AI for Local and Global Business Growth with Polly Barnfield OBE - VistaTalks Ep #162

Foreign.

Hello, and welcome to Vista Talks. Interesting discussions with interesting people from all around the world. I'm your host for today, Simon Hodgkins, and I'm very, very happy to introduce you all to Polly Barnfield, obe. Polly is the CEO and founder at Mabie. And Polly and the team at Mabie are on a mission to transform businesses with, let's call them, tailored AI solutions. Polly will tell us a lot more about this in a moment, but those solutions that can help deliver immediate and extraordinary results. So from working with UK government, FTSE organizations and around, I think at this count about 35,000 SMEs maybe are doing some extraordinary work. Let's find out more. Polly, you're very welcome to Vista Talks.

Simon, thank you for inviting me. It's lovely to be here.

So let's jump straight in. Polly, I mentioned that you're the founder and Chief Executive Officer at mabie. Can you start by telling our audience, explain to our worldwide audience a little bit more about maybe and the core of the work that you're doing, because it's fascinating and everybody's interested in AI, obviously.

Oh, it's so exciting, Simon. I mean, AI is just as a dyslexic that has, you know, my brain runs faster than my fingers. It's it. The world is opened up in front of us, isn't it? So maybe's vision is to help every business, regardless of size or technical capability, deploy AI safely and securely. We started working with AI over six years ago. We won a government contract where we had to collect the social media data from 3.9 million businesses and understand what it meant across over three and a half thousand high streets. So we started using AI extensively to do that before generative AI was a thing. And we used AI to then help us deliver contracts around training. Now in 2020, when AI, when ChatGPT, in 2022, when ChatGPT was released, I said to the team, do you know, the thing that we built internally is actually the thing that we should sell? Because being able to use AI to deliver services is so powerful and that's how we've served and trained so many businesses. And so we then set about building the tools that we were using to run our business as tools that other people could access. And it's really, it's so exciting seeing how AI can help us do our best work faster. I'm not a believer that AI is going to take jobs. Yeah, I think jobs will change in the same way that when motor cars were first invented, somebody had the job of walking along in front of them with a red flag, well, that job doesn't exist, does it? But do you know what I mean? It's, you know, it's an industrial revolution that's going to create change. But I think the creative thing is, what can you do with AI? And we're really passionate about helping businesses do that.

Well, Polly, thank you for giving us a little bit of an introduction to it. And it's interesting. I have spoke to a few people who were kind of in the a. In the AI space prior to the big AI sort of global awareness boom that happened not too many years ago now. But also what you said there about, you know, you're not a big believer that AI is going to take away all the jobs. And that was one of the initial concerns, I think, and maybe still exists in some areas, but I think it's change for business maybe to. It's going to take away all the jobs to how do we keep up with this stuff? How do we stay ahead and how do we actually use it to make a difference? So, on that note, let me, Let me ask you about a specific, if I can, because I know you're working and have worked a lot with Birmingham City Council, and I know that this involves content and importantly, languages such as Hindi, Urdu, Spanish and others. So can you maybe expand that project for us? How is AI transforming the content and the services in these sort of multilingual communities working with account, you know, a very large council in this case.

