In discussions I’ve had with CIOs via my weekly #CIOChat sessions this year, the top 5 priorities are:
And while there are differing opinions regarding the ownership of the analytics function, one thing is clear: CIOs need a better understanding regarding the potential for analytics and what is required to get data into a shape for their organization’s data scientists. CIOs also need very clear mutual direction established with business leaders – in other words, what questions should be answered with data?
Against this backdrop, “AI, Analytics, and New Machine Age” – published by Harvard Business Review earlier this month – is a timely, relevant compendium of HBR articles. The authors’ insights should have value for CIOs and business people trying to use analytics in the running their businesses.
Artificial intelligence for the real world (Tom Davenport)
Davenport contrasts the results obtained from large AI projects versus Robotic Process Automation (RPA). He is candid that ambitious moon-shot projects are less likely to be successful than low hanging fruit projects focused on enhancing business processes.
Davenport suggests that CIOs and business leaders need to look at AI through the lens of business capabilities versus the lens of technology. He says that process automation focuses upon the automation of digital and physical tasks using RPA. More importantly, Davenport says RPA is the least expensive and easiest to implement. This is an efficiency, (coupled with consistency and standard) time saving and integral part of any digital process.
Davenport contrasts RPA with “cognitive insight” and “cognitive engagement”. Cognitive insight uses algorithms to detect patterns against data sets with vast volumes and with this, interprets their meaning. Insights here are typically provided by machine learning. These supervised or unsupervised approaches are data intensive and detailed. Models are typically trained on a portion of a data set.
Models gets better, in other words, make predictions as they use new data. Clearly, the more data the better especially for things like facial recognition. Cognitive Insights often use a version of machine learning, deep learning which attempts to mimic the human brain to recognize patterns.
Cognitive engagement, in contrast to cognitive insight, uses natural language processing—chatbots, intelligent agents, and machine learning. These are typically deployed for customer service. Davenport suggest that CIOs need to understand each technology approach:
- Clearly define the business requirements/direction of travel
- Understand the tasks and strengths of each approach
- Create a portfolio of projects using more than one approach
- Launch pilots
- Use data from projects to implement business process redesign
- Work to scale up organizational analytical capabilities
Stich Fix on selling personal style to the mass market (Katrina Lake)
I personally found the Stich Fix story scary for the legacy enterprise that I know personally. Stich Fix is a fitting example of how start-ups using analytics can best existing market player. Stich Fix uses data science to deliver personalization at scale and transcend traditional brick and mortar retailers’ efforts at customer experience.
Its CEO say that it’s their goal to differentiate themselves through personalization. Its strategic choices, for this reason, are based on information from millions of customers. Its CEO says 100% of their revenue comes from analytical recommendations. Even more scary, she says data science isn’t woven into our culture; it is their culture.
Stich Fix attempts to deeply understand each item that it inventories and turn this knowledge into structured data. Because it uses data to better than legacy competitors, it understands what people want. This enables it to turn over inventory faster than conventional retailers and to buy the right things and even better get them to the right people. For this reason, it can sell inventory fast enough to pay vendors from client payments.
The Stick Fix CEO says its revenue is totally dependent on the recommendations that their algorithms deliver. To make this work, the company tracks 30-100 measurements per garment. It knows what portion of the population fits a 27-inch inseam and stock according to that proportion. In sum, it says that it’s put data science at its core rather trying to transform a traditional retailer—this is the basis for its advantage in the digital age.
Algorithms need managers, too (Michael Luca)
Luca suggested something interesting in his piece. He said that computer algorithms that perform tasks step by step create risks on their own and can even lead decision makers astray. He says that they are extremely literal. Luca says that they are black boxes—they can predict the future with great accuracy but cannot tell cause and event or why. This is similar to Stich Fix’s CEO who deploys humans to work with the data that algorithms produce. In that piece, Katrina Lake says, “we make unique and personal selections by combining data and machine learning with expert human judgement.”
Marketing in the age of Alexa (Niraj Dawar)
Dawar claims AI assistants will transform how companies connect with their customers. Dawar believes personal assistants will become the primary channel through which people get information, goods, and services. For this reason, he asserts that marketing will turn into a battle for the attention of the personal assistant.
