If AI can do more of the work, what should people be doing instead?

Transcript

Miguel: Hello, welcome again to a new episode of The Power of Possibility, a podcast from SPS. I'm Miguel, and today we're going to talk about how artificial intelligence is showing up in the workplace and what it actually means for employees and how organizations can use AI to support people, skills, and performance rather than creating fear or disruption. I'm joined by two leaders working closely with organizations in these kind of topics.

First, I would like to introduce to you Jess Larsen. She's the founder of Thriving Humans and with extensive experience advising organizations on how people and work and technology evolve together. Correct me if I'm wrong.

Jess: Thank you. Sounds good.

Miguel: Thank you. And we also have once more Nicole Mangarella. She's the head of global technology and innovation at SPS. Welcome, Nicole.

She leads workplace technology and innovation and initiatives focused on enabling employees and improving day-to-day experience at work. So, thank you for being here today. Maybe we should get started with the first question, an open question for both of you. You can decide who wants to start with this question, which is in in day-to-day office work, when what do you see AI removing real friction without reducing the value of human human roles?

Nicole: So. We see an increase in the number of employees overall that are using AI day-to-day. We we recently published  a survey in conjunction with Work Tech, where we found that about seventy-five percent of employees now report using AI at some level in some capacity in their day-to-day work. That's up about 16% from where it was last year. So it seems like more and more people are engaging with the tools and those tools are being made more readily available. And when we look at data from a variety of different surveys, and different researchs, it shows that a lot of people are using it still for basic summarization, research, rewriting and translation types of things. But we are starting to see more adoption of more complex AI workflows as businesses start to integrate it beyond just giving their teams access to the chat sorts of tools that are available.

Jess: Yeah, I would agree and add to that, I'd say we're sitting in the seat, talking with clients and working with clients and you know actually hearing what companies are encountering and often the conversations we're having are with AI leaders or people accountable for the change and sometimes what they're sharing is frustration around how to encourage that adoption. Like so the people who are closest to understanding the technology and introducing it into companies often can see that it promises these new capabilities and they've got more exposure to what AI tooling can deliver.

What we can see is there’s such a spread of experience, there’s such a spread of readiness and different journeys that companies are on. And what I say to them is both of those things are true at the same time. There’ll be people who are at higher adoption spaces and there’ll be people who are lagging behind, and you can learn from both of those cases.Jess Larsen

But often the people who are charged with introducing or transforming their AI enablement aren't actually in the delivery seats. So, they sometimes are quite removed from, you know, how do we actually make this work in marketing or how do we make it work in sales or how do we make it work in engineering or different parts of their business. And what we find is there's such a spread of experience.

There's such a spread of readiness and I guess different journeys that companies are on. And so often companies will be listening to what's happening in their own industry space or in, you know, peers that they're relating to. But I think what we can see, and so they listen to the kind of the media narrative of, you know, "there's this massive disruptive impact that AI is having", and then on the other hand, "AI is having no impact whatsoever and it's all a lie".

And what I say to them is both of those things are true at the same time, is that we have this huge variety of experiences happening. The risk is that people pick the bubble they're in and they decide that's the truth. And so if they get nothing else but this from this conversation, I would really challenge them to say, wherever you are on that journey, it's true that there'll be people who are at higher adoption spaces and there'll be people who are laggarding behind, and you can learn from both of those cases. You know, we we've talked quite a bit like why is it good to be out the front of the pack and why is it sometimes better to be at the back of the pack? yeah. So lots of a huge amounts of variation.

Miguel: Many companies still approach AI mainly as a as a cost-cutting tool. why can that way of thinking be risky in the end?

Jess: Okay, I'm gonna take this one first. So this is my sweet spot. I come from a career, you know, decades of experience in leading change initiatives and particularly the people dimension of those changes. And this is not the first time in my career that we're working in an era of technology affecting work. you know, those of us who've been around a bit longer like will remember cloud transformation journeys like ten, fifteen years ago.

Anyone coming from the transformation landscape or the change ownership landscape is really familiar with this concept of you have tech comes into a business and you have to make everybody pay attention to that and you have to somehow find a way that they're gonna care about it.

Even people who feel a long way away from it and often the problem is how do you make it relevant to them, and help them both understand what's happening and then actually care about it, right? And it's a classic change problem of, you know, learning as leadership, running these kind of changes of what is your story that you're telling? And by story I don't mean the lie. I mean like hopefully it's true story. But it it is a story in the sense of like we're humans and we are very attached to narrative.

Like, it makes sense for people if we can turn it into a conversation that people can relate to and they can find themselves in. And I guess tapping into that kind of emotional experience that people have through change is absolutely critical. And that's something we're definitely finding, you know, in supporting clients that we work very much with technology experts, often who can see the potential of what can happen here.

And then time and time again, they come to us and say, actually the real problem is this people dimension, like,  how the hell do we get these people to understand what's going on, to want to do these things, like actually make it happen, so that landscape of like behavioral impact, you know, how do you get those pieces of your business to pull together is the work we do.

