Forum#deeptech
Read time: 06'54''
2 August 2023
Meet Papercup, the startup using AI dubbing to make your videos go global

Meet Papercup, the startup using AI dubbing to make your videos go global

As part of our quick fire questions series – or QFQs – we spoke to Jesse Shemen, CEO of Papercup about making the world’s videos watchable through the power of AI, building a diverse team and the value of support from industry angels.

Papercup was founded in answer to a problem that had frustrated me for years – that 99% of the world’s video content was trapped in a single language.

Traditional dubbing services alone can’t solve this problem because they’re prohibitively expensive and labour-intensive – two issues that continue to persist. Vast number of content creators and owners were unable to reach lucrative international markets as a result.

I’d been interested in the capabilities of text-to-speech tech for a while, dating back to the robotic voice on GPS systems, and saw there was an untapped opportunity for anyone seeking to bridge the gulf between content creators and their potential audiences worldwide.

After months of research and development, it became clear that the market demand lay in an efficient and scalable dubbing solution that didn’t entangle creators in the complexities surrounding localization or require vast budgets. My co-founder Jiameng Gao, who was an expert in the field, and I then got to work developing and packaging the offering for lifelike AI voices that could meet this demand. Papercup was born.

Tell me about the business

Papercup is an AI dubbing startup providing scalable localization to the media and creative industries. Using state-of-the-art machine learning, we create AI voices with all the warmth, intonation, and expressivity of human speech. Our team of expert translators checks the accuracy of the translation and can alter the tone, pronunciation and speaker style which means that we achieve high levels of translation accuracy and high-quality dubbed audio.

Ultimately, our aim is to make the world’s video content watchable in any language.

Our customers use our AI tools to help unlock as much value as possible from their videos – whether it’s about monetising back catalogues, or helping newly-launched channels perform overseas right out of the gate. Our ability to help generate new revenue streams with minimal investment has been a significant growth driver for many of our partners.

We work extensively with media companies specialising across an array of genres including factual, information, and lifestyle content. For example, we partnered with Sky News and Insider to help them reach over 500 million of new viewers in the last 12 months in Spanish-speaking markets. We also work with Bloomberg to dub their daily news video; content that due to its time-sensitivity requires quick turnaround times and therefore isn’t a natural fit for traditional dubbing.

Beyond news, we’ve expanded into other forms of unscripted content. We work with The Jamie Oliver Group to localise its YouTube content for Spanish, German, and Brazilian Portuguese-speaking audiences.  We’ve also helped Fremantle bring hit talent shows like American Idol and Britain’s Got Talent to the Middle East.

We’re also now working with a number of high-profile YouTube creators to dub their content for YouTube’s new multi-language audio feature that allows YouTube content creators to add dubbed audio on their main channel.

How has the business evolved since its launch?

We’ve grown from two co-founders to a bright team of 60. We’ve landed partnerships with some of the world’s biggest media companies, raised $30M across three funding rounds and expanded our language range to cover many of the world’s most spoken languages.

Last month, we were named as one of theWorld Economic Forum’s 2023 Technology Pioneers. A surreal endorsement considering alumni include Airbnb, Dropbox and Spotify. It’s an honour to see the brilliant work the team does every day being recognised on an international stage.

Tell us about the working culture at Papercup

We’re doing something new — making the world’s videos watchable through the power of AI. The path isn’t well trodden so it takes people who are comfortable forging a new path. We’ve assembled a team I simply could not be prouder of; a team that I genuinely love working with. They are some of the smartest people I’ve come across and have helped create the building blocks of our company, Papercup. It’s important to me that people feel valued for the work they do and also that we invest in not only their growth but in creating a working environment in which people can thrive.

To that end – we have a hybrid working policy so people can balance work and life, and we also spend a lot of time together socially: The Papercup football team (which I attempted to star in but failed as an American), company breakfasts every Tuesday (bagels are a favourite), drinks on Thursdays. We have a university every week given by a member of the team on everything from ‘The Oldest Light in the Universe’ to ‘The Backrooms’. Our weekly newsletter, PaperThoughts, is written by a different member of the team (and lately ChatGPT!) each week; these newsletters are always telling – we’re a diverse bunch, all from different backgrounds and everyone has something to offer that the rest of us can learn from.

