AI, ANALYTICS, AND THE NEW MACHINE AGE Flashcards
Three types of AI
A. Process automation= are “robots” (that is, code on a server) act like a human inputting and consuming information from multiple IT systems. —> RPA is the least expensive and easiest to implement of the cognitive technologies and typically brings a quick and high return on investment. The tasks include:
* Transferring data from e-mail and call center systems into systems of record
* Replacing lost credit or ATM cards
* Reconciling failures to charge for services across billing systems
* Reading legal and contractual documents to ex- tract provisions using natural language processing.
B. Cognitive insight= second most common type of project. It uses algorithms to detect patterns in vast volumes of data and interpret their meaning. —> models typically are trained on some part of the data set, and the models get better—that is, their ability to use new data to make predictions or put things into categories improves over time.
Typical applications:
* Predict what a particular customer is likely to buy
* Identify credit fraud in real time and detect insurance claims fraud
* Analyze warranty data to identify safety or quality problems in automobiles and other manufactured products
* Automate personalized targeting of digital ads and provide insurers with more-accurate and detailed actuarial modeling
C. Cognitive engagement= projects that engage employees and customers using natural language processing chatbots, intelligent agents, and machine learning
* Companies in the study tended to use cognitive engagement technologies more to interact with employees than with customers.
* Companies tend to take a conservative approach to customer-facing cognitive engagement technologies largely because of their immaturity. —> ex. Facebook: messages chat robots couldn’t answer 70%
4-Step framework for integrating AI
- Understanding the technologies: companies must understand which technologies perform what types of tasks, and the strengths and limitations of each —> Companies will need to leverage the capabilities of key employees, such as data scientists, who have the statistical and big-data skills necessary to learn the nuts and bolts of these technologies.
- Creating a portfolio of projects: Evaluate needs and capabilities and then develop a prioritised portfolio of projects. This is usually done in workshops or small consulting engagements.
Companies conduct assessments in 3 broad areas:
A. Identifying opportunities= which areas of the business could benefit most from cognitive applications
* Bottlenecks= the lack of cognitive insights is caused by a bottleneck in the flow of information; knowledge exists in the organisation, but it is not optimally distributed
* Scaling challenges= sometimes knowledge exists, but the process for using it takes too long or is expensive to scale
* Inadequate firepower= a company may collect more data than its existing human or computer firepower can adequately analyse and apply
B. Determining the use cases= How critical to your overall strategy is addressing the targeted problem? How difficult would it be to implement the proposed AI solution—both
technically and organisationally? —> prioritise the use cases according to which offer the most short- and long-term value
C. Selecting the technology= whether the AI tools being considered for each use case are truly up to the task. Chatbots and intelligent agents, for example, may frustrate some companies because most of them can’t yet match human problem solving beyond simple scripted case
3. Launching pilots: Because the gap between current and desired AI capabilities is not always obvious, companies should create pilot projects for cognitive applications before rolling them out across the entire enterprise —> Proof-of-concept pilots are particularly suited to initiatives that have high potential business value or allow the organisation to test different technologies at the same time.
Systematic redesign of workflows is necessary to ensure that humans and machines augment each other’s strengths and compensate for weaknesses.
4. Scaling up: Many organisations have successfully launched cognitive pilots, but they haven’t had as much success rolling them out organization-wide. —> need detailed plans for scaling up, which requires collaboration between technology experts and owners of the business process being automated.
7 Questions on how sustainable competitive advantage from AI is:
- How much value is added by customer data relative to the stand-alone value of the offering? The higher the value added, the greater the chance that it will create a lasting edge. Ex. car manufacturers, which test them extensively before incorporating them into their products. It’s crucial for the systems to be fail-safe, and the testing data is essential to improving their accuracy. Whereas in the smart television market, the value of learning from consumers is relatively low
- How quickly does the marginal value of data-enabled learning drop off? how soon does the company reach a point where additional customer data no longer enhances the value of an offering?
* The more slowly the marginal value decreases, the stronger the barrier is. Ex. Smart thermostats products need only a few days to learn users’ temperature preferences throughout the day.—> In this context data- enabled learning can’t provide much competitive advantage
* Whereas when the marginal value of learning from customer data remains high even after a very large customer base has been acquired, products and services tend to have significant competitive advantages. (ex. Systems that predict rare diseases and only search engines as Google) - How fast does the relevance of user data depreciate? If the data becomes obsolete quickly, then all other things being equal, it will be easier for a rival to enter the market, because it doesn’t need to match the incumbent’s years of learning from data. —> ex. With casual social games for computers and mobile devices, however, the value of learning from user data tends to decrease quickly.
- Is the data proprietary—meaning it can’t be purchased from other sources, easily copied, or reverse-engineered? Having unique customer data with few or no substitutes is critical to creating a defensible barrier
ex. Speech-Speech-recognition software: Historically, users needed to train the software to understand their individual voices and speech patterns, and the more a person used it, the more accurate it became. However, he past decade has seen rapid improvements in speaker- independent speech-recognition systems, which can be trained on publicly available sets of speech data and take minimal or no time to learn to under- stand a new speaker’s voice. - How hard is to imitate product improvements that are based on customer data? It is difficult to build a durable competitive advantage if the resulting enhancements can be copied by competitors without similar data.