Nicole: Yeah, and I think the the cost cutting mindset it comes from how people looked at some legacy software transformation. Typically it was this process is very manual and it's very strict, it's very rigid, we need more flexibility. So we're going to change ERP systems, we're gonna move to the cloud, we're gonna create a data lake and get rid of all of these different data silos. And the goal typically was to calculate an ROI that would show that you had reduced cost in one area or another. And I think there's there's two aspects of AI that change that conversation.

One, the ROI is difficult to calculate on AI because the cost structure keeps changing. The cost per token has been going down, but the licensing models get more expensive depending on how you're deploying it and which tools you're picking. So, the complexity of trying to calculate how much cost you're actually reducing or at what point you're gonna break even is really difficult to do these days.

The other aspect of it is that while you may be reducing cost in one area, or you may have an opportunity to to take out headcount or optimize a team, you also have the opportunity to then reinvest that and expand what people are doing. And the example that comes to mind is IKEA deployed AI in their customer service space. They said, you know, if we put really, really good AI, and conversational AI, behind our customer support agents, we can reduce the time that it takes them to handle a request, and therefore they can handle more requests per day. And so we don't necessarily need as many people in support.

And so, you could make the decision to say, we're just gonna reduce that headcount and we're gonna take a cost savings and look at it that way. But what they decided to do is historically they had wanted to be able to offer interior design sort of consulting services, because if you're a furniture and a design company, how could you potentially grow your revenue? You could start a a complementary business and a complimentary service to go with it.

So, those people that they had in customer service who knew their product lines really well, knew their customers really well, they took that knowledge and that capability and used AI to then upscale and reskill that headcount into a new service for their business. And I think that's a really great example of AI as something that unlocks more possibilities for what you can do as a business, not just, I can reduce my headcount and I can reduce my cost.

Jess: And I I'm gonna add to that, right? Because this is the temptation that we see businesses walking toward this trap over and over again. I mean it we had talked in preparation on the disaster story that Klarna had of going out to the market and declaring they were losing, I think, seven hundred roles in twenty twenty two due to operational efficiencies, you know, largely driven by AI and then having to wind that back a couple of years later and recognizing the damage they had done like internally to their culture primarily. But I think most companies that I talk to can relate to that threat, that risk and some of them they've gone and made declarations, but a lot are sitting on the fence saying, we think we're going to achieve productivity gains with this, we don't know how to walk that tightrope of how do we kind of get from here to there without blowing up our company culture along the way.

Because ultimately the people who are going to be in that future state company are in the room already. You know, that's going to have a massive impact on that talent landscape, let alone the people that you let go that go out into the market that massively damage your employer brand. And so we've got a couple of case studies of businesses who've done that damage, and of course, there's a huge narrative externally. There's no shortage of articles like why not to do this? And what I would want to add, because the example you've given with IKEA is so nice, there are quite a few that we have seen, and certainly one of the businesses I was in, in fact, a couple of the businesses I was in, we did work, you know, we did a deep landscape of internal mobility, so looking really carefully at like what is talent in this organization and how can we make that more malleable? How can we make that more transferable? How can we take the the strengths that we have in the institutional knowledge and the types of behaviors that we have attracted in what we consider a really talent dense environment?

Like we think the people that work here are fantastic, but they may not have exactly the skills we need. To your point with IKEA's example, what are the potential landscape of opportunities that we could see coming up?

How can we evolve our people's skills towards that? How can we maybe take sacrifices so we don't go and pick the ready-made candidate from externally, but we do the work internally. We actually try to train that person up or we give them exposure to these opportunities.

In one business that I was in, we doubled the rate of internal mobility every year. It was a huge, like huge effort. It takes a lot of work. But the knock on of that was that engagement went up twelve percent. Like huge.

And people recognize it, and that was a drumbeat of messaging to say, this is what we're doing, this is why we're doing it. You've told us that you want more career progression. These are all the things we're going to do and all the things we're going to measure. And it was really hard.

It's really hard holding managers to the fire to say you have to exhaust the internal talent pools before you're allowed to look outside. And they're convinced that they have to get this external talent in, and I would argue that's the work that I would hope, that most businesses are like really digging deep on that stuff because if you throw those people out the door today, the message you're sending is ..., you know, you've got values on the wall, you're relying on the trust and respect that your people have for your ambitions. But once you start treating people as commodities and you send those messages of, hey, everyone here is disposable, you've lost a huge amount of people's engagement and productivity and performance and that is incredibly hard to recover from. Like it's always easier to get the first time. It's more than twice as much work to make it back if you damage it.

Nicole: It's interesting too when you look at how people are measuring the gains that they are starting to see in AI. And almost every measurement that I've seen in the space has been fractional hours saved from an employee's day. So you have depending on the department and depending on the level of seniority and depending on the job, it's nine hours a week, two hours a week, twelve hours a week. And it's it's not the full amount for that employee, but I think it was Shopify who came out and said we want this to be employees opportunity to stretch and to use the knowledge that they have about the company, about the culture, about what our customers need, what our partners need as a business, and use that extra two hours a week to play around with some of the AI tools and maybe create something that you would have never thought you had the potential to create.