Dawar claims this will occur because of the unprecedented convenience personal assistants provide. At the same time, he is candid that they will provide lower costs but also risks to consumers. These facts will change the game for brands and importantly, they will impact retailers, altering the relative power of players in the value chain and changing the basis for competition.
Dawar is clear that only a handful of general-purpose AI platforms will be left standing as competition heats up. But with this said, he says marketers who are today obsessed on creating an omnichannel customer experience will fade as AI platforms become a powerful marketing medium, sales and distribution channel, and fulfillment and service center—all rolled into one. Scary!
The fact is personal assistants can become a repository of reams of data about our habits, preferences, and consumption. As such, they will have a lot of influence over prices and promotions and the consumer relationship.
Given this, Dawar see the move being from trusted brands to trusted AI assistants. In sum, he believes that AI platforms will eventually be able to predict what combination of features, price, and performance with appeal to a given consumer at a given moment. As such, the issue for personal assistant providers will be ensuring accuracy, alignment, and privacy.
Why every org needs an augmented reality strategy (Michael Porter)
Porter says that there is a disconnect between data and the physical world today. He claims that augmented reality which superimposes digital data and images onto the physical world can close the gap. Specifically, he says that it can transform how we learn, make decisions, and interact with the physical world.
At the same time, Porter suggests that this can change how enterprises service customers, train employees, design and create products, manage value chains, and even compete. He believes augmented reality information platforms will address the fact that most data are trapped in 2-D screens and pages. By changing how data is rendered, augmented reality provides, in his opinion, a way to deal with the volumes of information available to us. Porter discusses, for example, superimposing digital images and data on real objects. He believes this will change the way humans process information and most importantly deal with what he labels “cognitive load.”
The truth about blockchain (Marcon Iansiti)
Iansiti believes blockchain has the potential to transform contracts, transactions, and records –the defining structures of our economic, legal, and political systems. He says blockchain provides an open distributed ledger that can record transactions between two parties efficiently and in a verifiable and permanent way.
In blockchain, contracts are embedded in digital code and stored in transparent, shared databases where they are protected from deletion, tampering, and revision. One thing that I personally hated about project management systems, for example, is people changing project dates to look good. Permeance is valuable.
Iansiti says in blockchain every payment has a digital record and signature that can be identified, validated, stored, and shared. He says blockchain works by establishing the following: distributed databases, peer to peer transmission, transparency and pseudonymity or requiring proof of identity to others. At the same time, Marcon says block chain allows for computational logic to be programmed into a blockchain including algorithms and rules.
With this said, Iansiti asserts that TCP/IP took 30 years to take off. We should not expect less for a system that enables bilateral financial transactions. Marcon claims that blockchain has the potential to dramatically reduce costs of transactions
Collaborative intelligence: Humans and AI are joining forces (H. James Wilson)
Wilson says that AI is becoming good at many human jobs. And while it is improving quickly, he says, in contrast to many in the press, replacing human workers is not inevitable. It’s like the human computers in the movie Hidden Figures—mainframes did calculations faster, but smarts were required to program and maintain them.
Wilson says that AI will, nevertheless, radically alter how work gets done and who does it. He says technologies larger impact will be complementing and augmenting human capabilities, but not replacing them. Wilson goes on to say that those who focused on workforce reductions will only see short term gains—AI should be focused upon enhancing humans’ complementary strengths. James says the catch phrase should be augment and not replace.
Humans have three critical roles with machines:
- Training (machines must be taught how to perform the work that they are designed for)
- Explaining (human experts need to explain their results)
- Sustaining (humans must make sure they continually work and ensure that AI systems function properly, safely and responsibility)
Machines in turn will help humans amplify their cognitive strengths and interact with customers and employees. This will free humans to focus on higher level tasks.
In a recent discussion with the #CIOChat group, they were clear that analytics and data are key requirements for digital transformation. CIOs clearly need to understand the potential and limitations of each approach. They, also, need to know how these technologies work together tin business and across the business supply chain. Additionally, there is no point in having AI and robots if somewhere else in the work process is an analog process that will slow down the work and become a bottleneck.
A moon-shot approach clearly is not the answer for everyone. CIOs have a role in these projects just like other transformative projects to act as the universal translator of business requirements. They also must be able to move and communicate smoothly between the business and technology. This means being able to communicate the potential and gaps for AI, machine learning, and other disruptive technology approaches.