Innovation Snapshot
Location:Edmonton, Alberta

Endless ideas at the click of a button

For entrepreneurs and small business owners, copywriting can be the most important—and time-consuming task—they face. Product descriptions and ads are how they communicate directly with their customers, but quality copy isn’t easy to create. Copysmith aims to make that task easier, by using machine learning to cut down on the time needed to write engaging articles, social media posts and ads.

Copysmith was borne out of a hurdle faced by many entrepreneurs and start-ups; co-founder Jasmine Wang was frustrated by the amount of time she spent writing marketing materials for her eCommerce store. As a researcher in machine learning, she knew it was a problem that artificial intelligence could help with.

Copysmith’s process takes advantage of a Generative Pre-trained Transformer language model, which allows it to generate writing that is difficult to distinguish from copy written by humans. It isn’t just a formula, however, that keeps spitting out similar content — Copysmith’s model learns, adapting based on what information the user gives it and what results they are happiest with. Over time, it takes those lessons to create content better tailored towards the user’s particular style.

Copysmith’s process builds its content one word at a time, rather than using pre-stocked phrases and sentiments. This ensures all the content made is unique and avoids plagiarism. The company currently offers over 30 templates, ranging from product descriptions to press releases and is currently working on adding new languages options.

Learn more!