But you know the business, you know the culture, you're invested in what we're doing, and then you get more innovation and better ideas out of that. I haven't seen the results from Shopify on the new ideas that have created with that, but it is it's interesting when you look at ... if you have an extra six hours of time, 10 hours of time, how do you help people then use that to amplify what they can bring to the company?

Jess: Well that's massive. I mean that's what every company needs. Like we definitely have worked with clients who have dedicated learning time. So they've taken from workforce planning capacity for learning, capacity for experimentation. And arguably, like anyone who's worked in change practice before knows like this stuff takes time. It takes resources. So any company that is proposing that they're on this journey and they're not allowing for that flex and stretch; its a really good point.

It's a really good point of linking together, okay, we're seeing the advantages and the gains. It has to then be allocated into this learning and growth because where companies have to go to make this stuff work is into this landscape of experimentation and and innovation, right? And we know innovation, like for it to succeed, you know, anyone who's worked in design or in, you know, tech's been a really good playground for this for a long time.

You have to have that ability to fail. And so ability to fail, the conditions for that are you need the resources, you need the time, you need the culture of trust, and you need like clear guardrails and sandboxing. So you've got agreement of what are we allowed to break, what are we not allowed to break, but those aspects as you're describing, where you're starting to see the productivity gains, like what a great model to load that then into the learning culture. Because I've definitely spoken with clients who are struggling because they haven't made that clear link of we're trying to get our people to this new capability set. We're trying to skill them, but actually we haven't allocated the time, you know, to do that. Or we haven't taken work off the table.

Miguel: You were talking about the the layoffs that we've seen in the in the big large companies, tech companies, while artificial intelligence keeps growing. And what message does that send to employees? I mean, how can it hurt companies in the long run?

Jess: Yeah, I mean, this is such an interesting one because I would say for years there's been this tension in the world of talent, you know, talent specialists and in the world of HR people, of that there has been not enough of the highest quality talent. So you're you're kind of coming at the talent landscape with a very competitive mindset. And then we have this tension of real job lossage.

Yeah, I definitely know of scores of people who are in the market now who didn't expect to be as an indirect or direct result of AI implementation. So they're absolute and I actually personally believe that the rate of layoffs is possibly higher than we're kind of seeing come through the stats because there's a bit of a shadow workforce of people going into consultancy or into, you know, contract work or kind of not picking up on the official records a bit.

And sometimes it's hard to track those numbers accurately. But regardless, they're definitely real. So the leg is like the tension that we used to have to say, you know, we have to do everything we can to protect our brand because it's that important.

That argument is actually being challenged a bit inside of companies to say, well, you know, if there's all these people in the market, does it really matter that much what our brand is? Does it really matter if we're hyper competitive, you know, actually maybe we're gonna get pick of the month and we don't have to struggle with that anymore?. I think that that's a bit of a dangerous to take for a couple of reasons: one is that with your employer brand, it matters what the external landscape sees you as, but it is more profoundly it's your internal landscape.

It's like this is your existing talent. Like the people who are already with you in the room. And in my experience working with you know, leadership bodies, you know, there's often they tend to overestimate the reception of the culture and how well people are finding things. And often, you know, you're calm if you work with engagement survey data or you work with sentiment data and you bring back this is what people actually think.

And you get leaders go why would they think that? You know, it's great here. They tend to have this bias of positivity, right? They think things because they are arguably in privileged seats. You know, they get more autonomy. You know, the things that are harder often down the tree is harder for leaders to accept, but it's the truth.

And I would argue that like the mode that most companies on these journeys are needing to go on, which is like almost a loosening of their structures and a figuring out like how can this technology, how can we bed this tech into our organizations? The answers to those questions are gonna come from the people in the room.

Like I've I've worked with leadership teams and AI owners of these implementation plans saying, I don't know how this tech's gonna like affect people in sales. I don't know how it's gonna affect like it's the people doing that work that need to calm and they need to come with the why, the arguably like what is it we're trying to solve for? They still own that business, that piece of business delivery. And so it's a always a partnership of understanding what the tech can do, but then still meeting that delivery need for the business. And I think that if we underestimate how important it is to emotionally take people with us on that journey, like we already have some really interesting stats.

We're looking at over the past few months have come out around you know active sabotage of employees around AI. It's on the is on the increase, which is hugely concerning. In the US it looks the worst out of all of the geographies. And the stats I saw are I think sixty-six percent of specific generation of the Gen Z generation reporting that they are telling the survey that they're actively sabotaging and that might be that they're you know disclosing data out to AI that they know that they shouldn't be under the company policy or that they're doing things to have the AI output look less successful than it actually is.

So having employees actively sabotage a change program is a major red flag. And I would argue at that point, you really want to be doing some honesty assessments of how can we repair trust in the delivery of this change.

And we're talking about people losing their jobs, you know, that's got to be one of the biggest threats that any of us are going to have at work. And yeah, it's really hard for companies to go and provide a hundred percent security to people, but arguably we know that the route to security is going to be the value that the person's going to be in the business model. And so with most of our clients, we're working with how do we help people in the room understand that emotional journey that people are on? How can they with honest reactions, you know, not trying to build an artificial story, but genuinely tell the story of we want you to be on that journey with us?

