How We’re Using AI at Austin Data Labs

How Austin Data Labs is using AI

AI in our platform

Our data science platform has always had a proprietary machine learning AI component. This predictive AI capability is available as a module to clients with a sufficient level of complexity in their end-to-end supply chain.

AI on our website

We use AI to help humans translate, code, and optimize the website through augmentation of human effort. Contrary to the current hype cycle, large language model AI, while useful, is far from error free or independently intelligent, and needs the human touch.

AI in our content

Large language model AI increases human productivity immensely. We utilize it as a tool for the tedious bits of content creation and tracking (such as creating outlines – a task LLM AI is well suited for). Once we have our AI starter kit tuned the way we like it for a piece of content (utilizing a complex proprietary prompt library) we bring in the humans to add their unique industry expertise, tone of voice, fact check, QA check translations, and more.

AI in our analytics

We’re augmenting Google Analytics with AI-based analytics, user heat mapping, and SEO tools. As with all of our AI use, we make sure a human is involved in the process.

AI for accessibility

We’re still working out the kinks with this one. The goal is to use AI to help us create a website that works for everyone, using AI to catch areas we may have missed, and to continuously optimize the site for visitors of varying ability and access levels. We hope to complete this implementation by Q3 2023.

Where we’re not using AI

We’re not currently utilizing AI for images – we have concerns about the way AI models are currently trained, in part, on the content of artists who may or may not have consented to having their art used by the public in this manner.

We’re not uploading proprietary or confidential data into any third party AI model like chatGPT. We take our customer’s privacy and intellectual property protections seriously.

See an AI-generated error on our site?

Let us know!

You cannot copy content of this page