What happens to the subscription economy when consumers stop choosing every service themselves and ask an AI agent to manage the decision?
In this episode of Consulting the Future, I speak with Giles Tongue, Vice President of Marketing at Bango, about the growing tension between subscription demand and subscription fatigue. Giles oversees marketing for Bango's Digital Vending Machine, which helps companies offer subscription bundles or make their own services available through bundles.
Subscriptions have expanded far beyond television and music. Consumers now pay recurring fees for gaming, fitness, news, productivity software, home technology, and AI tools. Giles cites estimates discussed in the interview that value the subscription economy at $700 billion today and project growth to $1.2 trillion by 2030. At the same time, he says consumers are becoming less certain about what they pay for, where they bought it, and whether they still use it.
That tension has brought bundling back into the conversation. The old cable model often forced customers to pay for channels they did not want. Giles describes a newer model built around choice, where customers select services from a larger marketplace, manage them through one account, and may receive a larger discount as they add subscriptions.
Mobile operators have offered entertainment services with phone and broadband plans for years, but the participants are changing. Banks, retailers, and loyalty programs are becoming distribution channels, while content providers are also packaging their services together. Giles points to examples such as Optus SubHub in Australia and the Netflix and Max bundle sold through Verizon to show how indirect subscription sales are becoming a larger part of the market.
The attraction is not limited to discounts. A service provider can use bundling to reach a large existing customer base and remove payment friction by adding a subscription to a familiar bill. For the subscription company, that arrangement can also change customer acquisition economics. Advertising usually requires spending before a sale occurs, while a distribution partner may take a share only when the subscription is sold.
There are tradeoffs. An intermediary can make discovery and payment easier, but it also places another company between the service and the subscriber. Brands need to consider how much control they retain over the customer relationship, data, positioning, and renewal experience. Consumers may gain one place to manage several subscriptions, although a larger bundle can still become expensive or difficult to understand if choice is poorly presented.
Our conversation then moves from bundles to AI because both subjects raise the same question: who helps the customer make sense of abundant choice? Giles says people already struggle to find where a program or sporting event is available. Voice search is one response, but AI agents could go further by finding a service, checking access, starting a subscription, and canceling it when it is no longer needed.
Giles cites Bango research discussed in the interview suggesting that one in five Americans would already allow AI to subscribe, manage, and unsubscribe from services on their behalf, with the figure rising to 40 percent among a younger group. Those numbers should be treated as Bango research until the exact report is linked, but the business implication is worth considering. If an agent cannot connect with a provider, understand its offer, or complete the transaction, it may choose a competing service instead.
Giles connects that possibility with his own experiment in AI filmmaking. He used an agent to turn a single line of natural language into a two minute video while coordinating specialist tools for images, moving pictures, voice, music, and editing. The result was deliberately left a little raw because the point was to show how much work the agent could coordinate with limited instruction.
The experiment also changed his expectations. Less than a year earlier, Giles was impressed that AI could produce a recognizable story. He now wants precise control over camera movement, visual consistency, and brand requirements. That gap between an entertaining experiment and dependable professional output is narrowing, but it has not disappeared.
Giles describes Claude as a workmate that helps translate a creative intention into the language required by specialist production tools. In his account, the person supplies the idea, context, and judgment, while AI helps express the instruction and coordinates parts of the workflow. The practical lesson is that successful AI use may depend less on asking one model to do everything and more on combining tools around a clear outcome.
The corrected production workflow includes Seedance 2.0 for video generation, Seedream 4.5 for image generation, Soul Cinema, and Soul for location work. Giles also discusses Claude, Higgsfield, Supercomputer, Nano Banana Pro, Kling 3.0, Seed Audio 1.0, ElevenLabs, CapCut, and Suno. The number of products is less important than the operating model: one person can direct a set of tools through an agent and produce work that previously required several specialist skills.
For subscription companies, the same operating model may arrive at the point of sale. Consumers could delegate discovery and administration to software, while providers compete for inclusion in the agent's choices. That creates an opportunity to remove friction, reduce forgotten subscriptions, and match services with genuine demand. It also raises questions about consent, transparency, commercial influence, data access, and how a person can reverse an agent's decision.
