Matching Human and Machine Intelligence

Management needs to look beyond data and algorithms to explore how artificial intelligence is transforming the essence of management

In the rush to prepare management for artificial intelligence, both Academia and CorpEd are spinning together programs of statistics, information technology, and information sciences to train future professionals for commerce and industry. Lost in the shuffle is the need to address both how AI is transforming management practices today, and how it will impact careers in management in the foreseeable future. Let's take a moment to underline the obvious- most managers will never become data scientists, but all managers need to look beyond the data and the algorithms to explore how artificial intelligence is transforming the essence of management.

What do business and engineering students need to understand about AI in Management? In the first contribution of this three-part series, we examined how AI influences the foundations of corporate strategy, the need to address the concept of "AI readiness", and the specificities of AI project management.[i] In this contribution, we will look at three additional foundations of AI in a business setting: the importance of fostering Collaborative Intelligence, the need to improving Managerial Decision-Making, and the implications of Digital Transformation.

Collaborative Intelligence

The future of business does not belong to chatbots, cybots and cobots, but to managers that have learned to adapt organizations to leverage new forms of human-machine collaboration. Collaborative intelligence refers to the design of networks of human and machine agents that play to their mutual strengths in fostering the crowdsourcing of knowledge, expertise, and preferences. Management alone can define the division of labor between human and machine agents and establish the rules that govern acceptable behavior. As algorithms demonstrate their value in automating repetitive tasks, managers prove their worth in understanding how and where to match human and machine intelligence. Management will need to encourage both communication and meaning in discussions that will be increasingly both asynchronous and virtual. Last but certainly not least, managers will be held accountable for resolving the inevitable conflicts that will arise when algorithms to deal with the complexity of human interaction.

If the goal of artificial intelligence is to enhance human potential, it's management's responsibility to discern what it means to be human. Training management to nurture collaborative intelligence can focus on how to operationalize an "augmentation strategy" efficiently and effectively.[ii] Case studies on the theme can explore the challenges of leveraging the network effects of human and machine agents in dotted-line organizations where managers and employees act as independent contractors. Workshops can focus on the practice of change management and challenges of designing patterns of transformation that regulate the pace of change. Far from the mechanics of data and algorithms, management must be prepared to encourage their ecosystems to leverage AI to co-develop human and machine intelligence.

Machine agents will never be any better than the computer program behind them

Automated decision-making provides a case in point. On the upside, algorithms promise us more efficient processes and more consistent decisions that will allow managers more time to focus on improving the business. On the downside, machine agents will never be any better than the program behind them. In cases where the agents rely on deep learning algorithms integrating artificial neural networks or support vector machines, the how and the why of each decision can prove very difficult to explain. Their output will be continuously challenged as being unfair, illegal, or erroneous. As a result, regulations like Europe's GDPR require companies to provide specific justification for each algorithmic-based decision affecting employment, health, purchasing or credit. Human and machine agents will need to develop collaborative intelligence in communicating both with each other and with consumers.

Managerial Decision-Making

Managers don't act on data, but on their perceptions of the what the data represents. Machine learning is about developing algorithms capable of learning from data, elucidating patterns and connections, and providing insights in supporting human decision-making. Yet in decision science, we learn that the major challenges to decision-making are human perceptions of the complexity, ambiguity, and uncertainty of the environment in which we take decisions. In the cognitive sciences, we explore how our pre-conceptions and prejudices distort how we see the problem and limit our ability to propose innovative solutions. Studying artificial intelligence in isolation is as short-sighted as it is useless, for practitioners need to understand the complexity and the diversity of human decision-making.

Programs on managerial decision-making can focus on where and when machine learning can help managers transform data into action. Coursework can help participants understand the role of modeling, the nature of specific decision environments, and the impact of cognitive biases in the data, algorithms, and logic. Case studies can elucidate the place of heuristics and algorithms in perception, prediction evaluation, and insight. As a whole, there programs can be designed to develop the participants' ability to "solve" business problems through the analytic method: evaluating the context, identifying the roots of the problem, evaluating the quality of the data, choosing the right methodology to address the problem, and creating the conditions for collective action.

