Artificial Intelligence in the real world – Harvard Business review

Question 1

Artificial Intelligence in the real world overview

Answer 1

The Problem: Cognitive technologies are increasingly being used to solve business problems but many of the most ambitious AI projects encounter setbacks or fail.

The Approach: Companies should take an incremental rather than a transformative approach and focus on augmenting rather than replacing human capabilities.

The Process: To get the most out of AI, firms must understand which technologies perform what types of tasks, create a prioritized portfolio of projects based on business needs and develop plans to scale up across the company.

Question 2

Three types of AI:

Answer 2

Process Automation, Cognitive Insight, Cognitive Engagement

Question 3

Three types of AI:
Process Automation

Answer 3

• Most common type is automation of digital and physical tasks.
  o Using robotic process automation technologies (RPA)
   Robot = code on a server
• Acts like a human inputting and consuming information from multiple IT systems
• Tasks are:
  o Transferring data from email and call center systems into systems of record
  o Updating records and handle customer communications
  o Reconciling failures to charge for services across billing systems
  o Reading legal and contractual documents to extract provisions using natural language processing.
• RPA is the least expensive, easiest to implement, quick and high return investment
• Well suited for working across multiple back-end systems
• Used for outsourced business processes
  o If you can outsource the task, you can probably automate it

Question 4

Three types of AI: Cognitive Insight:

Answer 4

• Second most common type of project
  o Using algorithms to detect patterns in vast volumes of data and interpret their meaning.
• Tasks are:
  o Predict what a particular customer is likely to buy.
  o Identify credit fraud in real time and detect insurance claims fraud.
  o Analyze warranty data to identify safety or quality problems in automobiles and other manufactured products.
  o Automate personalized targeting of digital ads.
  o Provide insurers with more accurate and detailed actuarial modeling.
• Difference between cognitive vs traditional analytics:
  o Much more data-intensive and detailed
   Better models
   Better predictions and putting data into categories improve over time.
  o Deep learning
   Moved data curation (process of creating, organizing, and maintaining data sets) from labor intensive to a machine identifying probabilistic matches.
   Can eliminate redundancies.

Question 5

Three types of AI: Cognitive Engagement:

Answer 5

• Engages employees and customers to use natural language processing chatbots.
• Tasks are:
  o Intelligent agents that offer 24/7 customer service
   Password requests
   Technical support questions
  o Internal sites for answering employee questions
   Topics about IT
   Employee benefits
   HR policy
  o Product and service recommendation systems for retailers that engage in personalization.
  o Health treatment recommendation systems that help providers create customized care plans.
• Goal is to handle growing numbers of employee and customer interactions without adding staff.

Question 6

Framework for integrating AI:

Answer 6

1. Understanding the technologies 2. Creating a portfolio of projects 3. Launching Pilots 4. Scaling Up

Question 7

Framework for integrating AI: 1. Understanding the technologies

Answer 7

• Which technologies perform what types of tasks; what are the strengths and weaknesses of each?
• Best to leverage the capabilities of employees (need statistical and big-data skills) to learn the basics of these technologies.
• Either have data scientists or analytics in house or create an ecosystem of external service providers.
• Make experts available to high priority projects throughout the organization; dictate groups to particular business functions.

Question 8

Framework for integrating AI: 2. Creating a portfolio of projects

Answer 8

• Systematically evaluating needs and capabilities to design a portfolio of projects.
  o Usually evaluated based of workshops.
• Identifying the opportunities:
  o Which areas of a firm would benefit the most from AI?
   Usually, data intensive departments where knowledge is at premium but not available
• Bottlenecks: knowledge exists but is not optimally distributed (information distributed among departments)
• Scaling Challenges: knowledge exists but the process of using it takes too long or is too expensive.
• Inadequate Firepower: companies may collect more data than its existing human or computer firepower can analyze and apply.
• Determining the use cases:
  o Evaluate the use cases in which cognitive application would generate substantial value and contribute to business success.
   How critical to your overall strategy is addressing the targeted problem?
   How difficult would it be to implement the proposed IT solution?
   Would the benefits of the implementation be worth the effort?
   Differentiate between short- and long-term value.
• Selecting the technology:
  o Examine whether the AI tools being the considered for each use case are truly up to task.
  o Are they slowing down any more complex production systems?
  o Today it is still wise to use incremental steps while planning for transformational change. For now, keep AI internal and prepare it for the ultimate goal.

Question 9

Framework for integrating AI: 3. Launching Pilots

Answer 9

• Create pilot projects for cognitive applications before rolling them out across the entire enterprise.
  o Proof-of-concept pilots (proves if concepts work piece by piece)
  o Avoid “injections” of projects by senior executives who are influenced by vendors.
  o In case of several pilot projects at the same time, create a cognitive center of excellence.
• Business-process redesign:
  o Focus on how redesign impacts the division of labor between humans and the AI.
  o Systematic redesign of workflows is necessary to ensure that humans and machines augment each other’s strengths and compensate for weaknesses.

Question 10

Framework for integrating AI: 4. Scaling Up

Answer 10

• Requires a collaboration between technology experts and owners of a business process.
• System has to be integrated into existing systems and processes.
  o Check whether it’s feasible.

Question 11

The future Cognitive Company:

Answer 11

• Companies already working with AI will find themselves as well positioned to reap benefits in the future.
• Professional services will become more valuable and less expensive.
• Goal is to free up humans to become more creative and productive.
• AI will not take over jobs but only processes.
  o Humans and AI will work together in the future.