We want to upskill you as we go. Like these are the pieces we know on that journey so far, but you're going to come with some of the answers to that. And so then you've got a collaborative change exercise going through. And I think I know from your background, Nicole, as we've talked about in tech transformation, like that work of buy and that work of emotional kind of engagement is underpins everything.

Miguel: In the end I guess that there are certain kinds of tasks that we should do in order to make employees to stay confident with AI. I mean what what new skills should employees start building now to stay confident? And relevant.

Nicole: I think that depends overall on the role and the the specific function. I will say, and and I saw the same study on the sabotage and I think it was as high as one in three at a company are actively sabotaging AI initiatives because sometimes the way those initiatives are messaged are, you know, it sort of feels like you're sharpening the axe that's gonna chop off your head. And it does make it hard to want to learn those new skills or to believe that learning that skill is going to help you advance.

There’s a study on sabotage that says one in three employees at a company is actively sabotaging AI initiatives because sometimes the way those initiatives are messaged makes it hard to want to learn those new skills or to believe that learning those skills is going to help you advance.Nicole Mangarella

And so I think there's critical thinking skills will never go out of style. I think the ability to learn how to interact with it in a safe way and in an effective way is is something that's baseline. There is a wealth of information out there for prompting tutorials. There are lots of different ways to learn about how you can use the technology, and there many of them are free.

In addition to the company sponsored initiatives that are going to be more specific to use cases. But from a skills standpoint, I think there is that real concern that if I don't learn these skills, I'm going to be the lowest performing employee. However, performance is measured within that organization. And so the best way to learn about technology is a hands on way to test it out, to try what's available.

I think companies that have created those types of environments where they've given access to tools and let people play around with it, it's helpful for people from an introductory standpoint, but then you do need to have that next level up for people who are, like yeah, I'm comfortable with this, I'm using it in my daily outside of work life, you know, how can I make sure that this is applicable to my job and how can I make sure that that's recognized within the business as me being more effective in my role and maybe ready for the next opportunity.

Jess: Mm. We've I was gonna say we've actually developed a mindset shift model for our clients and critical thinking is is up there. That's one of the pillars. And critical thinking, we talk about that as I liken this to when I've delivered leadership development work in the past.

You talk about when someone makes a transition from a frontline manager to a a manager of managers, that suddenly you're in a position where you're accountable for other people's delivery who are very capable, but you can't oversee everything that they're doing. And there's like a another layer through them. And we like the potential of a gen tech structure to that equation of what does that feel like? And when delivering that leadership development over the years, I've had many conversations with leaders trying to make that transition and be like, how the hell am I supposed to be successful? How am I supposed to trust these people? Like what how do we make this work? Right? Because it's a very uncomfortable position to be in to kind of let go of control, but then have enough governance and accountability that it actually is going to function successfully.

And I would argue that one of the tenets that we take our clients through is that the potential of AI enablement arguably kind of shifts everyone in our structures into that mode.

All of a sudden those capabilities that we used to consider just for the our senior leadership, everyone has to start adopting that thinking and those skills of how do I like be accountable for things that I'm not directly controlling. And I've heard quite a few themes of you know, preserving this really tight relationship to the human is always accountable for that decision.

And so how do you put people in positions that they feel comfortable with that? You know, how much governance then is needed through that chain? because that's definitely the world that some people are living in today. And I don't know that we've done enough cultural work and behavioral work with people to really make that as clear as it as it could be. And I suspect that we will see, you know, horror stories of things gone wrong, of outcomes. And I think people right at the front line, like certainly through engineering structures I've spoken to, you know, there's some real concern about how far agentic delivery is is going and and you know, the discomfort of being accountable for things that they're losing direct control over. And I think we've got more work to do as businesses with that of how do we design that out? 
 

Miguel: Now that you you've mentioned about critical thinking, which is a really a high task that we should keep, in a daily basis, there's like a feeling that sometimes when using artificial intelligence, employees can lose a little bit of abilities.

Jess: Yeah.

Nicole: Yeah.

Miguel: You feel a little dumb sometimes when when using too much artificial intelligence. So how can companies can balance this?

Jess: Yeah, I think this is huge, hugely risky. And I've definitely heard people increasingly start to talk about it, which is great. There there's a really good study actually in medical journals that came out about two months ago, and they have measured, I think it's radiographers or a specific skill set where they had started adopting AI and the individual's ability to assess reports. So what they would have done manually before, and then they started applying AI to that recognition pattern, is that then they tested the human's recognition pattern ability a couple of months. It was only three months through the test and it had dropped by I think it was four percentage points, which was considered statistically, you know, a really strong indicator of deficiency. So we know that from neurological understanding, that what we you know, our brain is a muscle, that what we don't use, our brain naturally recycles, it atrophies, we let go of those skills.

Same with memories. If we're not, you know, if we're not spending time in that cognition, then we lose those abilities. And I am very concerned as somebody with an an experience of designing work and designing workflows and working in parts of business where we design what people are doing and designing what behaviours, what skills do we need in our business, whether enough attention is being given to the risks of if we outsource this, are we okay with that, leaving the room in our company?

