[00:00:00] Karla: If you look at AI as simply, uh, a way to increase your productivity or reduce your staff, you will be left behind chat, GPT and generative AI reached 50% adoption in six months, and it's only getting faster.
[00:00:15] David: Carla, welcome to the show.
[00:00:17] Karla: The future means not doing boring work. 30% of your job you did not jump outta bed to do.
'cause that's the reality of working in corporate. Uh, whether it's processes or the have to dos, it's grudge work that fills us all up. If your teams aren't spending 30% of their time. What I call cognitively numbing work. Suddenly you can focus on the things that actually give you joy, that, that help you rediscover the reasons why you started this vocation or this profession or this expertise in the first place.
[00:00:46] David: When I heard your story, um, I was so impressed with the vision that you, you'd shown at such an early stage
[00:00:52] Karla: what set us apart and the advice that I would give to people who wanna create agents. It is start with the human. And what I mean by [00:01:00] that is take a very human-centered design to the way you create agents.
[00:01:04] David: What should a forward-thinking CEO be asking themselves right now?
Carla, welcome to the show.
[00:01:16] Karla: Thanks. Great to be here.
[00:01:18] David: We spoke a few weeks ago. You said something to me that I. I wrote down, I underscored it. I drew a circle around it. You said something along, along the lines of AI can bring joy. Back to work. Can you just sort of reflect on that and, and talk a bit more to what you mean when you say that.
[00:01:34] Karla: So, when we talk about ai, it's fraught with a lot of associations, right? Um, you know, uh, organizations talk about productivity. Workers are afraid about job loss. Um, employees are worried that their bosses are gonna ask 'em to do 10 times more work. 'cause suddenly they have these chat bots helping them to do things.
And one of the things that I believe is that part of the power of AI is. That the future means not doing [00:02:00] boring work. And when, when we think about our day-to-day as as employees, as corporate workers, as knowledge workers, I always get a laugh when I say, and I look around the room and I say to people, you know, uh, 30% of your job, you did not jump outta bed to do.
Um, 'cause that's the reality of working in corporate. Uh, whether it's processes or the have to dos, uh, it's grudge work that fills us all up, right? And so when we think about AI and what, what it can do, taking away some of that grudge work. And if your teams aren't spending 30% of their time and what I call cognitively numbing work or grudge work, suddenly you can focus on the things that actually give you joy, that, that help you rediscover the things, the, the, the reasons why you.
Started this vocation or this profession or this expertise in the first place. And so when people are finding a little bit more of that joy in, in, you know, 90% of what they do, you're getting so much more impact from your employees and in the [00:03:00] future, that's not gonna be a nice to have. That's actually gonna be a competitive advantage.
So imagine if a hundred of your, a hundred percent of your people. Enjoyed 99% of what they do, um, because we're able to focus on the things that create value, that create impact, that really leverage their special talent or expertise. Imagine what your company would be capable of.
[00:03:21] David: Can you give us some specific examples that you've seen perhaps within your own companies of this?
[00:03:26] Karla: No, absolutely. So, so we actually use agents. I, I run a company with 200 freelancers. Um, we're a small business, but we keep about 30 people visiting one point. And in any given year we do upwards of 500 projects for large companies. And, you know, the reality of that is, uh, there's a lot of meeting notes and work back schedules and, and things that are, you know, objectively boring but necessary in, in order to keep a.
Project going at full rails. And so in 2023, uh, I actually worked with our top 25 people, the experts, the creative [00:04:00] directors, the HR leaders, the project managers, the strategists, and then I helped create digital twins of them. Um, we, we worked out what their, what their, uh, workflow was, what were the frameworks that they used that, that really were.
Fundamental to delivering their expertise. And we created digital twins. So suddenly we had a hundred digital twins, AI agents, if you would, that people could call on to help them do their jobs. So the, um, project managers, uh, suddenly had assistance that could create meeting notes, uh, within five minutes.
Instead of having to look at your notes, read your notes, synthesize your notes, figure it out, what was important, and, and suddenly you've lost. 45 minutes just to create the to-do list, right? And so across a hundred percent of our people, uh, they are all using these AI agents to take the things that they don't love doing, but are important to do that work for them.
And then more importantly, they also use the agents to help. Help them do their most important and [00:05:00] most valuable work.
