Executive Summary:
Go behind the scenes of Oracle's groundbreaking
new user experience in this fascinating
conversation with Jenny Lam, Senior Vice
President of User Experience Design at Oracle.
Recording live from Oracle AI World 2025 in Las
Vegas, this episode reveals how Oracle's design
team spent years crafting the future of enterprise software - waiting for technology to finally catch up to their vision. Learn how the newly announced "Home with Ask Oracle" is transforming daily work through natural language commands, why the best user interface is sometimes no interface at all, and how Oracle's Redwood design system became the visual DNA of the entire company thanks to the backing from Larry Ellison. From her experiences building design systems at Windows and Alexa to leading Oracle's largest-ever design platform, Jenny shares invaluable insights about designing the "trust layer" between humans and AI agents, the power of telemetry vs. focus groups, and why falling in love with problems - not solutions - creates truly beautiful and functional design. Whether you're a designer, technologist, or business leader navigating the AI era, this episode offers a rare glimpse into the multidisciplinary craft shaping how millions of people will use enterprise software tomorrow. Stream it now, and be sure to LIKE, SHARE, and COMMENT!
​
Transcript:​
Mark 00:01
Greetings all. This is the enterprise edge podcast. This is Mark Vigoroso, founder and CEO. We are recording live from Oracle AI world here at the Venetian in Las Vegas. We are having a wonderful time, lots of exciting announcements and connections and conversations. And there's still another 24 hours to go. No, no, maybe, maybe 12 hours to go. But happy to be here with Jenny lamb, senior vice president user experience design at Oracle. This is something that I'm excited to talk about. It's maybe something that doesn't always get the attention and the appreciation that it deserves, and the work that goes in and the thought that goes into user interface design and user experience design when it comes to enterprise apps, especially in the AI era. So Jenny, thank you for being here. Thanks for having me. Great. All right. Well, your team made a made a small announcement today, I think, like three or four hours ago, about sort of the the main user experience when it comes to Fusion Applications, showcasing some of the new capabilities, like ask Oracle and some of the maybe design esthetic around the look and feel, the human sort of experience of using and utilizing these applications every day. You could probably talk for the entire time about this, but how about a short version of tell us more about that, what went into it and what makes you proud
Jenny 01:29
of it, right? So today, we announced home with ask Oracle. It's funny because this is a page that's very iconic, but it's actually a page that not many users spend a lot of time on. It's the one place where you're going to get all your push notifications from your agents doing an agent workflows. But it's also a place where, as an employee, if I need to find my pay slip or if I need to do a promotion for a team member, I can just use my words rather than surf the whole directory or hierarchy and just use my words in ask Oracle and get to the place I need. So that's a simple thing. You know, with natural language, we've been seeing that in technology for a while. But one thing it's it's new, is being able to do commands. So if I want to be able to get an answer right away. I can get that mask Oracle. Sometimes that answer comes in the form of text. Sometimes it comes in the form of data visualization. So depending on what kind of content is best suited as the answer is what's going to deliver and it's going to give it to you in the shortest path,
Mark 02:36
got it. And how long has this been in the works? If you can pinpoint a start and an end.
Jenny 02:42
Well, it's funny because I posted this announcement just a few hours ago, and I remember one of our technologists, you brought us this design a few years ago. Technology has finally caught up, so we worked on the design a few years ago. We've had it around. We've been waiting for the technology to catch up, and it's finally here. It says, very exciting, and the team is very relieved. It's here.
Mark 03:09
Awesome, awesome. So speaking of design, we were just chatting before you. You spent some time, good bit of time, at Microsoft, like 2020, years ago, 15 years ago. I'm not gonna date you, but I'm curious. You know, if you could transport yourself back then and compare like some of the core design principles, are there any that are just tried and true and remain true today, and others that are just so far defunct that you don't even think about them anymore?