I did some work recently on apprenticeship flows, which was super interesting of what's happening for early career, a whole different conversation. But arguably like similar problem in a different dimension of trying to get these early career people exposure to stuff and that some businesses are doing workarounds where they are they're consciously choosing to preserve some pieces of work and saying, well this could go to AI, but actually it's so important that we have that developmental path, that we're gonna protect it over outsourcing it. So we're gonna over optimise for our internal capability over productivity. And I think that question should be like front and centre for anyone doing this redesign and what are we trading?

Nicole: I do think I saw it, it wasn't the one in science, but there was one in in MIT that was basically the conclusion was Chat GPT will rot your brain. which I think is interesting because I can remember like how many phone numbers can you recall off the top of your head? There's people that you talk to every day that at one point you would have been able to do like that. And now it's it's kind of a skill that we've let go of, but if you wanted to build a website, you would probably be better equipped to do that than it would be to recall a bunch of phone numbers. And so I think there is in some cases it's good that we're losing some of these skills that have been necessary in the past but aren't as necessary in the future.

I think the biggest piece that is the gap is the quality level because AI can definitely do it faster, and in some cases it can do something that looks like it's five people's worth of work, so it's a bit cheaper, but the quality gap is pretty high, especially when it's unassisted. I think that's where there's a lot of promise in AI agents, but it hasn't fully been realized independently yet. It's more of a collaborative effort. They had there was a study recently that Scale AI and the the Center for AI Safety did around the ability for online agents to take tasks that would normally be put on like a task rabbit or a fiver and to do those tasks to completion.

And it was things like logo design, game development, make a video out of this information...They started the study in October and they called it the Remote Labor Index, and they found that even the best models out there, if you gave them the task independently, could only fully complete it to the right level of human quality 2,5%  of the time. So it was very, very low. And then they redid the study again recently, and the best model is up to about four percent.

So, I do still think that knowing how to use AI has become its own skill in and of itself. And it's really that human plus AI augmentation that ends up getting you the best results. So, if it's things that you don't necessarily care about the quality, you just want it done, you have a high tolerance for issues and for inaccuracy.

Those are the best use cases where it's okay, this just has to get done and if there's a little bit of an hour, it's fine. But I think that skill of how you incorporate that in is as basic as, you know, people using learning to use graphical interfaces or learning to use command line or all of the computer skills that have come before this. It's just a new way to kind of leverage that technology and it becomes its skill in its own while some of the other skills go away. I do think the part on losing creativity and the ability to communicate is a little bit overblown because the writing that comes out of it is still not the same level of quality. And even if you have it, it's a fantastic brainstorming tool.

Ultimately I think what the MIT study concludes is that it's really helpful for getting that starting point and moving things forward, but the quality just still isn't there in a lot of areas compared to to what you get from people.

Jess: I think the relational piece and I don't know if I'm understanding you properly, but some of the skills that we have as humans are going to become much more important. Like so when you ask what are the skills that we need, there's been a lot of commentary around the need for empathy. And then there've been some interesting studies demonstrating that in fact AI is quite good at demonstrating empathy if we train it well. And I would ask like which which I think is quite threatening for people, right? And I think there's a distinction there. The study I'm talking about was one where they in fact were looking at doctors - and I think anybody who's had problems with their doctors around empathy will relate to this- training an AI tool to elicit empathetic reactions and the experience of patients was far more positive in using the tools than it was using the human being, and there are good reasons for that right as you think.

I always think this when I go into see the GP. It's like, I think, God, poor them. They're sitting in this incredibly stressed schedule. They're trying to deal with things with like often quite high threat for people that where people's emotional reactions are really triggered, and then they have to be quite transactional with it. And so there's this combination of factors that makes that work really hard. It's really hard for humans to be empathetic, right? It's a skill we know there's a deficiency, you know, it's acknowledged deficiency for people in the medical practices, right, which they've done a lot to try to counter over the last few decades.

And then it's so interesting to find a tool that in fact is a mechanical tool which can in fact deliver an experience of empathy for humans if we put enough of a framework in there for it to reference. The bit that I that I see as a continued gap is the relational empathy. So, the experience of a human. And this is where we go into a bit of a woo-woo landscape that we don't talk a lot about in business settings, but I think I'm predicting is going to become a lot more important.

It's something like the distinctions we talked about a lot through the pandemic when we all moved to remote. There was a lot of conversation, certainly in the HR circles and people circles, of the quality of relationships at work and what that was doing to employees engagement and kind of nervous system regulation and therefore productivity.

So, people were operating in a landscape that with tremendously high threat. Like whatever their personal situation was, you know, there's a lot of people feeling super freaked out and going into bad places, and the quality of people's relationships at work suddenly became really, really important. And then we all had to do it virtually. And at the time there was a real surge of learning about how do we engage more empathetically remotely? So, we put our cameras on, we try create these like human moments of connection that trigger an emotional reaction in us as as animals ultimately, right?

And I would argue we know that like when people get together physically that there's a whole different impact in the quality of relationship that can be created almost instantly, and I say in the world of work we've got more to go and learning more about that, embedding that into how we're operating, right? And this is kind of some of the work we've been playing with at Thriving Humans. It's really fascinating of like who are we as human animals?