[00:05:00] David: Did you encounter any resistance when you were sort of introducing these people to building their digital twins?
[00:05:06] Karla: Building the digital twins? No. Um, but there was a little bit of fear, right?
Like, is it too complicated? AI is kind of scary. Do I need to learn, uh, a whole new skillset? Do I need to learn how to prompt? But truth be told, when I started to talk to people and, and I. I went up to them, and these are people that we had known for years. We'd worked with them. We spent, uh, we had executed, um, you know, probably dozens of projects with each person individually, and they had made hundreds of thousands of dollars with our business.
And I, and, and I, I went to our experts and said, I. Can I make you a digital twin? Can I make an expert that acts and talks like you? And by the way, in addition to having access to, to your digital twin, you'll have access to 50 other people, um, who have gone on this journey. And I would say the, the overall reaction was excitement because everybody was aware chat GPT was having.
But you know, it was [00:06:00] complicated. You needed to learn how to prompt, uh, sometimes it would hallucinate. Um, so the aspect of having. A hundred helpers in your, in your pocket, uh, was actually quite exciting for people.
[00:06:09] David: When did you go through this process creating the digital twins? It was a while ago, wasn't it?
[00:06:13] Karla: Uh, yeah, so we started in 2023. Yeah. So if, if you'd recall, uh, AI came into the public domain in, in, you know, early 23. So when Chat GPT became a thing, I'm a non-technical founder, so I don't code. I am a corporate executive marketer, and I started with. Uh, doing lunch and learns for people, um, and got deeper and deeper into the space.
So, yeah. So we started, uh, creating agents in, in, uh, the summer of 2023 when nobody talked about agents back then.
[00:06:42] David: So the digital twin thing, and obviously agent AI has come forward a lot in the last six to 12 months. What advice would you give to people? How, how do you make a, a good proficient digital twin versus perhaps, are there any key mistakes you've seen people make when they try to.
Try to do that.
[00:06:58] Karla: So what I would say [00:07:00] is what, what set us apart and the advice that I would give to people who wanna create agents is, is start with the human. Um, and what I mean by that is take a very human centered design to the way you create agents. Um, because a lot of people start with engineering first.
They focus on the technology and the automations and you know, the fancy apps that will link everything together. But where we start. Started then, and where we start today is we start with the problems and we start with the people. So I will sit down with an expert and actually understand their workflow at a really granular level.
So for whatever tasks they're working on, where do you start? What are the critical steps? Where do things go wrong and where do things go? Right? Um, and then use that as the basis for training. Um, an an AI agent,
[00:07:47] David: what would you say to. Skeptics, there are still skeptics out there, perhaps less than there were before, who actually fear AI will dehumanize the workplace, not enhance it.
What do you say to those people now? [00:08:00]
[00:08:00] Karla: Uh, I would say there's a risk of that depending on your point of view. And, um, I, I think where it begins is where do you see the role of AI and how it will work with people? So leaders, um, or, or um, uh, people who focus in AI too. Make things more productive or do more, or, uh, have less people are on the wrong track.
Um, and, and that's what's gonna cause, uh, you know, dehumanizing the, the content or the assets that come out. But leaders or people who look at AI as amplifying the experts that you have or the expert that you are, it, it's about creating a platform of co-ag agency where you are using. AI as a force multiplier, uh, for the talent that humans bring.
And humans continue to have a really large role in delivering whatever that expertise is. AI amplifies the human, um, by giving them superpowers, um, and help, uh, [00:09:00] uh, that is faster and sometimes smarter, uh, than the humans themselves. But then the human. Is part of amplifying the end product by providing our talent, our expertise.
Um, and our, uh, last part, uh, our last smile in that delivery.
[00:09:17] David: And you, uh, you mentioned earlier your, your kind of background was in perhaps corporate. Corporate marketing. Tell us a little bit about what you did before you founded adjunctive. And I, I'm really, when I heard your story, um, I was so impressed with the vision that you, you'd shown at such an early stage.
So I, I'd love to hear from you how you, your career path and how you. Sort of felt so confident to take this leap.
[00:09:39] Karla: No, ab absolutely. So, so by way of background, uh, I, like I said, I'm, I'm a non-technical marketing executive. Uh, I spent 25 years in, uh, corporate marketing. I was a CMO of two organizations. I was an EVP for an agency leading the, the business for one of the largest, uh, [00:10:00] automotive companies in Canada.
