OpenAI introduced Generative Pre-trained Transformer 3 (GPT-3) in 2020. It is an autoregressive language model that uses deep machine learning and NLP language models to generate text. This innovation has found a variety of use cases, especially in the NLP domain, such as copywriting, language translation, digital ad copies, chatbots etc. We list some of the commercial products that are built on GPT-3.

Algolia

Algolia is a search engine. When the OpenAI API launched, Algolia teamed up with the AI research lab to integrate GPT-3 with their advanced search technology to create their new ‘Answers’ product with better capabilities to understand the customers’ questions and provide answers to their questions. It also allows users to retrieve files, contacts, events or any object within the product through its APIs. 

After running tests of GPT-3 on 2.1 million news articles, Algolia saw 91 percent precision, and it was able to accurately answer complex natural language questions, four times more often than BERT.

Copysmith AI

Copysmith is an AI-powered content creation software. With Copysmith, the company wants to help people increase creativity and productivity. However, the company clarifies that the idea is not to replace human copywriters and writers.

“Using our unique combination of proprietary machine learning and OpenAI’s GPT-3 technology, we can provide you with the finest raw materials you need to smith your message,” says the company.  

Copy.ai

Copy.ai is a similar platform built on GPT-3. One can use it to write optimised blog posts in less time and produce digital ad copies on platforms like Google, Facebook, and Linkedin. It can also be used to create a website, sales copy and social media content.

The user has to upload their copywriting project (one can choose from emails, social posts, long-form blog posts) and then enter a few sentences about their brand and products. The  AI content generator gives the user multiple options for each campaign.Then Copy.ai’s editor can alsorewrite paragraphs. In the last step user would just have to copy and paste the output into CMS for publishing.

Replier.ai

Replier.ai uses GPT-3 to respond to customer reviews. The system automatically creates GPT-3 configurations for each individual business and business category. Replier.ai will automatically pick up the previous review responses and use those to create new ones. This is critical because that will allow the AI to always provide unique and custom responses that follow your unique business style and tone. 

It applies some pre-formatting of the input review that needs to be replied to, we run GPT-3, and then we clean the output – because GPT-3 is not always accurate. We built a pretty advanced AI monitoring system to understand what GPT-3 is generating for our customers, so we continue to improve and detect bad behaviours and add proper fixes.

Abtesting.ai

This platform allows the user to enter the desired URL and create a few variations of the title, copy and call to action. The company says that its AI can recognise the title, copy and CTA and automatically suggest variations.

One can integrate the company’s small javascript snippet in their webpage or install its WordPress or Wix plugin. “We create first experiment with some of the variations and run it using a multi-armed bandit approach,” the company’s description says. When the experiment reaches statistical significance, it considers the variations that worked best and uses an algorithm to mix them up and mutate them to run the next batch, as per the company.

Jasper AI

Jasper.ai is an AI writing assistant based on GPT-3. One can create original content with a high SEO ranking and content which is keyword-rich and plagiarism free. 

One can write and test more copy variations to increase sales and improve the return on ad spend. It helps to write in over 25 languages. Customers can scale up content marketing very fast and repurpose existing content to generate new results.

Quickchat

Quickchat is a multilingual conversational AI assistant powered by GPT-3. Quickchat chat bots can recognise and speak the user’s native language. They can automate customer support, online applications, searching through internal knowledge base etc. 

Users/businesses can also automate live chat conversations with the Quickchat AI widget. They can build AI Assistants for their clients or integrate them with proprietary software products. 

To be able to answer customer questions, chat bot needs access to information about the products. One can just upload FAQ’s, product descriptions, internal documentation and Quickchat takes care of the rest.  One can also connect Quickchat to the internal API, database or any other data structure to automate more complex processes using AI.