Skip to content
Contact us

Demystifying AI: Understanding Its Types and Applications

Artificial Intelligence (AI) has become a buzzword in today’s digital age. Many companies are eager to implement AI without fully understanding what it is or the business problem they’re trying to solve. This post aims to reset the conversation and delve into what AI is, its different types, and how it can be applied to solve business problems.
Understanding AI
AI can be broken down into three distinct categories: automation, predictive, and generative.
  1. Automation: This involves performing tasks that would normally be done manually by a person, without any manual intervention.
  2. Predictive: This involves using data to predict what’s next based on past results and existing data. It’s important to note that predictive AI is not creating something new; it’s merely analyzing all the available data to determine the most likely outcome.
  3. Generative: This is where new content is created, usually sourced from some sort of existing content. The content could vary from images to text, structured data, or knowledge bases. The key aspect of generative AI is that it starts from some point of existence and then utilizes algorithms to create something new.
Delving Deeper into Each Category
To better understand these categories, let’s explore some examples:
Automation
  • Processing email attachments: This could be useful in a finance department where you receive an email from a customer every month with an attached invoice. The attachment could be downloaded and uploaded to bill.com or Quicken automatically, a task that many finance departments are currently doing manually.
  • Monitoring application logs for conditional responses: If you start seeing a certain error message with a specific frequency, it could mean that you need to recycle your Redis cache in a web application. This is something that could be fully automated.
  • Applying brakes in self-driving or assisted vehicles: Many vehicles now have sensors that can detect when you are approaching another vehicle at too high a rate of speed and will auto-apply your brakes. This is another version of AI that falls under the automation category.
Predictive
  • Chatbots: Whether it’s on your websites or within applications, chatbots typically look at support desk history and product documentation to provide the customer with what it believes to be the right solution to help.
  • Fraud detection: If you’re traveling and using your card somewhere you’re typically not, you might get flagged and receive a text message asking if the transaction was really yours. This is an example of predictive AI using existing data to predict possible fraud.
Generative
  • Email generation: Almost all responses to clients can initially start with being generated through AI. The AI can summarize the last 20 emails and generate a response in a certain tone, getting a certain point across.
  • Code generation: Tools like GitHub Copilot are starting to appear that can generate code.
  • Image creation: AI can generate images based on a certain genre because it already has a whole library of images that have been tagged appropriately.
Applying AI to Business Problems
Once you understand what AI is and its different categories, the next step is to look at your business problems and start figuring out which category is the right solution. Whether you’re going to use automation, predictive, or generative AI will depend on your specific business case.
For example, if your business problem is wanting to make your customer success organization more efficient, then a chatbot might be the right solution. You could start with a supervised model that sits side by side with the customer success agents, giving them possible prompted answers. The ultimate goal would be to weed out the wrong answers, highlight the correct answers, and start moving towards a higher confidence level, eventually moving to a semi-supervised model and then an unsupervised model.
In conclusion, understanding what AI is and how it can be applied to solve business problems is crucial before implementing it. Whether you’re going to build it or buy something that’s already in the marketplace is something that can be figured out as part of the overall process. Remember, AI is a tool that can help you solve business problems, not just a flashy object to chase after.