Because we don't do enough when we think about work. We're very much in this cognitive headspace of solving practical problems, but not necessarily relating to ourselves as, hey, we're animalistic creatures. We have all these emotions going on all the time, which are impacting the quality of our thinking, the quality of our decisions, the quality of how we're communicating, and ultimately how we can succeed as a business. You know, but we don't talk about that a great deal.

In the workplace, there’s an opportunity there to say, how do we reconnect and preserve those moments of relational integrity? […] If the trust is not there, it doesn’t work. If the trust is there, you can do amazing thingsJess Larsen

I think it's very interesting that we're at this point of divergence that we have tools now that have been proven to deliver that empathy. But ultimately it does not create that relationship value. And we we've heard, you know, probably the disaster stories of, you know, particularly young people growing attachment to AI. We see that as a risk socially, you know, that's starting to be acknowledged. 

I would say in the workplace, there's an opportunity there to say, well, how do we reconnect and preserve those moments of relational integrity? Like how do we help people build deeper quality relationships at work with integrity and with trust? And all the factors we've talked about with change management, like, you know, anyone who's gone through hard change knows trust is the thing that has it work. You know, it's like where if the trust is not there, it doesn't work. If the trust is there, you can do amazing things.

It's this human experience of like, who are we as a team? Do I know this person? Do I understand them? You know, do I get them? Are they on the same page as me? Can I trust them? Like, do I get them as a human? So right back to our human skills.
 

Nicole: Yeah, we we see it a lot from a workplace services perspective in we do a lot of workplace ambassadors, which is, you know, they're the face that you see in the office that helps you access all the services that support you being able to do the work that you do without necessarily getting bogged down in the details on things. And the technology has been there for years to do most of it through digital ticketing systems. You can have a virtual receptionist... the tech is there.

You could. You probably shouldn't in some cases because, you know, an investment banker who's really stressed about getting a deck together for a meeting the next day wants to know and look at somebody who is accountable for what they're doing and AI is not accountable. So, that personal relationship of being able to say, Yep, I got you. We're gonna make sure this gets done. You focus on what you're gonna how you're gonna interact with the client tomorrow, get some sleep, you know, got the support structure in place.

Sometimes might just want to submit a ticket. But I think the variety of options that people want in a workplace now are I could do some of it with AI. I it sometimes I want a person involved in that. We always want a person behind the scenes. I think the human in the loop piece to get to the level of quality that we spoke about earlier is is still critical, but it is a matter of just because we could necessarily do it with tech doesn't mean we should.

Jess: Yeah.

Miguel: We've talked about the social aspect about AI, etc., but we just haven't talked about the implementation, right? And I want to know what should be first, changing the tools or changing like the process? The processes, the workflows.

Jess: Changing the tools or yeah, let me let me pose that back, maybe a different way. Changing the process or changing the tools, I think for me it's about learning the potential of the tools and then I think the thing that's super exciting with AI is the potential for design. Like I think with design is that we don't have enough of those skills in businesses in general. And I think now is the time. Any designers out there, now is your time. I think businesses desperately need the mindsets and the skills of how do we reinvent things? Like how do we reimagine like what could be possible? Because if we take an existing workflow and we take the pieces of it and we just put in automation, arguably, anyone who's gone deep in playing with AI capabilities will probably be like, "no!". You know, I've likened it to using Play-Doh. Like it's so variable. Like there's so many things that you can do that you couldn't do before. And so it's almost like seeing the potential of those things. And I find like what we were working with clients, some of it is upskilling on that potential, but some of it is just getting people in rooms and riffing. And so, you know, we actually bring design thinking journeys that we deliver workshops, we kind of do this work of like, what are your problems? And we look at these use cases and then let's just start riffing. And we start pushing, you know, do we run some innovation practices and things like, interestingly you were saying, AI is super good at coming up with ideas. I would argue that humans doing the hard work themselves on that is I think we need to do that. I think that we need to preserve that because, my experience of playing with innovation, you know, at one point I did some innovation training with MIT and we had this exercise where we were given, you know, half a day to come up with 20 ideas for such and such. 

And the next day there were we had to come up with fifty and it was like, "what?" The next day it was a hundred. It's like, "what?" And of course, the longer the list got, the more crazy you started getting. You're just like, right, well, let's just put it all down. And it's that level of ideation that you start seeing like no idea is about like just start playing. And our play comes out. And that those those muscles that when we've been working through the work environment of trying to be right and trying to be perfect and trying to do the best version of everything, you know, over and over again, that really stifles our natural innovation, our natural creativity. And I think what AI opens the door to is this whole different mode of creativity in business, because the cost has gone down, like the speed has gone down, and so trying to help our people think creatively, I think there's a lot of space for artists to come in the room for design experts, like we kinda need that thinking back in the room at the moment to challenge how we're thinking. Because if we just sit here with the tools and we start with that existing process and we redesign it, we're very likely missing out on, you know, umptean potential use cases because we haven't thought broadly enough.