Yeah, great question. So if I take you back to the beginning of our journey, when we started listening to all those businesses and, you know, where we cut our teeth on building AI, one of the big bits of data that came out of that was across the UK, only 19% of businesses create any content at all on a monthly basis. They don't. They don't say anything, they're mute. And that, if you like, has really set us on a path around counts, around cancer contracts to help SMEs create content and say something. And that's where the real collision of AI and generative AI came together. Because, you know, if you don't, if you don't have a presence, if you don't exist digitally, people don't find you. And so from a. From a high street agenda and from a place agenda and from a tourism agenda, if you're not being found, you know, actually a place has a problem because you've got to be sending out your digital signals for people to go, hey, yeah, I'm going to visit you know, Chinatown or wherever it is, you know, if you don't, if you're not. If you're not findable, people don't turn up. And so Birmingham City Council is one of many local authority contracts that we've delivered where we provide the tools and the training to help businesses understand why they should generate content, how to generate content. You know, five years ago it was about training them face to face. Now it's acceptable to do it all virtually. And the most exciting thing in Birmingham, because it's very multicultural, is they said, could you do it in multi language? And we said, yes. And we're now delivering. So we've got a massive support knowledge base that very much is geared to specific industries and we deliver responses in 13 seconds using AI. But it's not, it's not, it's not from chat GPT, it's from A. From a. From a specific knowledge base that, that we, that we've blended with our product. So it's, you know, how to achieve what you want to do with it. Schedule a post, create a reel boost that. Build an audience. Audience. All of those things that no business get stuck on. But that happens in real time or in 13 seconds, real time in the language that you input the question in. Now, have we had a native speaker check that the urge is correct? No, we haven't. But have we had a business using it? Tell us it's great. Absolutely. And the difference here is that if you've got. If you're a busy SME business owner and you're stuck, your choices are go to YouTube or go to Google and suddenly you've lost three hours because you start Googling or looking and you find everything and it's like, oh, I forgot what I started off looking for. Or in these business support packages, we deliver. You know, if you say, I don't know how to create my first TikTok or I don't know where to start, you've got an answer immediately. And through the way we run our AI, if you. If we sense you actually need human handoff that it then points you to. We run daily sessions and it points you to that. Now, to be fair, those sessions are run in English, but we've got. We can run the AI in the background. So they could be. But they can be auto translated. But the power of that, Simon, is massive. So from Birmingham City's councils and it's a tourism focused thing, you know, suddenly they're being able to provide business support to businesses who don't necessarily think digital is important. But from Birmingham City Council's perspective, they can up the volume of each business. That has a massive impact for Birmingham themselves because suddenly more digital signals out there, more signed posts, more people can be found. And our metric for measuring success on this is. Is Google more is. Does Google find more businesses? Birmingham. That's the success success metric. Yeah.

And it sounds like a really interesting and useful flywheel effect, doesn't it? Because as those digital signals, as you're describing them and more of these businesses in that community are found that that ultimately creates more sales, more revenue, whatever they're doing, but also probably leads to job creation and that has a knock on effect obviously to the wider community.

Absolutely, absolutely. I mean that's one of the criteria is jobs retained, jobs created. Yeah. And you know, it's just being able to empower local businesses to reach more people which drives economy full stop.

Yeah, it's great because obviously you think about this on a global scale. There are lots of communities, you know, here we're talking about a large Birmingham City Council where it's a very, you know, there are a lot of multilingual requirements in terms of providing that content to businesses. And it's interesting because I wanted to ask you a little bit more and dive a little bit deeper if I can, about tourism and about local businesses, because they often struggle, don't they, with that customer engagement, especially in multilingual context, you know, digital first content, sending out those digital signals. But maybe's AI solutions, obviously you're helping in areas like, like Birmingham to improve tourism related conversations, you know, to have these better conversations with more diverse audiences. Right. And we heard a lot of complaints initially when AI became really sort of popular, more well known that a lot of it was trained in, you know, English only, etc. Etc. But what lessons have you learned? You know, what's the impact of having that knowledge base that you've built up? That seems like a critical component here in terms of that knowledge base that you mentioned. And what lessons have you learned along the way with implementing AI in a diverse linguistic and cultural context? Polly. Because there must be a lot of learning that you've seen it maybe.

Yeah, it's really interesting. So, you know, we would never have been able to deliver the contracts we have from, with a people first approach. Does that make sense? Because there's just not enough. There's not enough time in the day, there's not enough, there's not enough that the contracts aren't the, the value exchange wouldn't work. We couldn't do it. And people will be waiting for an answer. And that doesn't work because when I'm stuck, I need help right now. And so the. And you could take that view of, oh, we can't, we have to check the answer before it goes out. And I'm a. I'm a bit of a it'll be all right kind of person. And I think it's that letting go with AI, that is the first thing. And so what you, you know, you could say, well, actually, we've got rid of our customer service team. We haven't. We've got one person that oversees it and manages the human handoff. But what we've done is invest much more in our content because it switched the dynamic and therefore we invested much more in creating our content. Because it's not just chatgpt telling you how to create a place, it's giving context of how you do that, in the sense of how you use the content scheduler, how you then manage a CRM. And it's linked with. In Birmingham City Council's perspective, we've tied it in with the maybe product that gives them all the tools as well. So it's both. How do I create the content, how do I schedule the content, how do I make it sure it's good? All of. All of those things, which are subtle nuances that, you know, a post isn't just an image and a piece of content, it's. It's. Is the caption right? Is the call to action right? Do I put a link in it? How do I make it a reel? There are so many questions that can block you as an individual, but if you can literally ask questions as you go along, it makes it possible to deliver it, then transpose that and say, it's multilingual. Boom, really game changing. So if the multilingual bit isn't perfect, is that better than just getting the answer straight away? And I believe the answer to that is yes. And the feedback we've had from businesses is, again, yes, because I haven't gone and got lost down a rabbit hole with. With YouTube and seeing a million things, not knowing what to listen to anymore. It's. And the fact that we blend it with daily calls so that they can jump on a call with one of our team in the morning every day and go, I've got this problem. It works really well. And I think every business has got a model inside it where AI can have a dramatic impact on a process. You know, where. Where are your customers getting stuck? Where are your team getting stuck and if you're not yet sort of adopting an AI first approach to things, I always say to a business, think about, think about the boring, you know, how could you banish the boring bit of your business? How can you banish the bit where customers are unhappy, your team's unhappy and AI is really good at doing, I call the hard stuff fast. Yeah, anything where you can write an SOP for it and give it the knowledge and invest in the knowledge, it does it really, really well. I mean, yeah, it's, it's your perfect employee for doing that thing that gives your people more time to do this. Because it's about human engagement that matters, isn't it? It's, that's what we're really good at. And the more facetime you can have with people, the better your business runs. That's how I see it.

