Artificial Intelligence for Boards

By Philipp Gerbert, Philipp Hartmann, Andreas Liebl, all appliedAI

Many top managers advocate Artificial Intelligence in Business – and then delegate the execution. We attempt to break this pattern, convincing boards to take charge and enabling them to do so. In this endeavor, we are grateful for the support of many progressive business leaders.

In Artificial Intelligence for Boards – Gearing up for the Future of Business we, jointly with our partners from Odgers Berndtson, describe why management boards should get actively involved in AI, what they need to know, and, most importantly, what are the core issues in each functional area.

Why dedicate your time?

As a manager you have a busy schedule - irrespective how large your organization. Being smart, you have understood long ago that Artificial Intelligence is important for business and put the topic on the strategic agenda of your team. When you checked again you possibly saw lots of activity, but little bottom-line impact. So where is the problem?

Have you ever considered that the problem could be you?

Yes, you successfully maneuvered previous technology waves: Mobile, Digital, and Big Data, to name just a few. You took them seriously and acted with determination, delegating the task to the unit in charge and holding it responsible for results. However, AI is a rather different animal, putting more pressure on you as the board to become educated and directly involved in a serious way.

How AI is different

There are four characteristics of AI, which in aggregate demand board engagement:

• AI is arguably the largest technological transformation business has seen. It affects all sectors and all functions. Beyond Tech, value pools in industries are shifting and some companies are already fighting for local or global dominance: Ant Financial in Banking, Netflix in Media, Nvidia or Uber/Didi/Yandex and more in Mobility, Ping An in Insurance, John Deere in agriculture to just name a few prominent examples.
• AI is not plug-and-play, so you cannot just ‘buy’ a pound of intelligence. Algorithms need to be trained on data, often your company data. Also, you are never quite ‘done’ with AI – so even buying expertise can only take you so far.
• Scaling AI is a headache. As ‘data writes software’, and there is ever more data, you are never quite done. Scaling AI often amounts to automating a workflow from developer to user with inadequately standardized systems support. Revamping the enabling infrastructure, also including governance, talent, cultures, data and ecosystems, cannot be done by isolated teams
• Finally, entirely new risks may arise wherever you do succeed in implementing AI. AI acts at unprecedented speed and scale, and every outcome is measured and documented. Mistakes and biases can multiply and blow up, with severe reputational or financial risks.

Summing it up: When the potential business impact is huge, you cannot just buy it, scaling requires a holistic approach and unattended risks can balloon, the board might want to take charge of AI.

What could help?

An avalanche of publications focuses on Digital&AI transformations and change, or on describing specific AI use cases. We take a more comprehensive approach to the future business environment, targeted at board functions:

• First the report describes what every manager should know about AI
• Then the report addresses individual board functions holistically. In order to do this, it assumes a functionally structured board but is easily transferred to other form of organizations – after all, every functions must be reflected somewhere.

 Let us illustrate what me mean, picking the CFO and CTO as examples:

Artificial Intelligence for CFOs (Finance):

• Of course there are numerous powerful applications of AI in Finance. Forecasting and planning are typical examples. The CFO needs to leverage those to increase performance (speed, accuracy, flexibility) and lower costs.
• However, the CFO also plays an important role in budgeting AI. This requires a different set of criteria from traditional IT invests. Given the speed of development and some level of uncertainty in outcomes, the approach to budgeting AI initiatives is more reminiscent of financing innovation.
• Also, CFOs are in charge of managing company risks. Here AI can help. For instance, it is a great tool for detecting outliers, pointing to financial or compliance risks. However, AI can also introduce entirely new risks. CFOs need to monitor those and trigger their mitigation.

Artificial Intelligence for CTOs (Technology):

• First and foremost, the CTO is responsible for the development of new products, incorporating AI. This puts a huge burden on understanding the potential of AI, talent management and managing the life-cycle of an AI-based product, including performance and safety assurance.
• At the same time, the CTO needs to understand and prepare for the business implications of AI in the offering. For instance, predictive maintenance creates large customer value. But, naively implemented, it can destroy the service profit pool. Thus, CTOs have to concurrently prepare companies technologically for advanced service offerings on top of maintenance.
• The incorporation of AI often requires developing new suppliers or forging new partnerships for a competitive offering. This new ecosystem in a partially unfamiliar domain needs to be created and managed professionally.
• Finally, the CTO is also responsible for the R&D process. There are numerous AI-based innovations, such as generative approaches to additive manufacturing or developing and testing on a digital twin.

The above examples provide a teaser for the discussion of two roles, but it should give a flavor of what the report aims to discuss in detail. Therein we cover the functions of CEO, CFO (Finance), CIO (IT), CTO (Technology), COO (Operations), CHRO (People), CMO/CSO (Marketing and Sales), discussing AI-triggered challenges, as well as typical problems/misconceptions and how to address them. In addition, we review under what circumstances an additional role of a Chief Data and AI Officer (CAIDO) might make sense. The discussion of the management functions should be equally rewarding for supervisory boards, as well as for managers at large. 

Enjoy the read! 

Access full report: Artificial Intelligence for Board – Gearing up for the Future of Business

appliedAI is a non-profit initiative of UnternehmerTUM that strives to enable companies to successfully apply Artificial Intelligence for creating business value.