And in 2016, I decided to take the leap, um, away from the, the comfortable, but. Somewhat repetitive, um, uh, arena of, of corporate world and jump into entrepreneurialism. And I created my own agency. And at that time, the, the, the, um, the vision around that agency was, uh, to create a. Better place to work for freelancers.
So it was a, a distributed agency. We had 200 freelancers. We would assemble them like Oceans 11 and provide projects, uh, for large companies. Um, and that, that agency, uh, that that service business worked very well for seven years. Um, we're quite successful. We were multimillion dollars in annual revenue and we worked with large organizations from banks to insurance firms, uh, and large consulting agencies.
Then in 2023, this little experimental chat bot came into the public domain called, uh, chat, GPT. [00:11:00] And at that time I was, I was wondering, oh gosh, um, is this the beginning of the end? Is there a best before date on the services business when this AI threatens to. Um, the things that we do. And so I went very deep into prompt engineering.
I, I joined a number of forums with machine learning specialists. Uh, I went deep into a number of different discords. The joke in my house with my, my teenage girls is that mom spends more time on discord, uh, than they do. But that allowed me to really learn. The, the craft of prompt engineering when everybody was learning it, nobody really knew how it all worked.
Um, because it was very early days and that allowed me to understand the science, um, of prompting, uh, fairly early on. And, um, because I had the luxury of 200 freelancers and expert consultants through the business that I was running, I was able to marry those two capabilities. And, um, that's when I spent time with my top 25 people understood their [00:12:00] workflows, the frameworks that they used, uh, and then used prompt engineering in order to create digital twins of my top 25 experts.
[00:12:07] David: What was it that gave you that drive to. To do that learning to kind of go into that sort of unknown place. It's not always easy as someone who's, you know, had success and been successful in a certain field to like, to take that leap. And like you said, you, you don't, you didn't have a technical background, but you were able to do all this learning.
What, what drove you to do that?
[00:12:27] Karla: That's an interesting question and, and I would say it comes down to core belief, uh, that that has always been constant for me. Uh, I've reinvented myself a few times in my career, so I've gone from an advertising agency person to becoming a client, to becoming an entrepreneur.
So the one thing that I would say is that change is the one constant that we can all depend on. Um, especially now because there's so much transformation and so much disruption that that's happening. What I found fascinating. With, with prompt engineering and [00:13:00] AI was understanding people require systems thinking, understanding technology, and AI also understands, uh, sys systems thinking.
So it's really marrying, uh, and understanding how of how workflows work, how people do things with a technology that can enable it.
[00:13:21] David: So when, when you mentioned chat, GPT came onto the market, it was a. Sort a chat bot at the time. Was there something specific it did or. Sort of said to you that made you think, wow, I really need to, to build something around this.
[00:13:34] Karla: W what really stood out for me with chat GPT was its, was its ability to intake incredible amounts of information, synthesize patterns, and then be able to come up with a, a great report and a great synthesis of the patterns in that data as good as any of our human. Leaders. And so if you could put that in a bottle, uh, [00:14:00] across a variety of different domains.
So whether that's marketing strategy or copywriting or creative, and help people with that superpower, taking in huge amounts of information, understanding it quickly, finding the patterns, and then synthesizing the answer. That's a superpower for any human who could, uh, apply that to their particular domain
[00:14:22] David: as AI has.
I guess evolved become more, more popular, more people are using it. What role do you hope to play in the development of the. Technology,
[00:14:32] Karla: I hope to be an elevator of people and, and mid-market organizations. Um, and we've been really lucky in, in creating this platform. So not only has it helped our particular agency punch above its weight, um, because like I said, uh, everyone of our team members is AI infused and AI enabled.
So there. Better, faster, smarter, and they do the work that they love, so they're that much more engaged. But we also offer this platform to our [00:15:00] clients and and other clients who come on board and we're able to provide that same level of enhancement to their teams. And this has been. Really meaningful to mid-market organizations.
'cause often with mid-market organizations, they are big enough that they need to manage a lot of people, but they don't necessarily have the capacity to do everything that they wanna do and maybe not the budget. So suddenly marketing teams, HR teams, IT teams who have access to that technology, uh, are able to punch well above their weight and take advantage of the opportunities in front of them.