Would you allow an AI agent to manage your subscriptions, and what controls would you expect before giving it that authority? Listen to the episode and share your thoughts.
Useful Links
Video/Ai/Workflow here
Reports https://bango.com/reports/
Bango LinkedIn
An odyssey of AI video experiments: 11 tools, full workflows, and the gap that remains
[00:00:00] Scale your business with agentic AI with limited risk. With Denodo's AI data layer, your agents are provided with real-time company data and guardrails for company protection. Create the business you always dreamed of with Denodo. And you can do that by simply visiting denodo.com to learn more. But now, back to my guest.
[00:00:24] What if choosing something to watch takes so longer that bedtime arrives before the opening credits? Sound familiar? I must admit, I have been there. Well, Giles-Tung, Vice President of Marketing at Bango, he joins me on this episode of Consulting the Future to examine subscription fatigue, streaming fragmentation and the return of bundling.
[00:00:50] And Giles will explain today why consumers value one place to discover, buy, manage and cancel services. But providers use bundles for retention as well as acquisition. That's just one area that we're going to cover on today's podcast because I also want to learn more about his experiment,
[00:01:13] which involved a two-and-a-half-minute AI film, 11 tools, 15 shots and one dragon that refused to behave. Yeah, it was a project that changed his expectations from producing something impressive to demanding exact creative control. It's a conversation today that will connect consumer convenience, AI agents, business models and creativity.
[00:01:39] But most importantly of all, help you, the listeners, understand why the subscription battle may be won through simplicity rather than an exclusive TV show. But I want to reveal any spoilers. So let me introduce you to my guest right now. Thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do? Yeah, hi, I'm Giles. I work at Bango, vice president of marketing there.
[00:02:08] I've been marketing for well over 20 years now. And at Bango, I'm looking after Digital Vending Machine, which is a product which sits in the subscription economy, helping companies that want to offer bundles or be bundled. And we're making all of that very easily. So we're working with some of the world's biggest streamers, for example, and some of the biggest mobile operators all around the world. Wow. It's a big time to talk about this because I think over the last few years,
[00:02:33] we've all gone from subscribing to maybe one or two streaming services to juggling entertainment, music, fitness, news, AI tools and so much more. So has the subscription economy reached a tipping point where consumers are just simply overwhelmed by choice and a little confused when they look down their bank statement each month? I think at a macro level, the subscription economy is continuing to grow.
[00:02:59] We've got estimates saying that if the size is now 700 billion by 2030, it will rise to 1.2 trillion. So the subscription economy continues to grow. And that's because of the subscriptionization of pretty much everything. Can you remember the last time you bought a digital product that didn't have a subscription option? When was the last time you bought something at B2B, you know, something you'd use at business that wasn't SaaS, you know? So it's becoming a subscription. So that macro level is continuing to grow.
[00:03:29] However, at a consumer level, things are getting a bit tight. People are trying to look after their household expenditure. And we have this thing called subscription fatigue where people are forgetting where and how they've subscribed to something. So people are starting to get overwhelmed a bit. But the average in the US now is around five subscriptions and a little bit higher in the UK, closer to six subscriptions that we all have. I'm well north of that, but that's because I'm constantly tinkering with tools. I've got over 20.
[00:03:59] Wow, 20. Man, I thought I had a lot. I think you've beat me at that 20. But they do soon add up, don't they? Yeah. Yeah. And as you said in your intro there, this is media and entertainment. We think of Netflix, Disney and all the rest of it. But then you're into productivity, gaming. You might have some gaming. You might have some things around the house, the ring doorbell. And suddenly you start to remember all these different things that you didn't realize were subscriptions. And they are all adding up.
[00:04:25] And what I love about what you're doing at Bango is you have a unique view across platforms, whether they be telcos or digital services, gaming platforms or anything in between. So what are the biggest shifts that you're seeing in how people actually discover, buy and manage subscriptions? And how does that differ from, let's say, a few years ago?