It's fascinating to hear you because what I love about the way you frame that is obviously we work with a lot of global massive brands who have budgets, they wouldn't dream of sending out content that hasn't been checked by a language expert, you know, and of course the type of content, you know, tourism is different to, let's say healthcare or life sciences or you know, that kind of thing. So the type of content obviously matters, as I'm sure you know, as we know. But what's also interesting is how those small businesses then are using, I'm going to use the term incorrectly, but it's almost like a self serve model. They can find the resources, they can get started on their multilingual journey using the AI tools, working with an organization like maybe because prior to that they don't have anything. So there's nothing happening in that space and I completely understand that. And I know, you know, I've mentioned a number earlier about, you've worked with I think over 35,000 businesses in this space. So I mean you're doing this at scale and it's interesting when you say that we couldn't do this without AI, it's just not possible doing it with a human first approach and I can see that obviously. But when it comes to digital marketing, sentiment analysis, the role that they play with these businesses to connect with their customers effectively, you're getting real time feedback, aren't you, from the customers who are saying this is superb, this is brilliant. And you must be seeing the capabilities of the AI improving over time too from when you started doing this to the latest models today. Am I right in saying that, Polly?

Yes. I've actually got two things I Understand this. I want to turn it on ahead and talk about how we can help, how we're helping global brands as well, from the other way around. But let's talk about training the model so models don't learn at the moment, each model gets better, but actually as regards running the system, it's how well you train it. So we listen much harder to feedback, both through AI, but also as a human being. It's the nuance, it's understanding. Are the questions they're answering, asking, being asked, well, where are the delays? Where is the unhappy customer? Is the actual content we've, is the base content we've given wrong? Are we giving an agent too much creativity? Do we need to turn it down? Because it's all about you've got to treat an agent like you treat a person that you employ. You've got to regularly check in with them, give them a review and think if there's a new model, how does that change how they behave? But I think what's really interesting and the work we've done with SMEs and working with them at scale and as you say, Simon, they've got nowhere because they're necessarily starting from zero. We have sort of taken the handbrake off and adopted that AI first approach and gone, let's test it. And it's been amazing because it's let us do that at scale piece. But if you flip that on its head and then talk about an organization that maybe employs 35,000 people or 100,000 people and you look at their needs and take the content stream as an example and you've got global brand, you've then got regional brands, you've got legal requirements, your image requirements, you've got copy requirements, you've got all of that and it's, it's a nightmare to deliver it. So we're working with a number of agencies and global brands to look at how you take the learnings from training, you know, a bottom up approach and apply it from a top down approach. Because you know, if you are a global brand and you want to be able to newsjack a story, or you want to be able to embrace AI in any way, shape or form, it changes so many business processes. So with all of this, the key piece comes down to documentation, the key piece comes down to what is it that you want to do, how do you want to do it and whether that's the knowledge base or whether it's training the agent. It's, you know, every image has to have this legal sign off it has to fulfill these criteria. This is the way as a brand we behave. And that means that you can then have a system that will say, no, don't even, don't even pass this piece of content onto the next human for part, for sign off, because these basic things are wrong. So you can improve the efficiency of the human machine tremendously by applying that same learning. And I'm really excited about some of the projects we're working on there as well. It's really, really exciting.

