will say humans are messy. Humans have bad days. Humans can make mistakes and do make mistakes. But with the AI, we want to make sure that the AI is not biased, that the AI has safety guardrails, that we know the AI has a hallucination defense system. In certain circumstances — especially when you’re calling on a policy or a law — we want to make sure this is not AI making it up.

It needs to be specific. That’s part of it. The other part is all that red teaming and knowing where all the pitfalls can be.

So there’s a lot of scenario planning to think through — what could possibly happen? You want to think through what are all the things that could go wrong and build the right knowledge and guardrails into the product.

Copy LinkHow AI could reduce stress and unlock human potential

EL KALIOUBY: Yeah, so final question. I am very passionate about this idea of human-centric AI — AI that is gonna help unlock human potential. It’s gonna augment and amplify our abilities, not replace us. But I do spend a fair amount of time thinking about this following question: in this age of AI, where AI can be patient and curious and empathetic, like Tuffy, what does it mean to be human?

JAN: I do think being human — if you think about all the human needs, like even Maslow’s hierarchy of needs — the positive part about having AI involved is a lot of our physiological needs, and even community needs in terms of interpersonal skills and understanding one another and being curious and all of that, we can accelerate or kind of meet those needs faster. That frees us up for those higher level needs of learning and innovation and problem solving.

And so I think the communication and community is a human thing, but I think it’s going to be accelerated and improved by technology, including AI. I do think we might actually find ourselves being more empathetic because we have relationships with AI that are empathetic.

I also think there’s another interesting thing in talking with neuroscientists. When humans are under stress, our brains actually shrink. So if we think about what’s happening to people in the world of work or even in our communities, the more stressed out we are, the less capable we are to use the most creative parts of our brain. And I think what we’ll see is that this technology, by solving some of the things that cause us stress, is going to make our brains actually grow and be more intuitive and more empathetic and more creative and innovative to solve the biggest problems in the world. I’m absolutely a human optimist.

EL KALIOUBY: That’s the perfect way to end our conversation — AI that is gonna help expand our brains, like literally. Thank you for joining us, Catherine. This was awesome. Thank you.

JAN: Thank you so much, Rona. Such a pleasure.

EL KALIOUBY: One of my key learnings in starting and scaling my company Affectiva – was that the biggest challenge was not creating Emotion AI, but it was dealing with people – and all of the human messiness that comes with it.

What I love about Tough Day is that their AI platform handles that human messiness in a safe and supportive way. They can manage these issues before they really escalate, which allows professionals to focus on the work.

Often the conversation about AI and the future of work is focused on job loss. But the conversation shouldn’t stop there. It’s so important for us to ask, how can AI augment our abilities at work? What are the areas of work that slow us down? Where would we rather be focusing our energy?

One of these areas is management. But there are so many other opportunities – from scheduling, and automating bureaucratic tasks to maybe even putting in lunch orders before you get hangry.