The pros and cons of integrating Amazon Transcribe with Nuxeo DAM

Looking to make your audio and video content more findable?

For clients with rich content libraries and a digital asset management system (DAM) like Nuxeo, our digital experience (DX)  and content services teams recommend considering an Amazon Transcribe integration.   

Amazon Transcribe uses state-of-the-art speech recognition to automatically — and accurately — convert spoken word to text. It’s a highly versatile transcription tool, with support for multiple languages and an abundance of great features like custom vocabulary and speaker identification. 

Integrating Amazon Transcribe with the Nuxeo AI add-on creates a user-friendly way to ensure your content is accessible and fully searchable. It offers your team the power to automatically generate searchable metadata, enhancing the overall findability of your video and audio assets.  

Of course, as with every integration, you’ll need to consider if the pros outweigh the cons and if it’s viable with your budget. Below, we take a look at what we feel are the top benefits of an Amazon Transcribe integration and offer some additional thoughts to help you decide if it’s right for you.  

 

The pros: It all adds up to greater operational efficiency  

1. Video and audio assets become searchable. You can choose to automatically introduce the WebVTT (.vtt) files generated by Amazon Transcribe into your Nuxeo DAM search index. This enables you to easily surface video and audio assets by theme, topic, or speaker. Your content and marketing teams can locate assets using keywords … or even a half-remembered soundbite. And with Nuxeo’s snippet feature, the search results become a hyperlink that sends you directly to the correct asset. Bonus: When assets are published to the web, these transcription files will be crawlable. That means improved SEO with no extra workload.
2.  Option to automatically create transcriptions on file ingestion. One of the other things we love is its flexibility. Configure your instance to create transcripts automatically on file ingestion into the DAM and you’ll have peace of mind that every single asset is searchable. Alternatively, you can configure for transcribe on request. The second option may be more suitable if your organization is looking to index only a few high-value video and audio files, or if budget is a concern. (See potential cons below.)
3. The transcriptions are highly accurate. Amazon Transcribe offers the opportunity to introduce custom language models and custom vocabulary that meet your organization’s specific needs.  People’s names, product names, acronyms, and technical terms can be recognized automatically and captured accurately in the transcript. And if you do need to make a correction? There are up to 10 alternates for each sentence, reducing the time your team spends on manual reviews and edits. 
4. Improved accessibility and compliance with subtitles. Any marketer will tell you that subtitles are a must-have today. They make your content accessible to the widest range of customers and improve overall engagement. Amazon Transcribe can generate subtitles from the transcripts, automating what has traditionally been a time-consuming and expensive manual process. And, even better, if you need support for livestreams or on demand, there’s just one service API that covers both. 

 

The potential cons: Associated development and operational costs

1. There is a charge per transcription. Usage is billed in tiered pricing per one-second increments. If your organization needs to hold substantially large amounts of lower-value video and audio content in the DAM, and your budget is limited, there might be the argument to opt for transcription on request instead of automatically transcribing everything. That said, it’s important to be aware that you may sacrifice some degree of findability and lose consistency in your metadata and your ability to surface assets.
2. You’ll need development support. To ensure a smooth integration with great end results, it’s worth investing in the support of a Java developer who understands Nuxeo. And, above all, someone who can partner well with your business users and IT team. In most cases, a basic integration is only a couple of weeks’ work. Configuring it with the search index or adding more automation will take additional time, complexity, and lift.  

  

Real-life success story: From static database to flexible, search-rich solution

iSoftStone’s DX team helped a large, internationally renowned 501(c)(3) organization integrate Amazon Transcribe with Nuxeo DAM. 

Our client’s marketing team had a rich back catalog of keynote speeches and other long-form video content, the majority of which it looks to maintain for posterity and/or leverage excerpts of for licensing. 

Prior to the integration, the team manually created transcriptions in Microsoft Word and uploaded them into a separate database with the video files. It was a bid to make the content of the videos more discoverable. Unfortunately, neither the Word documents nor the videos were searchable.  

Now, thanks to the support of iSoftStone’s Nuxeo experts, our client has a truly searchable and comprehensive record of the content it most wants to preserve and use. All the working hours spent crafting transcriptions by hand and struggling to find the right clip have been reclaimed.  

 

No Nuxeo? No problem 

We’ve chosen to focus here on Nuxeo DAM. However, Amazon Transcribe is DAM agnostic and our DX team has experience with ContentStack, Cloudinary, Tenovos, SAP Commerce Cloud, Shopify, and more. If you’re curious to find out if a transcription integration is right for you, feel free to reach out and talk with our experts.