Jenny 03:37
Yeah, I could. I could go on hours, but yeah, I got my start actually at Microsoft in technology. Prior to that, I thought it was going to go work as some design firm doing branding or something like that, but it was a really wonderful merge from my path. But anyway, when I this is my third time working on a design system for an enterprise. So I did one at Microsoft for Windows operating system. Oh, wow. And then I had a short stint at Amazon and did one for the Alexa devices using voice experiences. And this is my third one. And here at Oracle, this is the largest design system I've ever worked on. Going back to your question about, what are the things that are sort of evergreen. That's really about the designer's role of really understanding what the user wants to do, falling in love with the problem before jumping to solutions. So that's, I think that's always going to be there as part of the design discipline, when it comes to things that are like fads, they're probably the obvious ones, you know, where I think we overthink the mechanics of what users want to do and lead them down a click path when the UI doesn't even need to exist at all, as we're finding out in this age of AI, right? A lot of things can be automated. So sometimes the best user experiences, and the best UI is no UI. And you see that a lot in our Redwood design system. It's, I think a lot of people within Oracle still think that Redwood is a GUI graphical user interface, but it's a user experience platform, so and that includes when we don't want to show the technology when we don't want to show user interface and just automate it, and just do it for do it for them, because we know they want to do X, Y, Z, and we just do it for them. John, it does that make
Mark 05:30
sense? Fascinating. No, I love it. I remember Redwood when I was with Oracle, and it was still it was relatively new, but it's become quite pervasive in terms of just the brand of Oracle, right? I mean,
Jenny 05:42
yeah, that was the intention. Actually, it's funny, because Larry was the sponsor of redwood, and we originally had it as a redesign of the SaaS applications, and he's the one who wanted to make it for the entire company. So you'll see fingerprints of that user experience, whether it's a NetSuite OCI, the industry applications, even in the event experience, yeah, of course. So it's not just relegated to product. It's on the
Mark 06:11
jumbotrons. It's on everything. It's everywhere it's been out. It's phenomenal. Okay, cool. So you sort of jumped in, unknowingly, into the next question, which was in this world of AI, where you have a dual mode of, sort of human and machine, right? And you have human driven workflows, machine driven workflows, and then you have hybrid workflows, right? And I'm curious, what does that do in the brain of a designer when you're trying to think about harmonizing all of that?
Jenny 06:50
This is a hot topic in our discipline, in our nerdy little design world, yeah, but it is. It is really it gets back to what are the things getting to know what the user really doesn't care about, and that they bring some a lot of pain. Those are the things that we try to figure out first, and those are the things we try to automate. Then there are things like that. I don't think we have a healthy enough discussion about which is, what are the things only humans should do? And that's that brings them joy, you know, that we should remain sacred, you know, and that. And I think you start to see some of those more in the compliance realm, you know, like, Oh, I've got to sign off on these financial papers. You know, there has to be a human actually reviewing this, the human in the loop, but, yeah, in the mind of the designers. We're still figuring it out, but it's really going back to what is, what is the user going to do? And then how do the how do we want them to feel? Do they do we want them to not care about it? Then we can put those aside and automate or do we want them to lean in and really be engaged and CO create with the AI,
07:59
yeah, you know, it's so when in this world, which is just
Mark 08:04
now being birthed, right, where you have autonomous workflows, human workflows, human in the loop, semi autonomous workflows. And in that semi autonomous, sort of hybrid world, I've seen it described as this trust layer, right, which is the interface, or the sort of membrane between the human and the machine. When the machine there's a trigger or there's a gate that says, okay, the machine needs to check with a human now it has to be approved or validated before proceeding. That's the human in the loop. That trust layer has to be thought through in terms of, what is that? Is that? Is that an interface does? Is there like an alerting interface for the human? Is it? Is it going to be completely a deluge of agents asking for permission to proceed? Tell me a little bit. Is that a trust layer may not be the words, but does that makes sense to you, that sort of membrane
Jenny 09:01
between totally Yeah, and I in design, we have our design system supports what we call patterns. And so I think depending on which different workflow there, there is a professional that might be in front of their computer eight hours a day, five days a week, and they're, you know, processing invoices or something like that. In that case, they may, we do have a pattern that is more like monitoring, you know, and so they're kind of in this control plane, and they're monitoring things, but in maybe an HR role where they've got many different hats, you know, they're recruiting, they're also, you know, dealing With issues from the team onboarding. They may be multitasking, and that's where actually the homepage may come. The homepage may come to where that that's your hub, and you'll still notice that once we launch it, which is becomes GA early next year, is that you'll get notifications off of the O tag in the bottom right, and you'll get to see all the things that you need to prove. So those are pushed through a centralized place, and that is the fusion home. Got it. So that's as part of our design system, we have particular patterns. But you're right, the trust layer is is a necessary piece of the experience, and there's a difference between more assistive, like, I want to work with you and and notify you at the right level and not annoy you, versus the other stuff, where it can just, you know, I'm going to just check in at certain points.
