By Steffen Huß
Integrating AI into your company can be daunting. This is because AI helps in very different ways, each with its own set of risks, potentials, and demands for your organization. Therefore, understanding the following four types of AI wins can provide guidance and help you make more informed decisions.
This post will give an overview on what the 4 types of AI wins are. To go into the details, we will explore of these 4 AI wins in a separate blog post, including their specific risks and challenges in implementation:
- The Free Lunch: The easiest type of AI win. You just do a well-known task with AI support and get better or quicker results immediately. No relevant upfront effort involved. One example are complex Google researches that you speed up with tools like Perplexity.ai. Learn more about Free Lunches in this post.
- The Investment: Here you need to invest first before you reap the benefits. An example is implementing an AI that has access to your document and/or fine-tunes on them, like it seems to be the case for Harvey. Investments need to be tackled very differently than Free Lunches. Learn more in this post.
- The Piggybank: Often time savings with AI are small individually, but add up over time. In this case it is key to include these small AI wins in automated processes to minimise overhead and allow for scaling. An example is extracting data points from a document in the background before giving it to a lawyer for review. Learn more in this post.
- The Quality Improver: Not all AI wins speed up performing tasks. Some of them might rather reduce risks or improve your client’s experience. An example is a compliance check in the background as an additional layer of safety (blog post to follow soon).
These four types of AI wins help me tremendously to get orientation when discussing customers’ use cases. And this is even the case when more than one type is applicable, as it provides some structure to think about the peculiarities of the case. I hope you will benefit from the framework, too!
