Jen and I are really excited to dig in, exploring how, organizations like Sephora are navigating this this huge transformation. We do not have it figured out.We haven't figured out how to crack the code. So we're we're still, probably like many of you, figuring it out and trying and very much in the in the middle of things.

That's right. Thanks so much for having us. And I agree with Melissa. We're in what you call, the messy middle, and we're doing the things we encourage our our teams to do, upskilling, installing tools, prototyping, experimenting, getting our hands dirty like all of you. And, honestly, that hands on learning makes this moment not just challenging, but fun. And we gotta have a sense of humor while doing it too.

It has been really fun to just play around with these with these tools. And, at this as a senior director, I don't I haven't had much opportunity to, like, actually be building things. And so in the past year or so, that has been actually a really, like, inspiring and and, fun and challenging, place to be. That's right. So, usually, when when companies are talking about, enabling AI in in orgs and tech orgs, it's always about, getting more productive and being more efficient and, doing more and being faster. At Sephora, we're thinking about that a little bit differently. Jen, can you share a little bit about what we're thinking?

Yes. So you're right. We often hear a lot about which tools and what are the specific efficiency gains, what targets can we hit, whether it's that one tool that processes POs faster or whether it's a tool that provides the best meeting recaps. But remember that tools don't actually compound. So at Sephora, we're focusing less about what tool we're using, what agent we're using, and more about, how we think about the rework because rework is no longer considered a failure. It's actually more about learning, and the cost of finding things is radically lower with AI and with AI AI coding agents. So that mental model flips, and that speed of iteration is one of the strongest modes of the AI agent. We should capitalize on that.

So, Melissa, what did you personally have to unlearn as that shift starts to become real?

I feel like I have had to rewire my whole conditioning on how we build software. So especially in in product management. For, you know, twenty years, I've been optimizing for, like, how do we get all of the information that we need in order for the build to be efficient, protect the developer time, you know, minimize the bugs, you know, make sure that, avoid scope creep. So with AI, this has totally changed. I feel like we are trying to learn, but it's okay to just, like, build something and then discard it. So kind of letting go of that fear, of throwaway work, which is a little uncomfortable. That's one thing. I think the, we spend a lot of time traditionally, we spend a lot of time on, drafting artifacts. So having, like, a really thorough spec document or PRD, spending a lot of times transferring that into stories, and then, of course, the the all of the effort that goes into actually coding something.

And now AI makes all of that a lot easier, a lot faster. So it's really about shifting from, like, producing all of those artifacts to the fun stuff, which is, you know, what looks good, what what's a good interaction for users, and using your best judgment, being really clear about what what the outcome is. Right. So, yeah, I think that, that product in particular, product people are really uniquely positioned because we have been communicating what to build and why for our entire careers. And, we we still need to do that, and it's just about getting really crisp about about that clarity.

That's right. And I think one of the things too is, in the traditional model, if you miss a requirement, there is that debt that we incur because we have to then go back Mhmm. To build. And so I I do think that with AI, we're we're able to jump back into that iteration loop, faster. So when we think about AI and this model, like, how is the structure and the skill mix of product and engineering teams changing?

I've been thinking about this. We have been thinking about this a lot Yeah. Lately, because we, we've been structured in a way as I think most tech orgs have been for for many years where you have roughly you have one product manager and maybe eight to 10 engineers. And as we're introducing AI coding agents, the that ratio shifts. So now we're, like, one product manager maybe would work with maybe one or two engineers who are more like architects. They're really steering the AI agents. So one thing that I I'm really proud of Sephora for is that we're we're creating opportunities for people across the org to kind of reskill and focus on a bit more of that, like, upfront problem framing, communicating the clarity and communication that we've been talking about.

And this isn't, you know, totally flushed out at Sephora, but I think, having that vision and, trying it out in in small in small pods has been, really effective at kind of showing that these this sort of different way of working.

Yeah.

You know, not needing necessarily, like, perfect requirements to start building and, like you were mentioning, like, evolving and learning as you go, I think has been really, a really interesting shift.

Right. It reduces some pressure too in product.

It does. It really does. And I I think that showing someone being able to show someone, like, a stakeholder a prototype, like a quick prototype or, like, this is what we've built so far is so much easier for the stakeholder than than to, like, read a bunch of words on a document and try to envision what is gonna be built.

Right. So what what kind of people do you think, do you see really striving in this new model, Jen?

Right. I mean and we talked about being in this messy middle. So something we're noticing across these AI native teams that or the AI native teams that we're we're striving to be is that, the talent profile is also shifting, and it's shifting in a in in some really interesting ways. Like, first, as you said, product, managers are becoming even more central to the process, not less. But the focus is less about managing backlogs and coordinating team and really more about being clear and having that clear vision of the outcome. What are we doing and why is it important? And setting that clear intent and, defining the constraints that guide both humans and now agents. And so execution gets much easier, but then the clear thinking and decision making become the bottleneck.

And so product managers and product leaders who thrive are very curious, extremely adaptable, and comfortable navigating through this ambiguity and still can make decisions, not waiting for everything to be buttoned up before, being able to move forward. And we're also seeing a shift on the engineering side too. Engineers are moving from pure builders into architects and orchestrators. And because the barrier to the execution has decreased, you know, AI can handle huge chunks of the workflow now. Engineers, who thrive aren't actually the ones who want to control every line of code. They're the ones who are willing to let go of that control and focus instead on designing systems and managing outputs. And this means that across the board, this opens the door to a much broader set of people than the traditional model ever did.