Yeah. And it's sort of like becoming part of that. Whether it's that global brand, as you mentioned, or a small business, it's becoming part of the required workflow, then.

Yes.

And I, I loved what you were saying as well. Just going back to scale, for example, and not handing off to a human until the various parts of that workflow have been been sort of checked off or checked, you know, by the AI and the model and the training, of course, that you apply in the first place. But I, I'm reminded, as you were talking about the Chief Executive officer of Salesforce, Mark Benioff's statement, because he's talking an awful lot at the moment about the managers and the leaders of tomorrow, whether they're in small, medium or large businesses. You're managing not just human labor and inverted commas, but digital labor as well. Now, in terms of AI, which is kind of what you're touching on there, isn't it, Polly?

Yeah, and I, you know, I see it in the sense of that I think all of our teams should see AI as the colleague they wish they had. They should. They should see AI as the colleague. That means that they do get finished on time. You know, which one of us honestly goes home at night thinking, yes, my desk clear. I mean, does that ever happen then? World. My world. And again, honestly, as a CEO, my view is that every business is messy. You know, it changes, it evolves, it twists, it turns, it ducks, it dives. And if you can use AI to do the stuff that we're not getting done, that's getting done badly. I mean, you know, sorry, Mark Benioff, but who actually likes filling in a CRM? Well, actually, AI does that incredibly well. So the reason he's so articulate on agents is because actually, the reason that companies fail to use CRM, because people hate filling them in, agents do it really, really well. And so if you can start to think, what is it that you can get AI to do? That means the human, human performs even better. It's magic. And so, yeah, it's absolutely about thinking, how do you manage? It's another workforce that helps your human capital deliver a have a better life and deliver more. I think it's really exciting. It's really exciting.

Yeah. You mentioned agents there and I know a lot of businesses are experimenting and starting to use more agents in their and workflows too. Which brings me on to something I must ask you, because security and trust raise their head here, don't they? They are significant concerns for a lot of businesses who are considering AI. And I know that maybe you take this very, very seriously in terms of your, you know, being SOC 2 compliant, et cetera, and I know that you're building with safety in mind, Polly, but could you maybe discuss that a bit? And then how do you ensure that you and the team are making sure that it's secure, it's reliable, that maybe's tools, you know, because they're becoming so crucial for the businesses adopting AI solutions, security and trust are really paramount, aren't they?

Oh, absolutely. You don't want your secrets going to the wind. You've got to. I always say to somebody who says to me, well, we can just use Chat GPT. I go, yeah, you can, but just let's think about this. Go and ask ChatGPT what it knows about you today. Just go and pop it in in the, you know, the front end of the public Chat GPT, see what it tells you. Now imagine you tell it all your secrets and ask it the same question and it tells them all back again and anybody can find that out. How would you feel then? So you've got to remember that Open Chat GPT is being trained on whatever you put into it. So if you want to use it as a company, you need to be in a safe space, you need to have an enterprise account. So we have a SOC 2 agreement with OpenAI, which is the same as Mortgage and Stanley have, for example. That means if you're using it within the maybe wall, everything you do is yours, isn't it? The data is not shared with anybody, it's not used to train the model. You'll never tell it something and see it back somewhere else. And I think that in order to use data with AI, you have to do something called vectorize it, which means you can look at unstructured questions. Amazing. It's an incredible piece of technology, but it must, you must, must put your company data in a secure place. What we're really excited about is that we've nailed down security as a company. We've also about to Release multimodal. Which means that for those of you who've used Claude or Perplexity, you know, they're great, they do other things in other ways. It's really fun. And so we're saying, right, once you ask your data safe, it's all your secrets are never taken and used anywhere else. But actually what you want to know is when you ask it a question that the right model answers it so really exciting that it says, hey, we've used Claude for this because you're asking really creative problem or you wanted the web. So it looked at Perplexity and we are taking a local version of this week. Deep Seq's been released, you know, massive news about that. Would I do I want China to have my secrets? Not really. Sorry, China. So I wouldn't, I wouldn't tell Deep Seek something publicly, but we're running it on a model that lives on the server. So you've got the access to the model but without the data going anywhere. So just, I'd say you've got to be very careful as a business. Be very sure that wherever you're using AI, it's an environment where your secrets aren't going to the wind. Just, just say to yourself, do I want to see that somewhere else Again, else. No, don't put it in a public model. Make sure you're putting it in somewhere where you've got security, which is what we focus on.