[00:04:45] The biggest shift, I think, and it's probably the reason why we're leaning so hard into the subscription economy is the evolution from where once we were buying our subscriptions either from an app store or from a website. Now we're buying indirectly or through bundling. I mentioned that average of five subscriptions, two of which are, you know, 30 to 40%, depending on the UK or US, and now coming through indirect channels. So people are now buying or getting their subscriptions through could be their mobile operator, bank, retailer.
[00:05:14] So that's one major shift, this shift towards indirect. I think content discovery is a big issue at the moment. People can't figure out where to watch the content they're thinking of or they can't find something to watch when they've got these subscriptions. So content discovery is a big issue. And we did a report recently, which is around sports fans because they're the most affected by this. The fragmentation of media rights and particularly a lot of the streamers are trying to get live sports into their service.
[00:05:43] Disney just announced a deal with Formula E just this week. That fragmentation of sports is now leading subscribers to really struggle with not just the cost of these different services, but actually knowing where the game is. So that's a real issue. I think we've seen bundling being something that's been mobile TV and broadband for a while. Now we're seeing banks and retailers and retailer loyalty schemes come into bundling. So that's a big change.
[00:06:11] We're also seeing content providers work with and sell each other's content. So that's an interesting move. So Netflix and Max is the classic example. Those two bundle together through Verizon at a discount of 30 plus percent. And we've got services like Fox and ESPN bundling together, selling each other's services and providing value to the subscriber. So lots of big changes. And then how consumers are actually finding their content evolving as well.
[00:06:38] I think we'll talk about AI later, but around a quarter of people are now using voice control to search for content. So a huge amount is changing very, very quickly. It really is. I'm glad you mentioned bundling there because a few years ago, many TV viewers famously cut the cord to avoid being tied to content that they don't actually watch. And as you said, bundling is making a big comeback, but it does feel a little different from the old cable TV model. So what's the next generation of bundling look like?
[00:07:07] And why is convenience becoming just as valuable as maybe the content itself? And as I asked that question, I know when looking for content to watch myself, very often you spend 40 minutes trying to find something to watch. And by which time you think, oh, I must go to bed now. And you've not even watched anything, right? Yeah, that's right. So bundling used to be something that was used to help with customer acquisition. So you would sign up to a new mobile plan and something would be included within it.
[00:07:33] So you might get, I don't know, Netflix, Disney, something like that included within your mobile plan. Then they were using it cleverly to trade people up. So here's our standard plan. But for a little bit more, you'll get these things included. We've moved well beyond that now to super bundling where multiple different content are available to be paid for and managed in one place. Great example in Australia with Optus, their sub hub, multiple different, they've got over 20 different services there that you can pick and choose from.
[00:08:01] And the more you take, the greater the discount you get against those services. So it's great for the subscriber that they are now able to get these discounts and also to deal with that subscription fatigue by managing those services one place. But that kind of service is also helping with retention. So now we're seeing a shift from where bundling was all about acquisition. Now it's also about retention.
[00:08:23] So the more services that somebody takes from their provider, and I'm using that to describe mobile, broadband, TV provider, the more loyal they become to that first party service. So it's starting to become a tactic or a tool which is used to achieve a great many different strategies. Whereas once was acquisition, now it's retention, differentiation, loyalty, and so forth.
[00:08:49] So the world of bundling is moving quite quickly and being used in many different ways. And listening to you there, it strikes me that companies are not just competing on what they offer now, but on how easy they are to find access and pay for. So how important has reducing friction become as a competitive advantage in the subscription economy? Is this something you're seeing? It's almost a return to the one-click checkout almost where everything's just nice and easy and no hassle? Yeah, absolutely.
[00:09:17] So if we think of subscription brands, I like the media and entertainment examples because everybody knows them, the Netflix, Disney, and so forth. They're always looking to acquire new customers and also retain customers. And a great way of acquiring new customers is through bundling where, say, a mobile operator will offer a subscription. It could be a charged one. It's not always free. You know, sometimes you pay as an add-on. But then you're adding it to your bill, which is like the one-click checkout you mentioned.