Nicole: I agree with that in principle. I think the coming from the the technology transformation side, I think you don't start with either one, you start with why you're doing it. Because A) gets everybody on the bus, gets something that everybody can believe in, get you know, points you in the direction of what you're trying to actually accomplish, why you've actually gotten everybody in the room and taken their time and you want their ideas. And if you start from that why and that belief and you get alignment and buy in on that, then at this point, selfishly you sort of do both at the same time.

My background, I come from ERP implementation and transformation. So selfishly easily we got to rip out the guts of how the business worked and then redesign the processes based on the new system, based on what the capabilities were there, based on how the business worked, based on what they were trying to accomplish. It becomes very collaborative between the technology and what's possible with the technology and the process that you ultimately want. Traditional software, typically you would start with, here's my process design, and then you pick the tools at the very end.

I think with the flexibility and with the speed of the development of the capabilities of AI, you really do end up doing both interchangeably at the same time. But if you sit there and you look at it and you say, "well, we want AI in this process because we think that it's gonna help us be better", you need to go back to the drawing board and figure out why you think it's gonna be better and what that why is.

Maybe it could be "Hey, we think that we can reduce our headcount by 10%. We think that this will help us compete better because we can release products faster if you're in manufacturing", or we think that this is going to help us be more accurate and get better quality because the AI can do a first pass for us, and then by the time we get to the end of it, or for a lot of companies, there's also the opportunity of we think this will give us better data, coming out of that process at the end of the day that we can then use to figure out what are we doing well, what are we not doing well, where do we focus.

But I think it does happen really at the same time, but if you don't have the why as the guiding light and the thing that gets everybody on the bus, then you sort of just spend a lot of time and energy and you don't really accomplish things.

Jess: Yeah, I'd say what she said. Like do her thing first, and then do my thing. [laughs]

Nicole: But but I think it it's hard to do, to be perfectly honest. And you know, it's easy to say that and sort of say you should do this. It's really difficult to do in practice. It's it's not necessarily...

Jess: Easy.

Nicole: Comfortable! It really isn't, because you are asking, you need to involve the people who are doing the process first and foremost because they're the ones who are gonna know it best. And sometimes there's a lot of like, there's a lot of legacy knowledge, there's a lot of "no, we've always done it this way". There are like personal aspects of it that you need to get past, but that work is worth doing if what you want at the end of the day is a real transformation. If you just want, you know, we've got procure to pay process that we don't really like and we wanna optimize it. You can take that approach, but you need to be realistic about what's gonna come out of that versus something at a broader scale.

Jess: But I think you need the people delivering the process today, but you also need enough senior representation to answer the why.

Nicole: Absolutely. 

Jess: I definitely have been in the room watching processes get ripped up and realize that the operators who are very protective of that process, the question that they're trying to answer, actually the people that they're trying to answer for do not give a .....[laughs]  you know, like this. So, you know, I can't remember where it's from, but there's a really good business book that talks about that there was this particularly dense report they used to do on a monthly basis and someone decided "I'm just gonna stop sending that report to execs and see if anybody notices". And nobody noticed for like three or four months. And they're like, okay, so we don't need to do that anymore. Because you do have these particularly in more mature, like definitely enterprise landscapes, you know, where things are getting big enough and complex enough where people lose touch with, you know, like what who am I delivering this for and why? And what do they actually care about?

And it's back to business fundamentals of actually talking to each other and actually delivering efficient structures and all the rest of it.

Nicole: We went into a a huge investment company in the US to to look at their document processing and we talked to the business line owner about okay, how do you, you know, sort out documents, how do you prioritize everything? He gives us very clear rule definitions, very well documented. This is how it's done. We train all our people, our team is very tenured. This is the standard, this is what you should find. We go, great. We go out onto the floor.

We shadow the teams that are doing the work. We go, we shadow one person. "Yeah, okay, yep, that looks like the process". You've explained it in different words, but you're basically following the same rules. We go to another person and we find that they're doing it just slightly differently, but it's enough that we can't necessarily say this is the hard and fast way it should be done. And then it becomes a conversation of why did you sort of adjust the way that you're doing it? because the information that we get from our clients actually doesn't match what leadership thinks we're getting from our clients. And it's not anybody's fault. It's not a misstep, but it's something that when you do that sort of process exploration, you really start to surface and it becomes a okay, how do we make sure that what we're taking forward is gonna hit the right goal.

Jess: Mm-hmm.

Nicole: Be easy if everything was very, very clear.

Miguel: Now that you were talking about that, I just thought about the previous steps that companies should take in order to do the implementation of artificial intelligence...what do you think about how relevant it is to have a proper data foundations and governance, for example? I know that this is a tricky question, but...

Jess: [laughs] I know Nicole's gonna grab the battery about two seconds.[laughs] But it is obviously fundamental. It's you know, it's like it's a you're talking about technology that's sitting on top of a data landscape. I have seen some quite creative methods of like rates of clean up or not clean up, or like what can you get away with not scrubbing, and businesses walking forward without scrubbing and then learning that, you know, garbage in, garbage out, and there's no value in that.

