[Music] [Oliver Dorr] The theme of Adobe MAX this year is Creativity Unites Us All. I couldn't agree more. When people come together, that's when true creativity happens.
And there's a lot of discussion about AI right now. Some of it's exciting. Some of it's kind of anxiety-inducing. These are very natural human emotions. And like many of you, I've been reflecting on what it means for our work, our roles, our companies, and the future of our industry. And, ultimately, I believe that the goal of AI isn't to diminish human potential but to ignite it.
My name is Oliver Dorr, and I have about 20 years of experience at the intersection of design and technology. Not to date myself, but when I started my career, smartphones were not a thing. I was a flash developer, and I laid out websites using tables.
I've seen firsthand how the right blend of skills and perspectives can produce remarkable outcomes. I'm also a partner for technology at Work & Co. We were founded in 2013 to design, develop, and launch digital products. I've been at Work & Co since the beginning of the company, working at our headquarters in Brooklyn until just last year. This was actually my desk overlooking the Manhattan Bridge. People often ask me, "What does the name Work & Co mean?" And it represents both the work that we do and the company that we keep, our teams and our clients, because to us, both are equally important. We're also very proud to be a part of Accenture Song, the world's largest tech power creative group, who we found a team of talented and likeminded people who share our belief in the power of digital products.
Over the years, we've tackled a variety of projects that push the boundaries of what digital products can do across CPG and retail, even city transformation. You've probably seen this a lot this week, but just this week, we helped to launch a new Gen AI powered bottle personalization tool for Gatorade using Adobe Firefly, a breakthrough experience that was highlighted in yesterday's breakout session with leaders from Gatorade and Pepsi.
In the past, we've also worked with Apple on new digital experiences for their retail stores. We brought ecommerce to IKEA's mobile app for the first time. And for the New York MTA Subway, we created a digital map that finally settled the decades old debate between the Hertz and Vignelli designs, delivering a unified real time interactive experience.
What I'm here to talk about today is the architecture of working together, why it's crucial, how it's evolving, and what might still stand in our way. I want to share a new system designed to elevate how we build digital products by putting people at the center of it. And I have some predictions on the future of digital product creation and how we can prepare our teams to navigate it with hope and optimism. And we'll also have some time set aside at the end, if you have any thoughts, experiences, or questions that you'd like to share as well.
Collaboration. The action of working with someone to produce or create something. Now if you're like me, you're probably thinking, if only it were that simple.
There's a lot that goes into successful product development. I've highlighted just a few of the key elements here. But the truth is, these principles don't stand alone. They're all interconnected and rely on one thing to succeed.
Yes. Collaboration. It's what brings these pieces together. It ensures that teams are aligned and focused and able to execute at a high level. And without it, even the best intentions and strategies are likely to fall apart.
This is what collaboration looks like for me. It's a global team of over 400 designers, engineers, product managers, and strategists across three continents. And it goes further than that. We have a talent team who help to hire and retain the best. We have resource managers who navigate the complexity of assembling these teams. And we have our operations team who create an office environment where creativity can thrive. In our industry, when you focus on delivering solutions for your clients, it's sometimes easy to lose sight of the fact that we're actually in the business of people.
In our day to day work, you may not always feel the impact of scale like this, but think about your own company and your own team. We're all part of a team. Whether it's a dedicated project group or a discipline, these structural frameworks are needed, but they can also bring challenges that we need to navigate.
When I began my career in the creative industry, teams were often structured like this, standalone and siloed with waterfall processes dictating the flow of work. It felt and was like ships passing in the night. You had multiple teams working on the same project, but you didn't know who was involved, you didn't know why decisions were made, and you couldn't put a name to a face. Handovers consisted of documents and design files and wireframes, leaving each team to wonder, "How did we get here and what am I supposed to do with this?" I suspect maybe this still might resonate with some of you today.
Where one role stops and another begins, that's how organizations are structured, not how new ideas are realized. This is what creativity actually looks like. It's about overlapping skills and shared ideas. True collaboration happens when those boundaries fade away and teams build off each other's strengths. This is what it looks like when it's done right, breaking down silos to enable cross disciplinary innovation.