Strong system thinkers, people who are energized by iteration, not just delivery, and people who naturally move across product and engineering and design. Because in this model, the unit of work isn't the function, it's it's the outcome. And so at Sephora, the talent that, we're looking to enable and upskill and help make successful, all share the same through line. And so they're curious and adaptable and comfortable, comfortable with being uncomfortable. And that's what we're that's what we need.

And I think, we've been also talking a lot about resourcefulness. Mhmm. That's right. Just try like, how how do I do something? How do I figure out how to do something is, like, the new question because it's not like maybe the person next to you knows either. So

That's right. That's right. It's almost like, not to sound so doom and gloom, but it's almost like survivalist. We all need to find ways, to to move forward. Yeah. So with all this opportunity, right, this is the this is the the utopia view, but, there are some real risks too. So how do you like to keep teams grounded while moving quickly? And what do you think, Melissa, makes women's voices matter so much in this moment?

Yeah. This is, this is a great question. I'm glad we're talking about it because, there are a lot of risks I see for in AI. I mean, some of the, like, most practical things, that we're really, you know, hammering into our our teams and reminding everyone is just always verify the output. At at the end of the day, AI is a, probability equation, and it can sound really confident and give you totally incorrect information. This just happened recently where we were my team has been creating all both of our teams have been creating, these business cases as we plan for 2027. And, we created an AI skill where you can plug in a bunch of info, and then it'll spit out this the document. And, it was the the model gave these, like, this data, like, about, you know, something about, like, oh, every 50% of search queries are 10 plus words or more.

And when I asked Claude, like, what's the source of that data, its reply was that it, that was just proxy and just fill you're supposed to fill it in with your own data. So that's always a good reminder to not, to not, you know, just trust the output. Yeah. We don't want any work slap. That's right. And then there's, you know, there's a lot of there's probably a whole another talk, conversation about security, data privacy risks, so being really careful about your data and your company's data. But zooming out a bit and getting to, like, how do I help support my team through this uncertainty, I don't really know. I I don't I don't think any one of us knows what ultimately is, like, what AI is gonna do to our society. You know, there are these, like, utopian ideas.

You know, maybe we're working three days a week, and, like, everyone has all this time to focus on their creative passion projects. And, AI is gonna do all the grunt work. And then there's definitely these dystopian, you know, views of all the money is gonna flow to just, you know, four four companies, and there's gonna be a lot of inequality. And, you know, the environment is gonna be gonna be harmed. So I think that uncertainty, and just that I both of those are possible outcomes, but they're pretty extreme. So I think we'll we'll likely land somewhere in the middle as as we normally do. But, individually, we don't have a a lot of control over, like, where we end up.

So what actually grounds me is just asking myself and asking my team, like, what is within our realm of control? And I think the couple of things are, just making sure you're consuming balanced information, so, like, understanding both perspectives and then just how we show up as leaders for our team. So I'm creating, like, a psychologically safe space for us for both of those those, like, moments where there's experimentation and curiosity and excitement, but also making room for the skepticism and anxiety. And, I think the the these organizations, like Women in Tech, and Sephora's, Women Who Tech, they matter so much to me because it does create those spaces, those of psychological safety. And I think that we're doing a lot at Sephora in women who tech to have those those conversations. Like, we just recently had a conversation around gender bias and AI.

And that was, really engaging, and that just reminds me how important it is to not just be an AI billboard and say, AI is great. Use AI. But really be balanced about what is the reality here and what might some what might be coming up for some people that maybe is a little more anxious.

That's right. And I think it's a great reminder too when we're in safe spaces like Women Who Tech, that there are different learning styles. There are different, adapt ability styles too. And so, and it also gave gives us at least that's one thing that I take away from the womanhood tech community is just they're not, the members are not always just technologists, but there are different connections across the company. And so, so tying a vulnerability plus, like, plus very tactical ways that each of us can show up and in this moment of change, have been has been very, very helpful. So I love that you bring that up.

What, what would you I guess oh, we are closing out, but any last words, Jen, on what's, like, the mindset for product leaders?

Yeah. Yeah. I love that. So I think the the one the one thing, if anything, that, I would love to offer up to this to this group is, make sure that you're adopting a learner's mindset. So So just like Melissa said, no one can no one here can say that they know the future of what it what this looks like in twelve months, eighteen months, twenty twenty four months time. But if you, have a sense of curiosity and, you set yourself with this growth mindset, then you'll be able to set yourself up for future success and as well as help guide your teams. And that open and builds transparency with peers and creates that path that we can all move towards together through change and adoption, and also explicitly tell yourself to get comfortable with being uncomfortable. And so, that way, you're being intentional about what kind of building blocks you're setting for your future self, whether it's for your next performance review or for your next role, and that'll also model the behavior for your teams as well.

What do you think, Melissa?

Yeah. The flexibility, discernment, being able to accept not being able to control everything and, like, understanding where you actually have influence and where you could make an impact, I think are really, really important right now.