It's interesting you mentioned deepsea because obviously that's a hot topic at the time of recording this discussion and we saw the impact it had on other AI and chip stocks, for example, companies. But yeah, people, I, I do know some businesses that are using it, as you say, at a local level. They've downloaded the model. They're not, you know, the data is not going back to, in this case, Chinese servers. But it is interesting in terms of some of the claims that they're making. I believe there's some, even some more important reports coming out today, the day we're recording this, just in terms of its actual bias and some of its inaccuracies that are coming out. So that's going to take a little bit of study. But you're right, there are different models for different types of work and being able to pull those together is. It's a bit like being the orchestra of a band, isn't it? You've got to get it right. So it's good to hear that. And thanks for, thanks for talking a little bit about security and trust there,

which is massively Important. It is the most important thing.

I gotta change direction then I'm going to come back to what we're talking about here because there's a couple of things that I've got to ask you about. Right. Somebody told me a little bit of Tommy, that you won a gold medal a little while back now, a long time ago, for Team gb. Tell us about that first.

Well, in a previous life I used to ride a horse for a living and I was very privileged to go and winner European medal for three day eventing. So yeah, it was really fun. I, I, I love adrenaline. I run a tech company and before this I used to ride, ride horses and go fast for a living. So yeah, it's, yeah, a lifetime ago,

but yeah, so the adrenaline went from horses to technology.

Yeah, exactly.

It's still there.

Very good. Exactly.

Oh, thanks for sharing that, Polly. That's great. And the other thing I have to mention though is you've also raised over 30 million pounds for charitable causes and I know that you are honored with an OBE for your services to philanthropy. So could you maybe tell us a little bit about that because that's phenomenal. And also maybe tell us a little bit about how the experience in that philanthropy type environment and the great work that you've done there, did that shape in any way your approach to innovation, to business leadership, maybe in scaling maybe itself and driving impacts like tell us a little bit about that story because getting an obe, that's a kind of big deal.

Thank you, Simon. So, well, it links quite well to the gold medal. I was, I've been, I skipped medical school because I got shortlisted for a British team and much to my parents dismay said, no, I was going to ride a horse for a little doing. And after 10 years and hadn't had a whole load of fun, I realized that probably that wasn't the best career choice I'd ever made. Broken lots of bones, but massive fun. And I thought, well, I better go and get a proper job. So went to the Internet, was brand new at that stage of time and went to work for web design company. Spent about a year thinking, yeah, love it. And it was very computer literate because I run my own business with, with horses. But I, after a year thought, no, I can do this myself. So I started an agency and colleague when was going off to bake cakes for say the children one day I said, well, surely we can use the Internet to raise money. You shouldn't be baking cakes. You know, we're a tech company. And so we had this, this idea. I said why don't we, why don't we build a search engine that Because Google at this time Google. Google and Yahoo so battling out and every time you, you, you click on a search result you could revenue share. I said well let's have a business that we start and we give 50p every time somebody raises some money through our search engine we'll give 50p to charity and in our mem and arts as a business it says for every pound we get and we give 50p away. And insanely I managed to raise money to start this business. We became the fourth biggest search engine in the uk. Our model is interesting in that we said we didn't choose the charity, you did. So we took the data from 180,000 charities. I've always been very d. I've been a data geek and yeah, so we started that and we were using Google search feed and it was all amazing really growing fast became and I then went to San Francisco to see Google and they yeah, let's just say the meeting didn't happen and we had to then start using Yahoo and it was challenging and we, we pivoted and we built a business called give as you live that bought an affiliate revenue. Long story short we got down the roadside of charity through people clicking a button to buy a thing or search for a thing which was extraordinary. And I'll never forget the. The day I sat in my office and this emblet came with the royal crest on it opened it and said her Majesty the Queen and David Cameron reminded to give you no be. Oh and it's a secret, you mustn't tell anybody just fill in the postcard and say yes or no. So I rang my husband as you do darling I've got to say it's must be a joke, shortest joke. And so he said we'll say yes. So and he says better go buy a new frock. And so you take the box and send it off and it was a really sort of. You felt so I felt so on. I still to this day don't know who did it because it's a really big deal to get one, you know. So some whoever did that. Thank you so much. And then on the day I bought a lovely orange dress and a big black hat and we went to back palace and had a lovely lunch and it was incredible. So yeah, it's amazing really. It makes me feel quite emotional every time I sort of remember it. But yeah, whoever, whoever put me forward. Thank you so much. And as from a charity Perspective, it's, the business is still running, so it's still raising money now. And it's just lovely to know that it's. Technology is helping, you know, cause that matters to you. So whatever the welfare society or the Red Cross, you can use the give as you live product to raise money for it. And I think that's really exciting and that technology is delivering value back to society. So.