How are you funded?

To date, the majority of our funding has come from investors. Last year’s $20M Series A was a milestone for us in this respect. Along with the increasing revenue we’re deriving from our partnerships as a result, it’s helping us plan for and execute against significant long-term objectives.

We’re honoured to have an oversubscribed round led by Octopus Ventures and had follow-on participation from venture capitals such as Local Globe, Sands Capital, Sky and Guardian Media Ventures, Entrepreneur First and BDMI.

In terms of specific investors, Series A saw us attract several new angels, notably Des Traynor, Co-founder of Intercom, and John Collison, Co-founder of Stripe. They joined an impressive roster of other big names who have backed us from the start, including William Tunstall-Pedoe, founder of Evi (now Amazon’s Alexa), as well as Zoubin Ghahramani, Senior Research Director at Google Brain and former Chief Scientist at Uber. It’s a privilege to be able to draw on their support and expertise while we grow the company.

What has been your biggest challenge so far and how have you overcome this?

Before Papercup, localization options were either subtitling (cost-effective but far less engaging), studio dubbing (high quality but costly and time-consuming) or piecemeal AI tools (poor quality and resource required to manage the various tools).

Given the relative infancy of the AI dubbing category and the limitations of traditional dubbing services, a major and ongoing challenge has been educating the market on the time and cost gains of our technology, but also the quality dubbing it delivers. We know that once people hear our voices, they’ll agree that the quality is way above and beyond their expectations, but it’s a whole process getting to that point.

Our ongoing challenge, or maybe it’s better described as a sub-mission, is continually improving the quality of the voices we create, so we can tackle even more expressive content.

How does Papercup answer an unmet need?

As said, billions of hours of video content are trapped in a single language, which limits both content owners’ ability to reach broader global audiences, and viewers’ access to diverse and quality content.

We know that subtitling doesn’t quite have the same appeal for viewers. Some of the videos we’ve dubbed on YouTube have received 28x times the amount of views as their subtitled equivalents – that’s a staggering figure.

We also know that in certain markets, there’s a clear preference among some audiences for dubbed content. Mexico is a notable example, where 71% of viewers who currently watch just under two hours of foreign-language videos per week, prefer dubbing over subtitles.

What’s in store for the future?

It’s a good question – and one everyone wants to know when it comes to generative AI. While the ultimate long-term vision is to make any video watchable in any language, we are always trying to assess the short-to-medium-term steps that will help us make significant strides in this space.

In terms of groups of content owners, we’re currently expanding our work with creators. YouTube has recently launched its multi-language audio feature, which allows creators and brands to add dubbed audio tracks to their main channel, so audiences can watch in their preferred language. We’re working with a number of high profile creators to dub their channels.

On the digital media side, while factual programming like news and documentaries remains a key focus, we are constantly developing our models to deal with more complex content types, such as drama and entertainment.

And all this, of course, involves expanding the number of languages that we’re equipped to dub material into. Right now we cover many of the most-spoken languages in the world, but we’re always expanding our offering as global content markets continue to evolve.

What one piece of advice would you give other founders or future founders?

On a practical level, always remember that nothing goes to plan – and that’s okay.

You have to expect chaos, embrace the uncertainty and don’t feel like you’re an imposter because you don’t have the right answers. It’s the people who continue ploughing through that end up building something special that has real longevity.

And finally, a more personal question! What’s your daily routine and the rules you’re living by at the moment?

For me it comes down to family, exercise, planning and execution.

I start off my day with my three beautiful kids, which is always humbling and just outright fun.

Next is daily exercise to clear the mind and stay physically fit. I’m thankful my wife helped me establish a daily routine because it replenishes your energy and gives you a sense of confidence that is critical when facing a day of challenges.

First step at work is planning – what are my unequivocal three priorities and how can I move the needle on them? I then use my productivity app TicTic to set my 5 big tasks and up to 10 small, sub 5 minute tasks.

Then it’s time to execute!

Jesse Shemen is the CEO of Papercup.