Today’s interview is with Deon Nicholas, Founder of Forethought, an AI company and 2018 TechCrunch Disrupt Battlefield participant, that has a vision to enable anyone to be a genius at their job. In practical terms, they do this indexing and surfacing content that matters the most in the context of their work. Deon joins me today to talk about the difference between NLP and NLU, winning last year’s TechCrunch Disrupt Battlefield and how Agatha aims to make it easy to be a “genius” at customer support.

This interview follows on from my recent interview – You can’t transform something you don’t understand – Interview with Annette Franz – and is number 318 in the series of interviews with authors and business leaders that are doing great things, providing valuable insights, helping businesses innovate and delivering great service and experience to both their customers and their employees.

Here’s the highlights of my chat with Deon:

• Forethought builds A.I. tools that index information within an organization and then deliver information to employees when they need it.
• Their first area of focus is the customer support agent.
• Their experience suggests that they can help agents and their teams triage and categorize tickets as well as surface answers up to 20 or 30 percent more efficiently.
• The origin of their approach comes form how Deon learned how to study history at school. You’ll have to listen to the podcast to learn more.
• Most information in an enterprise is scattered across the organization or is stuck in people’s brains.
• They recently won Techcrunch Disrupt, which is a bit like America’s Got Talent for the technology industry.
• They have developed two products right now:
  • 1. Agatha Answers, which is a tool that basically embeds into the agent’s workflow and helps them answer questions more efficiently, including suggesting answers for things based in the information that has been indexed.
    • It’s focus is about taking new agents and helping them operate like super agents.
  • 2. Agatha Predictions, which intelligently categorizes, prioritizes and routes tickets to the right teams and people saving a lot of admin time and attention.
    • That allows agents to spend more time actually answering questions rather than classifying them and they can then use Agatha Answers as as their companion to all that.

• They have taken a Natural Language Understanding (NLU) approach to building their AI platform rather than the much more widely used Natural Language Processing (NLP) approach.
• NLU is classified as a hard AI problem.
• NLP is largely the category of tools used to process the English language i.e. is this a noun or a verb etc?
• NLU is a step on from that which aims to effectively understand how humans understand and communicate in language. That allows it to do things like question answering. So, given a document and given a question, it can answer a question as it pertains to the document/statement/evidence?
• Example: consider the statement “John picked up the ball, John went to the store, John put the ball down” and then consider the question: “where is the ball?”
• NLU could establish that the ball is in the shop where NLP would not be able to do that.
• NLU is really all about how you solve human cognition problems from reading comprehension to question answering to information retrieval.
• NLU took some big advances in 2017 with the release of a few different data sets that were published around a question answerin, reading comprehension etc. And, those have allowed researchers to really dive deep into this.
• Whilst Agatha Answers can offer a 20 to 30 percent productivity uplift, Agatha Predictions could offer another 20 percent uplift.
• They are using machine learning (ML) to learn from past tickets that have been labeled and knowledge base articles, limiting the load on organizations to train the system or create new data.
• Client examples:
  • A local services marketplace is using the Agatha platform to improve their support operations across a range of use cases from routing tickets to the right agents to ensuring ongoing safety of operations.
  • A large storage space company is using Forethought’s products to help them detect whether a ticket needs to be escalated to a senior engineer. Escalations can cost them thousands of dollars per ticket so doing this in a fast and efficient manner is saving them a lot of money.
  • A large payroll and HR services provider is using Agatha answers to help their agents respond to questions more efficiently, particularly given the complexity of the questions that they normally get asked which involve complex tax and HR questions.

• Check out Forethought’s upcoming webinar: Bot Or Not?
• If you’re doing at least 1-2 thousand tickets a week or more and you have a support team of at least 10 to 15 people, you can start to experiment with different forms of AI in your contact center or in your customer service team.
• However, before you do that have a think about what you want to do and why.
• Deon’s Punk CX words: augment, not automate.
• Deon’s company that epitomizes a punk approach to CX: Amazon
• Did you know that relentless.com redirects to amazon.com. That tells you everything you need to know about Amazon, right?

About Deon

Deon Nicholas is the Founder of Forethought, the 2018 winners of TechCrunch Disrupt Battlefield SF. To overcome the information overload in enterprises, Forethought builds AI-powered products that embed relevant information into employees’ workflows, starting with Customer Support. Prior to starting Forethought, Deon built products and infrastructure at Facebook, Palantir, Dropbox, and Pure Storage. He has ML publications and infrastructure patents, was a World Finalist at the ACM International Collegiate Programming Contest, and was named to Forbes 30 under 30. Originally from Canada, Deon enjoys spending time with his wife and kids, playing basketball, and reading as many books as he can get his hands on.

Check out Forethought, say Hi to them and Deon on Twitter @forethought_ai and @dojiboy9 and feel free to connect with Deon on LinkedIn here.

..

..