Now despite moving beyond the challenges of the waterfall model from a team perspective over the years, our tools and data still mirror many of its constraints today. While there's some overlap of these tools across teams, the volume of data created during a project and the range of tools used by teams often result in a fragmented system that forces teams back into silos, the very thing we've been trying to overcome.
And the cost is clear. We're duplicating effort. We have gaps in our shared understanding. There are multiple formats for the same deliverables, and missed opportunities to bring together diverse skills and perspectives across teams.
Steve Jobs once said that the computer is a bicycle for the mind, a tool to help us go further and faster. And I believe that GenAI can build on that dream, a tool to allow individuals and teams to spend more time on what they really mattered, collaboration and innovation.
Now the pace of change and transformation that we're all experiencing right now, I think, requires a new operating system.
After years of diligent research and building on decades of advancements in machine learning, 2022 brought GenAI into the mainstream with a bang. ChatGPT, image generation models, cogeneration tools, all arriving in rapid succession, asking questions in and of every industry. It was really head spinning, reshaping how we think about technology's role in creativity and problem solving.
In 2022, the conversation around GenAI was centered on a single question, what is it? It was a time of exploration and curiosity as we tried to make sense of these advancements. And the potential of AI seemed vast. The applications were still unknown to us, leaving industries and teams asking questions about how it worked and what it could ultimately do.
From something that wasn't even on our radar two years ago, to being part of nearly every conversation and every experiment we're performing today, how often does such a rapid shift happen? It's pretty shocking when you think about it. And as a result of that strategy and experimentation that we're doing at breakneck speed, the question has evolved from what is it to what value can it drive. The excitement now shifting towards real world application with businesses and teams focused on how these innovations can deliver real results.
We've witnessed an explosion of use cases with GenAI over the past 24 months, from multimodal content generation to real time video or AI driven prototyping, the potential is seemingly endless. But achieving these outcomes well, I think, requires a new approach to problem solving, one that acknowledges the unique modalities and implications of this technology.
Despite the popularity of agile methodologies, there is still a great emphasis on design as a discipline leading the way on ideas to be implemented by technology.
I've always believed that problem solving should feel more like a dance between capabilities rather than a handoff, especially when navigating a paradigm shift like GenAI.
Technology informs design and in turn inspires technology. When there are no existing patterns, approaching the opportunity from both sides at the same time sparks learning, opens new conversations, and helps us establish new ways to express the potential of this technology.
At Work & Co, we knew GenAI would require a similar approach. And we responded by learning as a single multidisciplinary team, bringing together technology, design, product management, and strategy.
I've observed that some organizations struggle to move beyond the proof of concept stage because a POC often fails to capture the true complexity of a real world scenario. That's why we choose to prioritize real projects where the complexity emerges, whether it's an internal tool or a project for your client, in actual implementations, the true challenges emerge, allowing you to learn and find solutions that work in practice rather than just in theory.
It's crucial for teams to be really honest with each other about the current and near term capabilities of this technology before moving forward. There's a lot of hype right now. If you don't get the basics right, your idea will never see light of day. And if it does, it likely won't be something that you're very proud of.
Finally, we believe ethics should be at the heart of everything we do. Our work should be a force for good. And we expect our teams to speak up and raise the alarm whenever they see that value question at any stage.
So these are the rules of engagement for our team, but let's see how it looks in practice.
We started small, very small, but meaningfully. Our goal wasn't to introduce yet more tools to our team, but actually to integrate this new capability natively into the platforms that they're already using every day. So we developed a plug in that would allow our designers to use multiple image models to generate prompts and place images directly within their projects at speed. It was a great start because it was a focused engagement with clear time constraints, an ideal starting point to focus on our team, what was truly valuable versus the art of the possible. And I think it was also a great way to get our team comfortable with the idea of using this type of technology in their daily workflow.
One of the first client engagements incorporating GenAI for us was with FreshDirect. They are a leading online grocery delivery service. And we designed a recipe recommendation tool. As users check out, the tool suggests recipes based on the items in their cart, identifies any missing ingredients, and allows them to add them directly to their basket.