I just wanted before I pass the mic, I talked to a client last week who's telling me that this is a specific project in their business and what they had been working on was translating data into essentially a semantic vocabulary which was agreed in the AI structure, so that the terminology that's being used, this is more about, you know, if you think what is this data? So we have a you know a data point that sits here that we call this, like what does that actually represent? So not just the data quality of is it accurate, but do we understand it? And are we using it consistently? And do we have consistent understanding? and where does it travel through our systems, et cetera? So, you know, I know all that data analysts will be either exhausted on the floor or rubbing their hands with the opportunity to get into this, but anyone I think who works in the landscape of data, like you know, as we all know, playing with the LLMs that, you know, whatever you put in, you yeah, you've got high risk on what you get back out. So yeah.

Nicole: Yeah. I mean, to your earlier question, good process design creates good data at the end of the day. I think that if you're redesigning things from a process standpoint, having somebody in the room who understands the impact of the the data that that's gonna produce really helps to make sure that you have something that you can use at the end of the day. I have yet to ever speak with a company that says we're really, really good with the way that we manage all of our data.

It's just it's not, you know, everybody feels like "oh, I should have my data in a better place before I start". You know, the best time to plant a tree was five years ago, the second best time is right now. Like you you can still start. It's not an insurmountable task. It is really difficult work because exactly as you said, you need people who understand what that data represents and what its significance is. I think some of the AI tooling that's available can help with that cleanup process. So it depends what you want to do with it.

AI, if you just put a rag model on top of a set of data, usually you'll find if it's a significant amount, you've got duplication, you've got conflicting information, you've got things that it will surface so that points you in the direction of where you need to focus your attention. What we've also seen is it kind of unlocks data that previously would have been too difficult for us to consider.

So for example, you have a lot of unstructured data out there. It tends to be one of the biggest problems for organizations is to be able to put some sort of sense around that unstructured information and get value out of it. We've seen as a very simple use case, AI be really helpful for getting better information out of ticketing systems. So that if you have, you know, a ticketing queue that's for shipping and customs questions,and people put in free text and they have a lot of different questions. Maybe they're disputing a customs invoice, maybe a shipment got lost, maybe they're trying to ask if they can ship certain things in certain regions. We can now feed that into AI models and get out reporting that says actually, you've got a lot of confusion about hazardous goods shipping. Here's how we can structure a training program, here's the groups that have those questions.

And so, it also it unlocks the potential of getting more information out of things that are happening in the organization. But it is exactly as you said, garbage in, garbage out. You need the attention and the governance around the information that you have to make sure that it's not spitting out bad answers, hallucinating, you know, coming up with numbers that don't make sense.

Jess: Yeah, I think the potential on the data side is so exciting. Like as you say, I've seen AI use cases of auditing landscapes where you might have, you know, even in internal communications, you've got different versions of guidance out in different places. Or, you know, you'll have your internal you know, internal communication portals are saying one thing and then you've got policies out over here and another country in the business it's saying over here and and, you know, that used to be a job of someone physically having to go around and find all this stuff and da da da.... and now it's a case where if you can connect the systems and you can set an agent to work to go and play with that and put it into a synchronized setting and say, well what's the impact of this, that, and the other, then you've got the ability to, you know, increase your hygiene like at speed.

So I think it comes back to your skills in your business of like who do you have in your business that has enough a awareness and understanding both of what your like legacy data landscape looks like and why, 'cause that story is often quite important as you say of like why are we doing this? Like there is a reason. [laughs] Or is there not a reason? You know, sometimes it's inherited and we can let it go now.

But also the what is that potential? Like what where do we want to go to? Like where what should we be aiming for in our data landscape? Because I think that those skills, organizationally are just becoming increasingly significant.

Like you know, what we're talking about is ultimately like the increasing digitization of our businesses. And I think what AI is promising is that it's digitizing industries that have been less touched by digitization in the past. And that's where we're getting more and more people, you know, into this narrative that before was much more in the knowledge based industries where, yeah, that's their bread and butter and you've got much more digitally native people.

But in my world, working also with industry spaces that where this is not the norm. And so you're having to do a big upskill, not only on AI, but actually just digitization and like why does data matter? You know, so it's quite fundamental stuff. And I think assessing the landscape of what's our starting point is really important. So you're not being super ambitious or not setting yourself up for fail.

Nicole: Yeah, it's also a really good point of the digitization aspect of it. There are a lot of instances where you use people as databases essentially. There's a lot of knowledge in people's heads. And what we also find a lot of the time is if there isn't a system and there isn't a method for getting that information into the system, then you you have to solve the system problem first before you can really realize the analysis that you want to do on top of it. And so sometimes it's "hey, we really want to be able to do this". We want to understand this about our workplace or our adoption rates or how people are engaging with our workspace and our services and things like that. But it points back to the where would you get that information from and what sorts of things could we pull together? Okay, we need to we need to create a home for this information so that we can do that type of of analysis.

Miguel: I'm afraid that today we don't have more time. There are a lot of topics that we can introduce, but that was the last questions for today's episode. Thanks a lot for being here today in this podcast. And I also want to thank you, all of you listening at home or at your offices. Remember that you can listen to us in Spotify, in YouTube, and also on Amazon, in Apple. Thank you. See you next time.