How Twofour streamlined unscripted production with AI + expert proofreading
Twofour is an award-winning television production company known for its innovative reality and factual entertainment programs. One of its standout projects is I Kissed a Boy (and its follow-up, I Kissed a Girl), the UK’s first gay dating show, which airs on BBC Three and iPlayer.
A key element of the show is the "tea room" interviews, where contestants privately reflect on their experiences. These unscripted, first-person narratives are crucial for storytelling, as they allow producers to search for and pull key moments into the edit quickly. However, to make this process work, Twofour needed precise, word-for-word transcriptions; a task that had previously caused significant delays in their post-production workflow.
Juggling high volumes of unscripted content and diverse accents
In order to carry out the task, Twofour tried multiple transcription approaches, each with major drawbacks:
- On-location live loggers: the team initially attempted to have loggers document interviews in real time, but weak internet connectivity made live updating unworkable.
- Manual transcription in post-production: they then hired loggers to manually transcribe interviews after filming, but this was painstakingly slow. They had to manually insert time codes every 10 seconds, making the process excruciatingly time-consuming.
- Seeking a faster solution, they tested other platforms’ AI transcription service. However, the accuracy was so poor, especially for regional accents, that the transcripts were almost unusable.
We looked back at some of the transcripts, and they were completely wrong. It was talking about something totally unrelated to what the contributor was speaking about, which rendered it useless."
With only 15 days of filming and an incredibly tight post-production timeline, Twofour needed a faster and more accurate transcription method that could handle diverse accents and unscripted dialogue while easily integrating into their existing workflow.
HappyScribe’s AI transcription + proofreading service
After testing HappyScribe’s AI transcription during Season 1, Twofour saw a clear improvement in accuracy over the previous provider. For Season 2, they decided to refine the process further by incorporating HappyScribe’s Expert Proofreading service, where professional linguists review AI-generated transcripts for maximum accuracy.
This approach led to three key improvements:
- Massive time savings: from manual logging to AI + Expert Proofreading, Twofour cut months off their transcription process. "Before, we were transcribing from October to January. With HappyScribe, we had everything done by the end of October," Nicholls explained.
- Better handling of regional accents: unlike other AI-powered platforms they tried, which struggled with various UK dialects, HappyScribe delivered far more reliable transcripts. "Even the AI alone was much better than the previous provider. But with Expert Proofreading, the accuracy was so high that errors weren’t a concern anymore," she added.
- A seamless workflow with easy file exports: HappyScribe’s ability to export transcripts in multiple formats (Excel, CSV, etc.) made it easy for Twofour’s editors to organize and search interview content efficiently.
"We found the right balance between speed and accuracy. With HappyScribe’s AI + Expert Proofreading, we now get near-perfect accuracy, and the process is so much quicker."
Faster workflows, reduced costs, and improved accuracy
HappyScribe completely transformed Twofour’s transcription process for I Kissed a Boy, making their post-production significantly more efficient.
Before using HappyScribe, the team relied on manual loggers, which meant transcribing interviews dragged on for months, often requiring contract extensions and additional costs. With HappyScribe, the team eliminated these bottlenecks and saved both time and money.
"Last year, we were still transcribing in January. With HappyScribe, we had everything sorted by October. The speed difference is huge."
Beyond just speed, accuracy was crucial. HappyScribe’s high accuracy rate meant that producers could search transcripts with confidence, ensuring key soundbites were captured and used effectively in storytelling.
The flexibility of file exports was another game-changer. Twofour’s editors could instantly locate key moments by organizing transcripts into a shared Excel document with individual tabs for each contestant. This made it incredibly easy to pull relevant dialogue into the edit.
"The fact that we can export as an Excel, CSV, and multiple other formats is amazing. It fits right into our workflow and makes everything so much more efficient."
Expanding HappyScribe’s use across productions
Following the success of I Kissed a Boy, Twofour has recommended HappyScribe to other productions, including Destination X and Educating Yorkshire.
"We’re so happy with HappyScribe that we’re actively recommending it to other shows. It’s the best balance of speed and accuracy we’ve found."
Results in a nutshell:
For Twofour, HappyScribe has eliminated transcription bottlenecks, improved accuracy, and saved both time and money, all while ensuring their editors can quickly find and use key soundbites to shape compelling stories.
- Transcription-time reduced by months: before HappyScribe, transcriptions took from October to January. Now, they are completed by October.
- Higher accuracy, even with accents: HappyScribe delivers high levels of accuracy, significantly outperforming previous AI tools.
- Faster search and organization: transcripts are easily exported into Excel, allowing editors to quickly locate key soundbites from across the whole series.
- Cost savings: eliminating the need for extended manual logging contracts helped reduce overall production costs.
- Improved efficiency across productions: Twofour has already recommended HappyScribe to multiple other shows, expanding its use across the industry.
"HappyScribe has completely streamlined our transcription process. It’s faster, more accurate, and far more efficient than anything we used before."
Cristina G
Cristina G is a former journalist who transitioned into PR and Communications before diving into growth marketing and ultimately product marketing. She loves learning and writing about how HappyScribe customers use the product to achieve their goals.