[00:09:44] So that is removing the friction entirely from that purchase. There's nothing easier than, yeah, add it to my phone bill or add it to my TV bill. So from the content provider point of view, they're benefiting from this existing payment relationship and really smooth checkout if we want to think it like that. But they're also benefiting from the enormous reach that these operators have. They have enormous customer bases. They have fantastic marketing machines. So they're benefiting from the reach, the easy checkout.
[00:10:12] But also there's a nuance that's a little bit hidden that people might not see. When running a direct-to-consumer business, customer acquisition is often advertising, which is something you would pay for before you've sold a product. Whereas with bundling, it's often a cost of sale. It's a margin sacrifice. You give away some margin to the operator who's promoting your product. So there's a cost which has occurred as you're selling. So it moves the cost around the P&L as well for businesses.
[00:10:39] So bundling's got a lot of benefits to finding, retaining, and also running your business costs and cash flow point of view as well. And one of the things I wanted to bring up today is I loved your recent experiment where you created a short AI-generated film. And in less than a year, your expectation shifted from being amazed of it worked to wanting precise creative control.
[00:11:02] Do you think that's exactly where we're heading with AI in business too, where the challenge is not just capability but better directing it more effectively? Yeah. Well, thanks for bringing that up. I spent some time tinkering around with Gen AI. Initially, it was just still images and now more videos, so creating videos. And I have to look back at the timeline sometimes because it seems ridiculously fast. But less than a year ago, I was able to stitch together enough videos to call it a little story.
[00:11:32] And I was kind of amazed by it and thought, well, this is incredible. I can't do any of this. I can barely draw. I'm not the kind of marketer that's got the pens and pencils or even the graphic design skills. But through AI, I'm able to do this. And what I realized this month when I was just tinkering around and playing and creating a video is my expectations have shifted from, oh, I can create a video. And oh, wow, that looks cool. So actually, I want it to really precisely do this thing. And I wanted to really specifically do that thing.
[00:12:02] So my expectations have moved very, very quickly. And actually, it's that nuance there, that little gap, which prevents it from being something I'm using sort of professionally at work. But it's closing. That gap is closing really, really quickly. So there's little techniques that I'm learning, which is the difference between here's me having a bit of fun and I've stuck myself on a dragon and ho, ho, ho, that's funny. Doesn't that look cool?
[00:12:26] So now I'm working with a brand and I've got specific guidelines and it has to do these things, has to be precise. And I think actually that's the gap that's now just about to close. And one thing that stood out from your workflow was a comment you made that Claude didn't replace your creativity. It actually translated it.
[00:12:47] So do you think that's a useful way for business leaders to think about AI more broadly as a collaborator that helps you express ideas, bring them to life rather than just going away and generating something for us? Yeah, definitely. I mean, when I think of AI and listen to the conversations, particularly in the early days, it was often about replacing. It's going to take our jobs and this kind of thing. Actually, I think it's enabling us to do more and better. And this is an example of where I wanted to create something.
[00:13:16] I had a vision for it. I was pretty sure the AI models would be able to do it. And now I'm thinking more and more sophisticated about, oh, what if I could do what in my natural language would be a drone sweep or something like that? And I look it up and I find some, oh, it's actually called this or a dolly pool or whatever. This language that I'm not familiar with around camera work. But I'll give it to Claude and I'll say, I've got this image. It's going to look like this. I want it to do this. I want the camera angle to move like that.
[00:13:46] Create the prompt that I'll now use in this particular tool, which was C-Dance 2.0, to create that image. And now it's creating a really precise prompt that's mirroring exactly what I wanted to do. But in the right language, the language that is more professional, but also one that will instruct the model to do exactly what we want. And I think that's probably the way to use AI generally as a kind of companion that's going to help you get to the outcome you want to,
[00:14:12] rather than replace your thinking or replace your ideation moment. It's helping you get to the outcome rather than literally doing it for you. You've experimented with more than a dozen AI creative tools while also watching AI reshape digital services through your day job. And I'm curious, from your experiences, where do you think we're still maybe overestimating AI?
[00:14:37] And where are we actually underestimating how quickly it could change the way we work and ultimately create? Yeah, I think probably where we're underestimating is the role of agents. So I'm studying this at the moment to educate myself a bit better. But going back to the video creation two weeks ago, I put a one-line prompt in and it created a two-minute video, more or less what I had in mind from a one-line prompt.