[00:15:31] David: Can you think of a particular. Sort of success story where someone's really punched above their way based on the use of these agents or technology that you can share with us.
[00:15:40] Karla: No, absolutely. So, so we work with one incubator. Uh, they're a fairly large AI incubator in Canada. Um, but they're not-for-profits, so their teams are fairly small.
Uh, their marketing team was about 20 people, and because they were an AI incubator, as AI came into the fore, they were very, very high in demand. And [00:16:00] so they didn't have the capacity to answer. All of the opportunity and all the requests that were coming in from potential clients and, uh, and um, the public, uh, for their resources and enabled with ai.
We worked with the Vice President of Marketing Communications and she, uh, really embraced the capability, the issue, redrew the processes, uh, the workflows that they did, and all of a sudden that team was. Able to generate the impact, um, uh, and answer all of the, uh, all of the requirements, uh, but we're able to do it in a managed way that didn't burn out their people.
And I thought that was incredible.
[00:16:39] David: What's gonna be possible over the next 12 months? What are some of the evolutions you expect to see within the technology?
[00:16:46] Karla: Oh, gosh, that's a big question. Um,
[00:16:49] David: maybe I should say two months instead of 12 months. Yeah.
[00:16:51] Karla: 12. Exactly. Well, the reality working in this industry is decades happen in days.
Yeah. Uh, when I look at what's possible now, [00:17:00] even a year ago, I never thought that it would be possible. Um, it, it was just recently announced that the first fully AI hospital was created in China. Um, and so that is a fully AI enabled, um, hospital in terms of, um, uh, management, administration, um, uh, physician enabled, et cetera.
Um, uh, you can now create video, uh, with audio. Through a simple text prompt, we've created a sales BDR agent, uh, where our clients can come in and speak to an interactive video based agent that is a digital twin of our top sales person, and it can answer all of their questions around security, around what Adventive does, uh, and all of their possible questions.
A year ago. I never thought that would be possible. So, and uh, what we're also seeing with the rise of these new models is incredible expansion, uh, in intelligence. Um, they ran [00:18:00] the, the state-of-the-art models, uh, today, uh, through the Mensa Norway test. Um, and one of the models scored 148 iq. Another one was 135.
Genius level for humans is 130, and that's not where they're stopping. This is the worst that AI will be ever. And so if I look at 10 months from now, what I expect is, uh, a continued increase in the intelligence of these agents. So not only will it be like having a hundred experts in your pocket, it'll be like having a hundred PhDs in your pocket.
I think it's gonna have an incredible impact on, um, production. Uh, the, the creation of, of video, of, of copywriting, of content and the ability to hyper-personalized and create very targeted, uh, and very relevant content to all of, of your audience.
[00:18:53] David: What's been the biggest hurdle you've found in terms of helping leaders?
[00:19:00] CEOs understand and adopt these agents.
[00:19:02] Karla: I think you nailed the, the underlying problem as adoption. Uh, a lot of leaders look at this technology and see it as a technology opportunity or a technology problem, and, uh, it becomes the realm of it. And, uh, you know, they might spend a million dollars creating a custom chat bot, but then they're disappointed that less than, you know, 15% of people adopt, uh, the ai, uh, despite, you know, the million dollars that they've spent on the technology.
And the reality is, is this is as much a. Talent opportunity and a talent challenge. As it is a technology challenge, it is only as good as the experts in your organization adopting it and using it effectively. And so when I speak to expert, uh, to leaders, uh, the real opportunity is to understand the best technology for your organization.
Your. Case, but equally important is understanding the people and understanding the workflows at the [00:20:00] front line. Where are the friction points? Where are the opportunities? Where can your, your, um, uh, workflow actually be transformed by using ai? Uh, and, and that's only gonna be. Done. If you speak with the frontline people who are actually executing for you, once you understand that workflow, then it's showing people that this is about opportunity, not job risk, and that they can play a fundamental role in helping transform their organization because the reality is organizations don't change.
People change driving. Um, uh, engagement in that change is, is the real unlock when using ai.
[00:20:41] David: And I imagine one of the challenges you face there, especially if you're speaking to those frontline staff about their processes, you really need to have buy-in from them in terms of them actually wanting the technology to come in.
