AI has become a “buzzword” of sorts.
On one hand people are afraid it will take their jobs, on the other people believe it can magically help them make more money, or contribute to societal progress in the form of curing cancer or custom tailoring education to children. These questions will not be addressed in this blog post, but we will examine at a basic level what AI is and how Spark Talk is using it to improve lead generation.
At a very basic level, software is designed to take input and produce an output. This may be as simple as you clicking on a button on a website – the input is your click, the output takes you to a new page or submits a form. This may be as complex as multiple sensors on a self-driving car taking in immense amounts of input/data and deciding the car should stop because the car is stopping ahead. AI or “Machine Learning” is the process of taking inputs and allowing the machine to interpret the data better over time. There are multiple methods in which this can be done, but typically it involves statistics and categorization. For example, let’s say you watch several romantic comedies on Netflix. Netflix has a set of algorithms that will determine, statistically, that you like romantic comedies with a high degree of confidence. Therefore, Netflix will suggest romantic comedies for you on the home screen.
At Spark Talk we use AI in two ways:
First, we use Natural Language Processing (NLP) to interpret responses from leads. As you interact with the leads and “close” them, or move them to “not interested” we use AI to update our NLP models to learn which responses produce positive results, and negative results.
Second, we use AI to help determine which types of people are interested in your product or service. For example, let’s say you sell medical devices to doctors. Our AI will begin to identify specific characteristics that increase or decrease the likelihood of interest.
For instance, we might be able to tell with 92% confidence that orthopedic surgeons are more interested than the average doctor. We might discover that hospitals with over 500 staff are way more likely to buy your product than smaller hospitals with less than 500 staff. You may think your intuition can do these calculations, and that is certainly true in some cases. However, there have been a number of instances where we discovered that companies were actually wrong in their target audience assessment.
This is one of many reasons our system dramatically outperforms entry level sales employees. And…our system will keep getting smarter and smarter over time.