Demystifying Artificial Intelligence – An Executive’s Guide to AI


Kitae Kim

Principal Sales Engineer

As I mentioned in my previous blog, this time I will do a deeper dive into the differences between Deep Learning and Machine Learning, business use cases for Deep Learning, and how Apttus products leverage the technology.

What is Deep Learning? And what’s the difference between Deep Learning and Machine Learning?

Deep Learning is a subset of Machine Learning based on artificial neural networks, which are computing systems inspired by the structure and function of the human brain.

The differences between Deep Learning and Machine Learning are:

1. Deep Learning typically requires less involvement from humans. Hence, we can say that it can make intelligent decisions on its own

2. Its performance increases as the amount of data grows, whereas traditional Machine Learning methods’ performances can reach plateaus much earlier (even if a large data set is provided).

3. Deep Learning requires high-performance hardware and lots of labeled data to train but takes less time to make predictions.

Despite the benefits, we cannot say that it’s superior to traditional Machine Learning because there are still many situations where traditional Machine Learning provides more accurate results or are the best considering other factors such as data size, cost, time, etc.

Using Deep Learning in a Quote-to-Cash (QTC) process

Deep Learning shines when it comes to complex problems with a large amount of data to process. Nowadays, the most common uses for Deep Learning are image or text recognition and natural language processing (NLP) where it often outperforms traditional Machine Learning methods. Interestingly, natural language processing and image recognition are increasingly gaining importance in QTC process. Thus, it’s expected that more and more solutions using Deep Learning will be introduced imminently in the QTC space.

Currently, Apttus offers two products utilizing image recognition and natural language processing: Contract Lifecycle Management (CLM) and Virtual Assistant – Max.

CLM – Importing third-party paper with Machine Learning powered by KIRA

This functionality uses text recognition and natural language processing.

AI Flow

1. Ingest Agreement – This includes the process called Optical Character Recognition (OCR). It turns scanned hard copy legal or historic documents into texts, which can be used for data processing. Typically, the structured scanned documents can be converted using the traditional Machine Learning method, but with the help of Deep Learning, even handwriting and signatures in contracts or badly scanned documents can be recognized more accurately.

2. Identify & Extract – Recognizing clauses and fields requires natural language processing as analyzing and reviewing contracts need the ability to process words. Natural language contains lots of different information such as nuance, sentiment, etc. Additionally, there are many features to use for predictions such as the similarity between words, and repetition.

Bearing in mind of all these facts, Deep Learning works well compared to traditional machine learning methods.

Virtual Assistant – Max

Apttus Max also leverages natural language processing; understanding a user’s command is its first and foremost capability. As we have already seen, Deep Learning in Max extracts features such as the similarity between words in the sentences from a user and classifies the question(s) as one of the actions it should take. The actions include finding the information, cloning quotes, or creating agreements.

Accelerating digital transformation will bring more data into organizations, resulting in finding different business use cases for Deep Learning such as price or discount % recommendations, white space analysis, … etc.

What is right for me?

As you’ve seen in the blogs, there are different types of Artificial Intelligence, and depending on the problem you are trying to solve, the Artificial Intelligence technology to use can vary widely. Each technology has its own pros and cons; some AI technologies require more data and use ‘black box’ methods, but generally they provide more accurate predictions. On the other hand, others need less data and reveal which inputs contribute the most to outputs. Also, some require high-performance hardware, but others don’t. But the most important thing is whether an organization ha successfully deployed digital technologies as data is the key for Machine Learning and Deep Learning.

At Apttus, we provide a world-class, industry-leading Quote-to-Cash solution are experts in the AI space. We continuously support our customers and prospects with our solutions in their digital transformation and machine learning journeys.

Watch Our Webinar On Leveraging AI For The Quote-To-Cash Platform!

Leveraging AI For The Quote-To-Cash Platform

Trending Blogs