Enrich your data with our range of human-annotation services at scale.
It’s simply not enough to give a computer a large amount of data and expect it to learn – data has to undergo preparation for computers to find patterns and inferences within it. That’s where we come in. We preprocess data to make it usable for machine learning. “Annotation” refers to any metadata tag used to mark up elements of a dataset. Adding meaningful metadata to the original dataset provides a layer of rich information to support machine learning.
Human-powered data annotation
Access the expertise of qualified annotators with our community of 1 million+ AI experts
We can quickly process hundreds of thousands of data rows so your models get the information they need to work in the real world. We harness the intelligence, skills, and cultural knowledge from our global community of contributors to create the highest quality data.
AI training platform
Our proprietary AI training data platform is capable of processing all kinds of text, image, audio, video and geo-local data with the power of our community members
Our quality assurance system features built-in validation, spot-checking and a workers seniority system to ensure the highest quality data to train machine learning applications.
Need things done in a specific way under strict guidelines? We can work with you to make sure our team of experts gets your project done within your requirements and timeline.
Data types for all of your machine learning needs
Our global AI Community, along with our TELUS International AI training platform, provide high-quality data annotation services for your machine learning use cases. Together, our people and our technology enrich and process millions of custom data points across text, image, audio, video and geo datasets.
With a background in natural language and linguistics, we are well equipped to handle any kind of text annotation project. Whether you’re training an entity extraction system or sentiment analysis tool, our curated community can accurately label text data in 500+ languages and dialects.
We provide text classification and text annotation services in addition to labeling entities within parts of speech.
Data annotation and labeling services
Our extensive data annotation services, alongside our global AI Community, will improve your machine learning models to create better AI systems.
Using human annotators, we can quickly classify your data into predetermined categories across any data type. Notably, this includes audio / sound classification to support NLP applications like chatbots, automatic speech recognition, text-to-speech and more, as well as text classification to analyze and tag content while understanding the subject, recognizing the intent, and analyzing the sentiment within it.
Available across multiple data types, our transcription services drive more value from your existing data. For example, our team can digitize text that is pictured in an image such as a photo of a receipt or handwriting. This then supports optical character recognition (OCR) models. Likewise, our video transcription services convert what is spoken on video into written text for subtitling or captioning. This makes online videos more searchable and accessible because it provides a better UX and boosts SEO.
Entity annotation and linking services
A named entity embodies a single distinct concept, often with a proper name, like “Japan” or “Elvis Presley”. However, the number of ways to classify entities is almost endless. We’ve spent years building solutions that can locate and label entities in raw text data. Likewise, entity linking involves locating and disambiguating named entities in a text through the use of a knowledge base. The purpose of entity linking is to add metadata that not only identifies the appearance of a named entity in the text, but also which specific entity it is.
Tying entities to a unique entry in a knowledge base is an important step in the construction of many machine learning text algorithms.
Sentiment analysis services
Also known as opinion mining or emotion AI, sentiment analysis determines whether a text is positive, negative or neutral by extracting particular words or phrases. Sentiment analysis provides helpful insights that drive effective business strategy. Language is often vague or highly contextual, making it very difficult for a machine to understand without human help. As such, human annotated data is essential when training a machine learning platform to analyze sentiment.
Accurate localization is essential if you want to achieve success in the global market. At TELUS International, we combine the expertise of our AI Community with intuitive tech to solve even the most complex translation needs. From providing parallel text corpus to localizing software to evaluating your machine translation quality, our network has helped the world’s largest companies globalize their business.
Data annotation improves conversational chatbots
Even today, chatbots used for customer service often lack a human touch. Why? Because the unstructured nature of conversation is particularly difficult for machine learning. Our client wanted a conversational bot to engage its users. This required finding a partner that could deliver a wide range of text data reflective of natural conversation and accurately labeled for emotional intent. TELUS International was selected to create this AI training data. The results:
- 5,000+ Q&A text sample
- Each set labeled for emotional intent
Diverse global AI Community of annotators and linguists
Data annotation languages and dialects
Locales covered across the globe
Secure onsite global delivery centers if required
AI starts with data: Facing the challenges of data collection & annotation
Discover useful insights into the challenges of data preparation to ensure that your next artificial intelligence project is a success.
Upgrade your AI
Partner with our AI Data Solutions experts to customize the exact project to advance your machine learning needs.