Mark 10:35
That makes sense. And I'm sure that's a bit of a moving target, right? As as agents mature and proliferate. The volume, the frequency of that interaction between humans and agents, might fluctuate or change over time,
Jenny 10:49
right? Yeah. There's a notion, too, of a like an Uber agent, like a supervisor agent, take care of the other agents.
10:56
Oh, yeah, that's a good idea. Yeah, Commander in Chief,
Jenny 11:01
yeah. So there is that pattern too that we're looking at where it's you're training a supervisor, yeah, he's gonna take care of those agents, and you won't, you won't get the deluge of notification. Cool.
11:14
That's awesome. I love it all right. So one of the
Mark 11:19
things that I've heard a little bit this week last week too, at Sweet world, is this idea of making AI accessible to non technical users, right? Whether you're like an analyst, who's a financial analyst or business analyst, you're querying a database with conversation natural language, like you're describing marketers. Maybe they're building campaigns again, maybe through a conversational interface. But I'm curious if you think that there are, I guess, risk. I don't know if that's a too strong a word, but if there's a risk that you're putting a a burden, or, you know, a responsibility, upon a non technical user to understand business logic that was previously sort of delegated to other people, developers and product managers and things like that, and and is that some people call it a cognitive burden, like, are we burdening non technical users to now understand a level of logic that they've never had to understand
Jenny 12:31
before? I've never thought about that. That's a great question. Yeah, you know, the first thing that comes to mind is, no, you should never burden the user with anything. But then also, the burden really should become on how we set up their their tools, right? The configuration, the business rules and policies that are, are they built into the architecture of the tool they're using? So they don't have to think about that and that they come from an official source. So I think, I hope I'm answering, if I'm approaching the question the way you were asking it, but that, but that the architecture and stuff behind the scenes is set up in a way that won't, that won't fail them, you know, also burden them or get them to do something they didn't mean to do, you know,
Mark 13:24
yeah, and I think, I think what I'm seeing on both sides is that the on the user side people are sort of reinventing the ones that want to lean in are sort of reinventing themselves, figuring out how to use AI and understand AI in the context of a perhaps a long standing function, and on the other side, some of the development and thinking going into agents is sort of a forgiving interface, like it's not a it's like can interpret the fallibility of humanity in terms of any kind of conversation you're going to be absorbing from a human and do what you meant me to do. Maybe you didn't say it the way you should have said it, but I'm going to do what I think you meant right, and be accurate most of the time, right? And that kind of intelligence in the agent, I'm also hearing a lot of that. So it's, it's, it's more of a, you know, the spirit of what you're asking me to do, and maybe not just the letter of what you've said, right? And maybe there's a combination there, but it's interesting to see it. It
Jenny 14:31
totally is. I think that's a good thing, though, right? Yeah, to have that flexibility, oh yes, yeah,
Mark 14:36
yeah. And to have to have the necessary fail safes and safety nets, especially early on, so that if there is a misinterpretation, that it doesn't go too, too far down the
14:47
line. Yeah,
Mark 14:50
cool. This is fascinating stuff. So I have a couple questions. We have a little bit of time, so we might be able to do all the speed round questions, which is, which is fun. But before we do that. I'm curious about your method and your team. As you, as we look behind the curtains, are you guys doing, like, a lot of focus groups and user groups and research and I mean, how do you get to the right answer?
Jenny 15:14
That's an excellent question. It's a constant iteration our team. I couldn't be more proud of them. It's a multidisciplinary team, and it's unique in any than any other setup I've been in, you know, at previous jobs and previous companies, because we have engineers on our team. So I think typically designers just have the designers and researchers, maybe together, but we have designers, researchers, data scientists, Oh, wow. We have engineers, excellent engineers, who are building the platform pieces and product managers, writers. We even have sound designers, yeah. So we have a really great multi disciplinary group that work together to build things at scale and that build things that are accessible and and responsive and work in a variety of devices. Your question
Mark 16:12
just that's sort of the method to the madness, like what you know when you done focus groups, or you're standing behind two way mirrors or, I mean,
Jenny 16:20
so that is that. So back to your earlier question, like, what worked in the past? Yeah, that doesn't really work so much now. One of those is really around focus groups, yeah, I remember
Mark 16:30
doing those in the late 90s, yeah? Or click, you're watching people click on a screen.