What I appreciate about this example is how the technology is essentially invisible to the user. No one's using this feature saying, "I think it's powered by GenAI." It just feels like part of the shopping experience. And for our designers, it meant integrating this new functionality into an existing design system. For our engineers, it involved powering the experience that provides reliable, structured output from the large language model into a checkout process to simplify it rather than complicate it.
Next, our team set out to integrate a natural language interface without sacrificing the effectiveness of modern user interface. At the time, most language models were sort of a text in, text out approach. You wrote something, you got something back in text. But we believe that the future will be a hybrid solution combining unstructured text, voice or vision based interactions with structured UI elements. We called this conversational commerce. Now with no established patterns, it took many iterations with our design and technology teams sitting side by side, learning every single day. And rather than collaborating and design files, we actually worked in code based prototypes and iterated there. And we learned a lot during this project.
Prompt engineering, essentially instructions for how a language model will act, is more of an art than it is a science. Managing context is difficult during a long conversation. Is something that you mentioned 15 questions or statements years ago still relevant? And if so, how does a language model understand that? There were no best practices for combining data across different sources for a language model to understand and aggregate. And sometimes the large language model will simply fail because this technology was still in its infancy. I'm mentioning all of these challenges because I think sometimes when you're on LinkedIn or YouTube and you've seen these videos about people building applications in a matter of hours and the death of design and development, I think my answer to that is, one, it's not true, and two, I think it's a bit disingenuous. To get this right takes a lot of work. To build something that just works is just the beginning. The real challenge lies in making it production ready, where quality should be the top priority.
We've also worked with travel and hospitality clients for years. So when we set out to create an AI powered trip planner, the potential to us was clear. The idea, however, was pretty ambitious. You may have done this with your friends and family where you want to organize a trip together and one of you might be the orchestrator for taking all of that information and trying to plan something. We tried to take that idea and pass it off to an AI system. Looking back, it's a bit crazy.
Agentic AI or AI agents, you might be hearing that term a lot recently.
It's a popular term. It refers to systems that can reason through tasks and make autonomous decisions with minimal human input. And as we dove into these new emerging agent frameworks, we quickly realized something. There are no design patterns to represent memory and reasoning, planning, or the use of tools with these AI models. And that raised an important question for us. Do any of those complexities even need to be represented? Or should they just work invisibly behind the scenes? This turned out to be an exciting thought experiment for us. And while these frameworks at the time showed real promise, the technology wasn't ready for the level of real time autonomous decision making that we had in mind.
Again, we found those boundaries. Now 18 months later, the landscape is already shifting. And you're going to be hearing a lot more about this topic in the weeks and months to come.
You'll recall, I mentioned the challenges of silo teams working in different tools. And at Work & Co, we've always believed that if we're going to be disrupted, we should disrupt ourselves first. That's why we created CodeSail. CodeSail is a platform that takes product requirements and generates server side code in multiple programming languages to power modern applications. It can read third-party documentation and generate new APIs from them. It can also test and fix its own code, and it makes the development faster and more reliable. Now I lead a technology group at Work & Co of more than 100, and so I'm very aware of the implications of building tools like this. It was very important to make sure our team understood and embraced the opportunity in front of us. The goal is not to replace people. There's a lot of rote, mechanical, typical sorts of work, glue code that teams need to write that doesn't excite them every single day. Our goal with a platform like this is to be able to offload that type of work so they can focus on more unique and complex challenges. And for us, CodeSail has been one of the most rewarding projects for our team over the past year, not just to think about the future of product development but our attempt to shape it as well.
And finally, as I mentioned earlier, we had the opportunity to partner with Gatorade to bring an exciting new feature to life, a GenAI powered design tool for their iconic bottles. With this tool, athletes can personalize their bottles with nearly endless possibilities, drawing inspiration from their favorite sports, hobbies, and unique styles. To build this, we integrated Adobe Firefly, the image model, directly into this tool, making it simple for users to search for ideas and choose from their own preset options and create something truly their own, offering athletes a fun and approachable way to express their creativity while staying true to the Gatorade brand.