Otherwise you're gonna get them perhaps being a bit nervous about sharing what it is they truly do, where they truly spend their time. It could be a very difficult [00:21:00] challenge.
[00:21:00] Karla: Absolutely. 'cause, 'cause a AI comes with equal parts opportunity, uh, and equal parts risk, uh, but equal parts fear. And the reality is technology is changing at an exponential rate.
Organizations and people change at a linear rate. And the unlock for organizations is how do you have your people, uh, e embrace the change, be part of the change as opposed to fearing the change? 'cause they're worried that their job is on the line.
[00:21:28] David: When I hear your story, I kind of consider you sort of an early mover in the space in terms of understanding agen ai, efficient at building it, presenting it.
How do you see yourself kind of being able to maintain that sort of early mover advantage you perhaps had as we, as we go forward into the future?
[00:21:46] Karla: So part of part of that is, um, playing in two worlds. One world is understanding at, at an in-depth perspective the pains and gains that [00:22:00] organizations want.
And so we've been very fortunate in that we have acquired, um, uh, 20 enterprise level clients and have begun, uh, engagements with them. And so that gets us in the organization and it gives us a, a. Front row seat, uh, on the workflows that they're using AI for, and the opportunities to transform those workflows with ai.
And so we stay close to our clients to provide value and understand at a granular level where AI can help even more, um, whether it's within that department, other departments, interdependent changes, and what requires transformation versus adaptation. So. Understanding the human factors that go into running a department and the workflows marry that with staying on top of the technology.
So we're actually partnered, um, with, uh, the Vector Institute. We are working with them to develop the next generation of agen workflows. Um, and so we're part of a [00:23:00] program where we're creating, uh, orchestrated multi-agent workflows where agents, uh, are being managed by other agents. We're working with them around governance.
Evaluations, advanced use of memory. Um, and that lets us level up the value that we create with, uh, with the agents that we have today.
[00:23:20] David: And, and do you think, like you said before, because you're not a, a technical founder, and I do the air quotation marks, because clearly you've, you've, you've studied up and you are very knowledgeable now in the technical aspects, but do you think perhaps not coming from.
That background has allowed you to kind of maintain that human-centric approach, and that is something that's. Helping you guys to be so successful?
[00:23:43] Karla: I would fully agree, uh, because I think a lot of people who came from a technical background, so developers and engineers looked at this technology like software and software is binary.
It's off for, its on, there are very clear rules because if you opened Excel [00:24:00] and it made the formula slightly different day to day, um, that's not really great if you're a finance, uh, person or if you're a data analyst. If my answer is. Sort of the same. Um, and developers come into this with a rule-based, uh, view, uh, whether they're, you know, using Python or React or any number of languages, what I've seen is technical people who write with ai look at it with an engineering mindset.
Uh. People like me who came from non-technical or uh, creative or human factors point of view, uh, actually began to talk to the ai and we applied the systems thinking of how people react to instructions, to tasks, to goals, uh, and married that with. Different way of utilizing the tool. And so it forced us to look at the human factors, uh, as well as the way the AI worked, which was non-deterministic.
Um, and so really it was a blend [00:25:00] of technical systems, understanding versus development and creative problem solving, which is the human side of things.
[00:25:10] David: And I think yours is, again, your story is particularly inspiring because I think there's a lot of. Perhaps would be entrepreneurs out there following the AI trends are excited by its capabilities, but are perhaps holding back from kind of really launching into starting a company.
'cause maybe they don't have that kind of technical expertise. So I think it's, it's great to see that it's very much possible and not only is it possible, it may just be. A competitive advantage.
[00:25:36] Karla: I agree. Um, but what I would say though, um, is that's the entry, that's what created our early IP was when people began to work with our platform, it felt like it was talking to humans.
So not to anthropomorphize the, the agents themselves, they are algorithms. They, they are, uh, machines. Um, they don't have feelings and they're not conscious. But for the user, it felt as familiar [00:26:00] and as easy as. Talking to a colleague on teams or, or WhatsApp. However, if you wanna scale at enterprise level, the engineering does become important.
Um, and when I talk about agents, particularly with enterprise, they're 25%. Um. Prompting and, and AI and agents. But the 75%, which is hidden to the user is actually security governance, uh, um, evaluations, observability. And that's another critical aspect for organizations who are wanting to implement ai. Is, is that content that's below the line is actually hypercritical to, um, really pay attention to moving forward.
