We use cookies to ensure that we give you the best experience on our website.

AI for Sales, Buzzword or Reality and so What?

By Rahul Kapoor, Co-Founder, KlozeAI

Ok now that we have established that selling is going to exist at Zovo. Let’s chat a bit about what role AI could play in enterprise selling, what would be a natural fit?

And the question that’s been on every salesperson’s mind: Can AI help sales do anything of real value in 2018? Why are people making a big deal about a software that recognizes cats and wins jeopardy; what I do is far more sophisticated and quite unrelated?

Before we jump into the questions above, let’s get a few definitions out of the way. We have found that these simple definitions can instantly make you look good at dinners and cocktail parties:

  • Artificial intelligence (AI): algorithms exhibiting human-like intelligence / rational behaviour.

  • Machine learning (ML): algorithms that can learn from existing knowledge. ML is a set of algorithms that lead to AI.

  • Deep learning (DL): a type of learning algorithm that replicates how human brain learns. DL is a type of ML technique.

Now that we have a common understanding of some terms related to AI. Let’s first answer a context setting question before we get to the real question we raised above:

What is AI good at in 2017?

In a nutshell, AI in 2017 is generally good at solving problems where large sets of training data exist. To further clarify this let’s discuss some real-world examples across various categories where AI has some good traction:

Search, Natural Language Processing

  • Google Search has learned a lot from your search and click behaviour and gotten very good at recommending accurate search results. Underneath Google is a massive ML engine that is sifting through the massive amount of web data and marrying it with customer behaviour data to learn what satisfies a customer. Amazon uses the same principle to recommend products and Netflix movies.

  • Siri voice recognition is continually getting better at understanding what we are saying and taking action based on that, it is continually training itself on a large pool of accents and pronunciations from all Siri users.

Vision (Perception)

  • Tesla and Google are all working on cars that can drive themselves that are based on computer vision deep learning algorithms that are good at identifying things and taking appropriate action.

  • Harvard’s skin cancer detection technology is near 100% accurate in determining skin cancer based on deep learning algorithms.

Robotics

  • Boston Dynamics has robots that are now performing simple tasks of moving boxes. Other companies have developed robots for picking fruits and vegetables from fields.

Game Playing, Logic

  • AlphaGo defeats World Go champion Lee Se-dol in May 2017, sorry Lee.

  • IBM Watson defeated the human jeopardy champ in 2011.

  • DeepBlue defeated the chess champion in 1997. Clearly logic and game playing is an area that has seen significant advancement in the area of AI.

In summary, machines in 2017 are good at learning in situations where there exists a large volume of training data e.g. identifying cats in images when provided with millions of pictures of cats as training data. Now that we know a bit about the state of art of AI in 2017. Let’s address another related question that begged to be answered:

What are the big limitations of AI in 2018?

Without getting too technical here are the some of the shortcomings of AI today:

  • Using the past as a predictor of the future: Most AI/ML rely solely on prior data to make predictions for the future. AI/ML would not have predicted the movie Hangover to be a standout hit given the cast, storyline etc.? AI/Ml will most likely miss “black swan” events.

  • Causation remains a mystery: AI/ML is good at predicting relationships between things but can’t explain why certain things are related.

  • Common sense remains uncommon: AI/ML stays best at solving limited/narrow problems the notion of generic learning is still at bay.

Now with our newly acquired knowledge of what works and what doesn’t with AI in 2017. Let’s address the question we raised at the beginning of this long (but hopefully interesting post).

Can AI help sales do anything of real value in 2018?

I believe the answer to be yes but the way I see it the power of AI lies in augmenting human intelligence rather than replacing it. Let me explain. All experienced sales people know the world of selling is complex, unstructured and driven by relationships, emotions and personalities. In 2017 AI is not good at any of those things however, there is hope, AI is really good at solving narrow problems especially where prior large training datasets exists and sales is full of those scenarios e.g. which prospect should I call first, what is the most relevant information about a customer, what is the most relevant content for a customer, what is the best pricing for a customer etc.
In short, AI + HI (Human Intelligence) = Augmented Intelligence in sales is ready for prime time.
  • In my next post, I cover a new layer of applications that will exist to power sales of the future.

Signup for a free version of KlozeAI on

Signup Now
Request Demo

Related Posts