Technology

Three types of artificial intelligence (AI)

Thomas Davenport from MIT and Rajeev Ronanki has conducted a study of 152 AI projects. They divided AI into the following categories:

  1. Process automation through the use of robots
  2. Cognitive insight
  3. Cognitive engagement

Category 1: Process automation through the use of robots (bots)

In the study, this was the most common application of AI (47% of projects). This type of AI is becoming more advanced due to the emergence of more sophisticated code and integration/data capture from several IT systems. These result in advanced algorithms that can perform more intelligent tasks than a few years ago.

Examples of applications:

  • Categorize which customer inquiries (in CRM) are to be invoiced by reading through the contract and comparing it with the text in the customer’s e-mail
  • Put the right risk classification on the customer (in CRM) by retrieving updated credit information and comparing it with market news about the customer
  • Dispatch ordered goods by updating the eCommerce and CRM systems and managing customer communication

Category 1 is the simplest and cheapest form of AI. It is possible to secure both a quick and high return on investment.

Category 2: Cognitive insight AI projects that utilize machine learning was the second most common AI category in MIT’s survey (38%). The projects use algorithms to detect large amounts of data and then interpret what the patterns indicate.

In sales and marketing, the applications can be:

  • Predict what a customer will buy when she visits your online store
  • Design and publish digital ads tailored to your primary audiences
  • Prepare sales forecasts based on a weighting of information capture along the customer’s purchase journey
  • Detect real-time credit misuse
  • Change salespeople’s priorities by preparing lists of priority leads

Category 3: Cognitive engagement AI projects that use cognitive engagement use technologies such as chatbots, intelligent agents, and machine learning and accounted for 16% of MIT’s study projects.

In sales and marketing, the applications can be:

  • Intelligent agents who 24/7 answer potential customers’ questions about prices, functionality, delivery options and services in the customer’s language
  • Product recommendations based on dialogue with potential customers

Among these is the Swedish bank SEB. The bank has an intelligent agentĀ  that they test on customers.

Related Articles

Leave a Reply

Back to top button