And what separates strong ai. Entrepreneurs and, and, and, uh, platforms is that they're sweating the human factors. They're making sure that the interaction is human and easy and, um, intuitive for the end user, but they're [00:27:00] spending a lot of their time on what's below the hood to also make sure that it's safe, risk-free, and secure.
Because as we move into this era of. Agent and agent to agent interaction, the risk goes up exponentially. Uh, we worry about things like, uh, tool poisoning, um, like agents failing and, um, not following rules about error handling, about providing our clients with observability on the choices the agents make.
How errors are managed and how risk is managed.
[00:27:30] David: Kind of leads perfectly to the next question about what should. A forward thinking CEO be asking themselves right now.
[00:27:37] Karla: I think they should be asking themselves about how they want to transform their business model to really thrive in, in a new era of human and AI collaboration.
Because that is going to be the order of the day. As the AI becomes part of all of our businesses, part of all of our processes, part of [00:28:00] the day-to-day of all knowledge workers. If, if you use a computer, AI will become a fundamental part of your life. The Pandora's box is open and, and that is gonna be the reality, uh, more and more every day.
As this technology becomes smarter and, and, um, uh, more embedded in, in. Virtually every program that we use. And so a forward-looking CEO needs to look at a number of different elements, first and foremost, talent. Um, how will they make their talent feel like they're part of the journey? How will they take away the fear?
How will they manage change management? Um, how will they understand the friction points, um, that that talent feels, and take that away with AI so that that talent embraces it. They also need to very significantly look at risk management. Uh, what are the risks that AI can embed into their business?
Whether that is, um, uh, observability, whether that. DeepFakes Cybersecurity. There's a whole host of different risks that an [00:29:00] agentic future holds for every organization. So having the governance protocols and the experts that can help you manage that risk, uh, is, is hypercritical. Like I said. The third part is having a transformative vision.
Um, ai, if you look at AI as. Simply, uh, a way to increase your productivity or reduce your staff. You will be left behind by the disruptors that are actually fundamentally changing business models and ways of doing work differently in your category. Take this for an example, right? So if we take a look at various technology revolutions, uh.
A, you know, adoption of electricity, uh, took, you know, close to a hundred years, uh, adoption of the internet. Um, the computer, the mobile phone, uh, ranged from, you know, two decades to five decades. Chat, GPT in generative AI reached 50% adoption in six months. And it's only getting faster.
[00:29:58] David: So we [00:30:00] have a, we have a tradition on the podcast here.
As we, um, as we get towards the end, there's two questions we like to ask. First one is, what's one thing you've changed your mind about recently?
[00:30:10] Karla: So I used to believe. That the larger the organization or the more senior the person, the lower the capacity for change. Large organizations are, are built on processes and the more senior, the more expert you have, you know, the, the more that you should have all the answers.
So you're not gonna take risks to change. And you know what? Change is hard. Change means. You're gonna fail that there's a higher risk of not knowing the right answer. Uh, and so most people avoid it. Where I've changed my mind is, uh, that with the right parameters, with the right attitude, large organizations, um, and senior people have an incredible capacity.
So if they're put in the right environment with psychological safety, uh, with the right communication, uh, with the right opportunities, [00:31:00] change can be fast.
[00:31:01] David: The second question is, what's one thing you've not changed your mind about and is kind of a core belief that you live by and want to share with others?
[00:31:11] Karla: I'll build on the last answer, which is that change is the only constant. Um, and that especially in this day and age, uh, where transformation happens daily, uh, change is never, uh, gonna slow down and it's only gonna get faster. So your ability to learn continuously is actually more important than what you know today.
And I believe that, uh, insatiable curiosity and willingness to learn is actually gonna be the new and most important technical skill, uh, in the future.
[00:31:44] David: Thank you so much for your time today. It's been an absolute pleasure speaking with you.
[00:31:48] Karla: Oh, it's been a wonderful discussion. Thank you for having me.
[00:31:50] David: Now we have it.
Carla Cosson, the founder of adventive. If you enjoyed today's episode, please like, subscribe or leave a comment. Even more importantly, if you [00:32:00] know someone who could benefit. From the wisdom that Carla shared with us today, please go ahead and share the episode. I'm David Jepson. See you again next time for CEO: Behind the Scenes.