[00:15:01] And what I did was sat there for two hours watching an agent go through the motions of creating videos, asking me to approve it, create an image, a video, a voiceover, et cetera, et cetera. And this was an agent just doing all this for me. I couldn't believe what I was seeing. And I left it there a little bit raw and posted it up on LinkedIn because I thought, people have got to see this. I've done almost nothing apart from use some natural language. And the agents pulled together probably 10, 15 different tools, images, video, voice, music, et cetera, to create this thing.
[00:15:31] And if you apply that same thought process to other areas of your work, whether you're creating, I don't know, presentations or writing code or whatever, what we're seeing now is agents that can pull together different tools and different workflows and do it all basically for you. And you think about the capability that that enables for individuals. I'm sure we're underestimating that. So how many people are actually getting deep into agents? So I'm looking into this at the moment. That's kind of the underestimation.
[00:16:02] And I think relating it back to subscriptions and so forth, one in five Americans would already let AI automatically subscribe, manage, and unsubscribe from services. And that rises to 40%. So in the subscription economy, we may get to a situation where agents are running people's lives around subscriptions or in parts of lives that go beyond subscriptions. And that changes everything. You know, that changes discoverability in AI and AEO and how do agents operate?
[00:16:32] Can they connect to your tool? Let's forget about media and entertainment. Let's say it's a productivity tool. Can they connect and trigger and start your tool? Or are they going to be able to start somebody else's tool? And if you can't do your one, they'll do somebody else. And you've lost that business. So the role of agents throughout the subscription economy and beyond, I think, is going to be huge as well. Wow. Incredibly cool. And if I was to pull out a virtual crystal ball here, ask you to look ahead.
[00:16:59] And whether we're talking about subscriptions, AI-generated content, or digital experiences, what do you think consumers will value most over the next five years? And I appreciate that's a big question considering just how much has changed in five years already. But will it be having more choice or having someone or something intelligently simplifying that choice for them? How do you see this progressively evolving over the next five years?
[00:17:26] Yeah, I think we've probably had in our heads this world of, you know, I've come home from work, the light switch on, the food starts cooking. I was put on the TV. It finds something for me to watch. So I didn't even have to think about it. It already knows about me, my behaviors. It can identify that there's somebody in the room with me and they might have a different preference to what I have. And then they'll sort out, do I have a subscription for this particular thing it's found or not? And it will switch that on and switch it off and manage it all.
[00:17:54] This doesn't feel five years away. It probably is five months away. Where do we get to in five years? I have absolutely no idea. But I think the important thing is to sort of keep on top of it, keep listening to podcasts like this so that you can at least see what's just around the corner. Never mind, you know, sort of around the estate and up the hill and five years away. Love it. And for anybody listening, maybe they wanted to follow in your footsteps.
[00:18:20] I think the best way of trying to understand this technology is exactly what you're doing. Rolling up your sleeves, experimenting and seeing what you can create. I think I'm right in saying this, that when you created your short film, I think it was Higgsfield.ai. Can you just very quickly walk me through the tools that you use to create that? And then what I'd also do is I'll embed the video so people can see what you can create and how you've done it. Yeah, I'd love that. So I've evolved very quickly as we do.
[00:18:48] And I've ended up with Claude as my kind of workmate using co-work particularly. And I'm doing another course on co-work to understand how to optimize for that. But that's the bedrock for the production in terms of writing stories and getting the prompts right and all the sort of collaboration work. And then Higgsfield.ai is the creative tool. Higgsfield has some of its own proprietary tools which help on the production side, including the agent I referred to, Supercomputer.
[00:19:16] So you can use Supercomputer and kind of forget what else is happening. It will do it all for you as an agent. Or you go in and dig into the tools yourself. So there I'm using GPT-2, Nano Banana Pro for images. And I've probably forgotten a bunch of others. And then CapCut to stitch it all together. There's Seed Audio 1.0 for voiceovers or 11 Labs for voiceovers. CapCut stitches it all together, puts the captions over it. Suno for music. So Suno creates music, again, from a really simple prompt.
