The Encord data development platform goes beyond business analytics to become “the world’s only multi-modal AI data development platform.”

On Thursday, the company announcement new multimodal data annotation capabilities for classification audio And documents — all in one interface. The update expands Encord’s existing support for medical, computer vision, and video data.

Also: I have tested many AI tools for work. These 4 actually help me get more done every day

Now, AI Chatbots And image generators are relatively common. But it’s much harder to generate compelling video or audio than it is to generate text. The AI ​​industry is increasingly focusing on multimodal capabilities, especially with the release of features like ChatGPT. Voice mode.

To refine an AI model, you need quality – and sometimes hyper-specific – data. Text data doesn’t provide the nuance these complex models need, and accuracy is even more important in high-stakes settings like medicine. Builders need platforms that can annotate and evaluate all kinds of data: video, audio, images, charts, reports, retail listings, PDFs, etc., ideally in one place. Several Encord customers use the medical image platform such as MRIs to develop better models to help doctors.

Having high-quality, well-annotated audio data allows you to create speech and emotion recognition models, and can even identify sounds. Video and audio AI products require increasingly sophisticated data support to achieve human-like realism, whether in terms of transcription accuracy or lip sync. For example, the AI ​​text-video synthesis platform Synthesis uses Encord to develop training models for its realistic AI avatars.

The Encord update includes new annotation and curation features for documents, audio, vision, and medical data. With multimodal annotation, AI teams can customize an interface to review and edit different file types side-by-side. Currently, different types of data are often siled across multiple services and platforms, adding time and cost to data annotation. Encord already supports key data annotation categories such as entity recognition, translation, summarization, text classification, and sentiment analysis.

“It is time-consuming and often impossible for teams to gain visibility into large-scale data sets throughout model development due to a lack of integration and consistent interface to unify these siled tools” , the company said in the press release.

Also: Organizations face growing pressure to accelerate their AI projects, despite lack of ROI

With Encord, AI teams can filter their data to identify and organize exactly what they need to build a model. Its assessment dashboard can also flag data that is hindering a model’s performance so teams can remove or replace it.

“On average, Encord customers use 35% smaller data sets, leading to models performing with 20% accuracy,” an Encord representative told ZDNET via email.

In a demonstration, Encord co-founder and president Ulrik Stig Hansen told ZDNET that he believes the company’s focus on quality and centralization will ultimately enable general artificial intelligence (AGI).