It was extremely exciting to be involved in a project like this, to apply GenAI to something tangible, something you can hold in your hands. From building a UI that leverages Firefly and curating themes and categories for the universe of imagery, to developing a pipeline to generate tens of thousands of these images that are unique and our goal to be high quality as well. And it presented a new challenge literally every single day. That's a challenge that our team thrived on. And most rewarding, it puts the power of GenAI directly in the users' hands. Users who may not have interacted with this type of technology until today in a very simple and intuitive way.
If you'd like to check out this online, it went live this week at gatorade.com. I'd recommend it.
We're still in the early stages of GenAI. This is the first innings with this technology. And we've made progress, but we've also made mistakes as well. Both have taught us a lot. First, discipline specific work limits potential. The best outcomes happen when teams work together, when perspectives collide and new ideas emerge.
Experimentation is critical because there's just so much that we don't know.
Innovation is moving faster than we can keep pace with, so we need to embrace a trial and error approach to continue to learn and keep up.
A great way to do this is with prototyping, which is more than just a tool or an output. It's a forcing function for collaboration. It's what our teams do every day. It brings our teams together, side by side, to test, discard, and refine ideas, and it accelerates learning and drives towards better solutions.
I believe that learning should always lead to sharing. Every team that touches this topic is going to gain valuable insights, create opportunities for teams to connect with each other within your broader organizations. And be candid. Push for the unvarnished truth about the challenges that they're facing. Find those boundaries. Honest exchanges will drive growth and strengthen your entire team.
And finally, quality is non-negotiable. My colleague Nick Law at Accenture Song recently said that mediocrity is now free with GenAI. Speed doesn't mean lowering standards. For those of you using these tools today, iteration has likely become your best friend. Iteration means discarding a hell of a lot of ideas before landing on that gem that you had in mind. Every team member must take responsibility for keeping the quality high, even when working at speed.
And I'd like to leave you with what I believe is coming up for us in the future, and ask a simple request for everyone here as well.
In 1962, Arthur C. Clarke, a technology futurist and fellow Brit, in his book Profiles of the Future, said that "Any sufficiently advanced technology is indistinguishable from magic." GenAI feels as close to that as anything I've seen in my lifetime. And I believe that our opportunity is to apply it without losing that magic.
The way we interact with technology is evolving. It's becoming more intuitive and personalized. Instead of rigid, one size fits all solutions, we're moving towards experiences that adapt to us and our needs. Accenture Song calls this fluid experiences, putting flexibility and user control at the heart of every interaction.
We're already seeing examples of this in our work today. In the top left, 3D environments with digital twins of Defender Vehicles, Apple Vision Pro Apps for the PGA Tour, and Fortune Analytics, an AI powered tool that sifts through years of business knowledge, gathering insights from trusted sources and presenting them in an easy to understand format. These aren't just future concepts. They're real and they're alive today.
I expect and predict that we will see projects that use GenAI to solve some of our current challenges in smarter and more integrated ways. Let's look at a few of those possibilities.
First, think about every project artifact, whether it's meeting notes, your client notes, design files, architecture diagrams, becoming globally available and accessible to both teams and systems. By breaking down the walls between these disciplines, AI can empower us to make smarter decisions and respond more efficiently whilst also catching inconsistencies before they become bigger issues. It can foster a connected environment that supports every discipline across your entire organization.
In the future, I believe that we will move from different tools per discipline to unified tools that empower teams to architect, design, and prototype, and deploy together. This integration will mean fewer interfaces and fewer handoffs, leading to greater collaboration and faster iteration.
And the way our teams look are going to start to change as well. While it might seem like science fiction today, we're approaching a future where collaboration with AI will facilitate our work in areas like production design and code reviews, automated testing, presentation decks, thank God, and more. We will interact with these tools to the same communication channels that you do with your teams today, including chat and video and voice. With the goal of offloading tasks so our teams can focus their energy on topics where human creativity really flourishes.
As we navigate this paradigm shift, please remember that your perspective, voice, role, and work shape the teams you work in and the companies you work for. Every project you launch sets a standard.
So I encourage you to use that influence. Use your creativity to unite us all and make your work indistinguishable from magic.
Thank you so much.
[Music]