As long as we purchase and consume, the essence of marketing will never be captured in an algorithm

B2C Marketing provides multiple examples of the complexity of managerial decision-making. AI modeling and simulation techniques can certainly be leveraged to construct relevant buyer personas and to predict consumer behavior trends.[iii] Yet as long as human beings purchase and consume, the essence of marketing will never be captured in an algorithm. Each marketing decision requires understanding customers' needs, appreciating the temporal impact of the of community, culture and economy on consumption, and aligning each organization's product and service offer. Decision environments, cognitive biases, and consumer preferences inherently influence how individuals and organizations view the data at hand. B2C Marketing isn't a question of data and outcomes, but of the processes of perception, prediction, evaluation, and insight that influence consumption and production. [iv]

Digital Transformation

The term digital transformation suggests that digital technology provokes fundamental changes in how businesses deliver value to their stakeholders. Although the concept originally reflected a vision of transforming « atoms to bits » it has evolved over the last decade to encompass a mindset about how managers and their stockholders view their businesses. Digital transformation is not about hardware, software and data, but rather technology's role in providing management insight into what's core to the business, and how their organization works with customers. As importantly, experience demonstrates that digital technologies don't change companies and markets, people do.

The current objectives of digital transformation: the use of AI and machine learning, the adoption of digital operating models, the creation of new digital partnerships, and the evaluation of a company's digital assets, underline the importance of management training in any digital initiative. Coursework can help participants understand that digital transformation isn't an « IT project » any more than the responsibility for its success lies with the Data Science or IT teams. Case studies can provide ample room for debate and discussion around the role of management in setting an appropriate vision for each project, as well as pertinent methodologies and evaluation metrics in specific contexts. Workshops can provide the necessary room to explore the need to foster cultures of risk taking and overcome organizational barriers to change.

If these companies view data as nothing more than a reflection of the reality of their market, they see data-driven decision-making as nothing less than their vision of how they deliver value to the customers

Platform business models like Airbnb, Uber or Amazon provide ample food for thought. Although each make extensive use of artificial intelligence, the value of each case is neither in the data they capture nor in the algorithms they use. They view data as nothing more than a reflection of the reality of their market, they see data-driven decision-making as nothing less than their vision of how they deliver value to the customers. [v] When analyzing the impact of digital transformation inside these companies we need to look beyond their digital infrastructure to explore how their data practices support their business models, business processes, and revenue streams. The value of their business can not be calculated as the sum of their material and immaterial resources, but in how their vision of digital transformation has encouraged both their customers and employees to think differently about doing business.

What are the lessons learned?

Training in Artificial Intelligence today needs to help managers look beyond data and algorithms and explore how AI has transformed organizations and markets. The future success of business endeavors doesn't depend on AI, but managers capable of leveraging digital technologies to enhance human potential. In this contribution we have addressed three themes that should be at the core of any Data Science program: Collaborative Intelligence, Managerial Decision-Making, and Digital Transformation. In each vision, management skills rather than just technical savvy prove fundamental in aligning the promise of AI with each organization's strategy, resources, and culture. In the final contribution in this series on the role of management in AI we will look at the importance of Digital Ethics and AI's influence on innovation.

Lee Schlenker is the Principal in the BAI http://baieurope.com and Professor of Business Analytics and Community Management. His LinkedIn profile can be viewed at www.linkedin.com/in/leeschlenker. You can follow the BAI on Twitter at https://twitter.com/DSign4Analytics

- - - 

[i] Schlenker, L. (2020), The Place of Management in an AI Curriculum, Medium [ii] Wilson, H.J. and Daugherty P.R., (2018), Collaborative Intelligence: Humans and AI Are Joining Forces, HBR [iii] Tjepkema, L. (2018), Using AI for Marketing: How Machines Optimize Decision-Making, Emarsys [iv] Schlenker, L. (2019), Putting AI First, Towards Data Science [v] Feng, L. et al. (2017), How Airbnb Democratizes Data Science With Data University, Toward Data Science