Yeah, well, congratulations. And obviously it's an honor to pick up an OBE and to be invited to the palace.

Yeah, it's lovely. It's lovely.

So that's, that's, that's phenomenal for all the great charitable work that you basically created and continues today. So phenomenal. So I hope you don't mind me just raising both the gold medal and the obe, because we have to mention them, Polly. And look, let's finish off, let's come back to AI to finish off our conversation today, because one of the things that people are really challenged with at the moment, whether you're the CEO of one of those big global brands, whether you're running a small business, or whether you're a business leader within an organization, today you may be a little hesitant about fully adopting AI. Jumping in with both feet. Have you got any advice for how people need to start integrating AI into their operations to help them ensure that growth, to help them also maintain that human touch in maybe their international growth and business strategies?

Yeah, absolutely. Number one, make sure that wherever you're doing it, you're putting your data in a safe place. Number two, start. And number three, pick a, pick a thing that the team hate doing. Don't overthink it. Pick something small. You know, this is an elephant that every business has got to eat. Just chew off a tiny bit. Because the minute you have a little win with changes that it changes the thought process. If you try and overthink this and try and think, what's the. How do we change the system? By the time you've decided what's going to happen, AI will have changed again. And it's, it's a very, it's a very interesting way of working. You know, we have a deeply technical team. I'm really privileged to have. You know, we've got a really, really great tech team. And I find that every time that we tackle a new business challenge, it's actually the companies that move fastest with this are the ones that start small and they go, this is the thing that's really hard. It takes a lot of time, or if we could fix it, it'll Drive revenue or if we fix it, it'll make our team happier. So just find something. It could be a process. It could be, you know, it's just, it's just one thing. And don't overthink it. Learn through that one thing. You know, choose something that's going to, you can deliver in six weeks and think about it in the sense of what's the value it's going to deliver. And yeah, think of it like hiring a. Yes. Just, it's just, just I want to fix that thing. Don't, don't over overcook it. That's the key.

I really like that. You know, pick the thing that's annoying the heck out of everybody, everybody and start. Because when you get that small success, you'll go, oh my goodness, this is phenomenal. And it'll help you in many areas. And look, before, before I do finally wrap up Polly, is there anything that's on your radar? What's keeping you up at night? What are you thinking about? What are you excited about over the coming months ahead? What's on your, what's on your radar?

So I had a really great team meeting this morning and we've set a big, set a really big goal for ourselves, which is that we think the future of our product is that. So we're building agents, we're building, we're building an agentic platform right now that's multimodal and it does so much. It's got plugins from every, everything you can imagine. So Google Analytics, Facebook, Adobe, everything. So, so literally, if you imagine what good looks like for us is that Simon comes to the box and says, says, hi Simon, what do you want to do today? Simon says what he wants to do and it chooses the model, it chooses the agent, and it gives Simon the result. So the, the business, the business infrastructure and the value that the business, the people in the business creating is about defining the process of what, how Simon does his job. Well, because imagine if you hadn't got to go and find the file, find the thing, find that, find the data. You haven't got to log into 10 different products to do the CRM, get the analytics. So how can you blend data and AI in a smart way so every can do their job faster? And that's our vision and we're busy executing it. And yeah, we're really excited. We're really excited.

Yeah, it sounds, sounds really exciting. We'll have to check back in Polly and see how things are going as we progress through the year. Look, it brings me nicely to the end of today's show with Polly Barnfield obe. I want to ask people to make sure that they tune in again to see and or listen to our next episode of Vista Talks too, where we'll be discussing some more interesting topics with very interesting people from all around the world. Just like Polly. But Polly, thanks for being on the show. It's been a real pleasure to talk to you today. Thanks for sharing all those wonderful insights.

My pleasure, Simon. Speak again soon.

Sa.