Factors that affect companies’ ability to overcome the challenge:
* Whether the improvements are hidden or deeply embedded in a complex production process, making them hard to replicate
* How quickly the insights from customer data change. The more rapidly they do so, the harder they are for others to imitate. ex. Google maps - Does the data from one user help improve the product for the same user or for others?
* When data from one user improves the product for that person, the firm can individually customise it, creating switching costs —> provide a barrier to entry by making existing customers very sticky
* When data from one user improves the product for other users, this can—but may not— create network effects —> provide a barrier to entry by providing a key advantage in competing for new customers - How fast can the insights from user data be incorporated into products? Rapid learning cycles make it hard for competitors to catch up, especially if multiple product-improvement cycles occur during the average customer’s contract. But when it takes years or successive product generations to make enhancements based on the data, competitors have more of a chance to innovate in the interim and start collecting their own user data.
Cold start/ Chicken-egg problem
Businesses aiming to build regular network effects need to attract some minimum number of users to get the effects started, and those aiming to achieve data-enabled network effects need some initial amount of data to start the virtuous cycle of learning.
The AI factory
- AI factory treats decision-making as a science. Analytics systematically convert internal and external data into predictions, insights, and choices, which in turn guide and automate operational workflow
- With weak AI, the AI factory can already take on a range of critical decisions. In some cases it might manage information businesses (such as Google and Facebook). In other cases
it will guide how the company builds, delivers, or operates actual physical products (like Amazon’s warehouse robots or,Google’s self-driving car service). But in all cases digital decision factories handle some of the most critical processes and operating decisions - Components essentials in every factory:
1. Data pipeline= process that gathers, cleans, integrates, and safeguards data in a systematic, sustainable, and scalable was
2. Algorithms= generate predictions about future states or actions of the business
3. Experimentation platform= hypotheses regarding new algorithms are tested to ensure that their suggestions are having the intended effect
4. Infrastructure= systems that embed this process in software and connect it to internal and external user
Removing limits to scale, scope, and learning with AI
- AI-driven processes can be scaled up much more rapidly than traditional processes can, allow for much greater scope because they can easily be connected with other digitised businesses, and create incredibly powerful opportunities for learning and improvement—like the ability to produce ever more accurate and sophisticated customer-behavior models and then tailor services accordingly.
- Collision= when AI-driven firm competes with a traditional firm by serving the same customers with a similar (or better) value proposition and a much more scalable operating model —> As both learning and network effects amplify volume’s impact on value creation, firms built on a digital core can overwhelm traditional organization
- Collisions are not caused by a particular innovation in a technology or a business model.
They’re the result of the emergence of a completely different kind of firm. - Collisions firms suffer from critical mass and cold start problems
- Ex. Amazon collides with traditional retailers and Uber with traditional taxis services
Rethinking strategy and capabilities with AI
- As AI-powered firms collide with traditional businesses, competitive advantage is increasingly defined by the ability to shape and control digital networks —> Organizations that excel at
connecting businesses, aggregating the data that flows among them, and extracting its value through analytics and AI will have the upper hand
* Organizations that can’t leverage customers and data across those boundaries are likely to be at a big disadvantage. Instead of focusing on industry analysis and on the management of companies’ internal resources, strategy needs to focus on the connections firms create across industries and the flow of data through the networks the firms use.
* However, this will also cause dislocation in many occupations —> not include job replacement but also the erosion of traditional capabilities. In an AI-driven world, the requirements for competition have less to do with specialisation and more to do with a universal set of capabilities in data sourcing, processing, analytics, and algorithm development
Leadership challenge with AI
- The removal of operating constraints isn’t always a good thing. Frictionless systems are prone to instability and hard to stop once they’re in motion
- Digital operating models can aggregate harm along with value —> A mistake can expose a large digital network to a destructive cyberattack
Dark aI scenarios and malevolent AI applications
- A small number of tech giants could monopolise the AI economy
- Algorithms can manipulate our buying and selling patterns and transactions
- AI can help widen the gap between leaders and laggards, the rich and the poor
- Micro bits of personal data could help concoct sophisticated deep fakes
Methods to combat AI
A. Accountability as a means to enforce data privacy standards: Although the present state of accountability in AI is far from perfect, the G7, G20 and OECD have served as leaders in the creation and deployment of such principles
B. Placing more emphasis on Explainable Artificial Intelligence (XAI): a movement known as Explainable Artificial Intelligence, which argues that autonomous systems should be able to explain their motives and actions.
C. Addressing the issue of bias: A number of reasons can influence algorithms to make unfair decisions, starting with the data itself. Data is compiled by humans, not AI. This is why it is important for data to be collected and processed by a team of diverse individuals, or the information fed to the algorithm will present skewed preferences —> implement fairness constraints
D. Enforcing ethical codes and standards
E. Not releasing on AI for all the answers: reliance on machine learning algorithms to supply us with the right answers can quickly lead to dangerous outcomes —> ex. In the case of war: having AI decide who to kill or when to engage an anti-enemy aircraft