[00:19:46] And I very rarely needed to go to one or two versions with Suno. It's pretty amazing. So I don't know how many that takes us up to, nine or ten. But that's some of the story. But it's not as complex as it sounds. Once you start tinkering away, you'll realize that one's slightly better for one job than the other. For example, Kling does a really nice job at taking a starter image. And then you animate onwards from there. Whereas Seed Dance will imagine the whole thing.
[00:20:13] And you can feed it elements, which it will refer to. So you can have a starter image, but you don't give it as the starting frame. You just give it as a reference, if you like. They call it an element. So that's the kind of mixture. And the workflow, as I say, it's all jumping backwards and forwards from Claude. And one other, just to complete it. I think you mentioned before we started recording today, Soul Cinema for locations. Is that right? Yes. Yeah, Soul Cinema for locations. That's right.
[00:20:41] So you've got within Higgsfield a bunch of their proprietary tools. Soul Cinema, Soul Location, some other Soul Cinema. Yeah. So there's a bunch of tools that Higgsfield have built themselves and some AI in there, which they build themselves, and all these other sort of third-party AIs, all available within Higgsfield, which is why I really like it. And guess what? You pay for it with a subscription. And I think that is a powerful and somewhat poetic moment to end on.
[00:21:06] And for anyone listening wanting to find out more about Bango or connect with yourself and learn more about how you created your films and have a conversation with you there. Where would you like me to point everyone listening? Yeah, please jump over to LinkedIn. Both Bango and myself, Giles-Tung, where you'll find me by searching on LinkedIn. And also Bango.com. We've mentioned a bunch of data today that is coming from a lot of our reports, which are all free to download on Bango.com. Awesome.
[00:21:35] I would have links to absolutely everything there. And I'd love to hear from anybody listening on the kind of things that you're creating and how you've done it. Because I think conversations like this can just enable anyone to create and get those ideas out of their head. It's really that the barrier has been lowered. As you said a moment ago, you can't draw or anything. I think I'm very similar as well on stuff like that. And to be able to get those ideas out of your head is something quite remarkable, especially seeing it all come to life like this.
[00:22:04] So I'll also include your video. I'll embed that in the show notes over at the podcast episode at Tech Talks Network. But more than anything, thank you for bringing all this to life today. Thanks for having me along, Neil. It's been a great fun. Thank you. I think Giles gives us two ways to think about the future. In subscriptions, value includes finding, paying for, managing and cancelling services without turning the monthly bank statement into a piece of detective work.
[00:22:32] And in creative AI, capability is becoming far less surprising. While control and exactness determine whether an experiment can actually become professional work. And I think his experience with Claude is incredibly useful. I know I'm going to be following in his footsteps and having a play there because the AI translated his idea into the language another model needed.
[00:22:59] While Giles remained responsible for the story and the direction throughout the entire process. So a big thank you to him for bringing streaming bundles, AI agents and a stubborn dragon all into one coherent conversation. That takes some doing. And you can follow Giles and Bango on LinkedIn and find Bango subscription research over at bango.com. But I'm going to put all this into a useful link section and I will add that to the end of the blog post.
[00:23:29] So go over to Tech Talks Network. Go to Consulting the Future Podcast. There'll be useful links at the bottom on this episode. And you'll be able to see everything. And I want to hear and speak with you. I want to see your experiments, what you create. I think there's something beautiful in that when you're not in the creative industry, but you can go out there and create an AI short film with your unique ideas. Yeah, I'm giving you a little bit of a job to do at the weekend and a little homework. And I also got a question for you all.
[00:23:57] Would you trust an AI agent to choose, activate and cancel your subscriptions for you? Or does that surrender too much control? As I said, I'm at techtalksnetwork.com. Please connect with me on LinkedIn, just at Neil C. Hughes. You can also connect with me on Instagram at Neil C. Hughes too. I want to hear from you. And if you go to Tech Talks Network, there's even a way you can send me a voicemail. So if the idea of typing everything out fills you with dread, send me a voice message. But I've taken up far too much of your time.
[00:24:27] Thanks for listening as always. Bye for now.