Jenny 16:34
We don't do so much of that anymore because it's, it's, it used to be, like, designed to build, and now it's, it's prototype to productize. And so we do a lot of prototyping. It gets up faster. We get it with you, real users. We ship it to them, real users, with their data and their context and then their environment. And you get such better feedback that way. So it's kind of the startup model, you know, just shipping quickly and learning from it and iterating, but we do have a built in service that comes with the platform in Fusion called signals. I don't know if you've heard of it, and we just are launching it to customers, but it's telemetry so they can see the usage, and we use it for observability to see paths and where they're getting stuck in a loop, you know. Or if we launch something and they're not using it, why aren't they using it? Or, you know, where are they coming from, what device, what browser, what resolution, and also, we're going to get feedback on AI, and you know, how many times are they coming back to it? So that's really powerful. That kind of quant is very powerful, and we're offering that to our customers now so they can see that too. Wow. So I a lot from that. We do participate in CADS. We co design with our customers all the time. We observe factory floors, like in manufacturing, or observe like how, how their workers are dealing with a 20 foot experience of a big screen on the wall, or their man paper manuals you know, that they're dealing with. So, yeah, we do that kind of qual research in the field, but most of the really good stuff comes from the telemetry.
Mark 18:18
It's like I used to be in I used to work on IoT. That sounds like an IoT use case. That's a
18:23
manufacturing floor. Very
Mark 18:25
cool. It's awesome. Well, it's, I mean, it's like, straight from the source, right? I mean, it's like, you want to stay grounded in what's working and what isn't. I mean, someone who's invested is a customer in their own environment, wanting to think through, is this going to work for me, right? Not some sort of, like, generic conference room
Jenny 18:43
pilot right in there, and they're in your space, they kind of feel bad, exactly. So if you get a sugar coat, yeah, in there, like, oh, one place we're doing what's really well is in healthcare. Okay? So that's,
Mark 18:57
that's fascinating. That's awesome. All right, I'm gonna, okay, we're gonna do these speed rounds before we run out of we're gonna do them. Okay, all right. So last week, Evan was on stage with sweet world, and he announced dark mode. And I have to know, do you prefer light mode or dark mode?
Jenny 19:17
It depends. So if I were a developer, it's obvious my answer would be dark mode, but it depends on what task I'm doing, okay, yeah, or
19:25
if it's light or dark, out
Jenny 19:28
that too, like I don't And also, like, on a plane, I don't know, like that bright screen, like screaming at me and lighting my face. I want to hide
Mark 19:36
my face in bed trying to go to sleep. Your last couple scrolls.
19:41
Yeah, exactly.
Mark 19:42
Okay, I have to think about that. I don't have a preference. Best piece of design advice you've ever received.
Jenny 19:51
Fall in love with the problem, not the solution. Love it.
Mark 19:55
It's interesting. I have this next one now. I look at it differently now that we've had our conversation. But what is the designers kryptonite? Endless stakeholder feedback or no research budget?
20:09
No research budget?
Mark 20:15
I think Oracle has has some research budget. Well, spent. Okay, so design tool of choice, figma, sketch, or something Oracle built, or anything else,
Jenny 20:27
or figma, or figma all the way, figma all the way, yeah, got it.
Mark 20:31
This is interesting again, because you said you have engineers on your team. Favorite design crime committed by engineers when they just quickly fix the UI? Oh, that doesn't happen. I don't have any trouble believing No. I don't buy Yes, but, but there are engineers, perhaps, maybe they have access to just quickly fix the UI. But I know that. I mean, basically, it's like I said before. I think a lot of people under undervalue what has gone into the UI and the user experience, right? And they don't think about how it all fits together and accomplishes a goal.
Jenny 21:11
Yeah, there's a lot of hidden, hidden steps and processes in place and thoughtfulness that goes into any design, not just software design, you know, like the devices we use the chair we sit in. Yeah, that's great. Can I tell you my favorite quote? Yes, you can please. It's Buckminster Fuller, and I hope I get it right. He says it in the beginning, I don't worry about it being beautiful, but in the end, if it's not beautiful. I know it's not right,
Mark 21:45
so beauty is a byproduct of being right. You got it. Love it. I love it. On that inspiring note, we are out of time. Jenny, thank you so much for being with us. You can't see us, but we're smiling. We're happy to be here. Thank you all for being with us. We ask you, as always, to repost, like, share, comment, let us know your thoughts in the enterprise edge community. We'll be back, I think, two more times before we call it. Call it quits this time around. Thank you all for joining and we'll catch you next time on the Enterprise edge podcast, Thanks all. Bye.
