Artificial intelligence and business models

Sat 30 January 2021, Philippe Ropenga Digital assets Artificial intelligence

It has been seen in a former post that digital assets were not only a matter of data but also a matter of person. Pieces of property are part of an estate because a person values them.¹ An estate relies on a close connection between a person and a thing. This is what makes a house a home. This is also true for a business, a trust, or any asset, including digital ones. A person has to decide how to allocate resources such as time or money. Resources are allocated to develop value and can be a source of conflict between stakeholders.² Value is a core notion in management studies while lawyers have to deal with conflicts over rights to property. Let us see that law consolidates value instead of freezing it (1). Managers must be careful when they try to standardise a disruptive service (2).

1 Consolidating instead of freezing

Law is an efficient tool when it is used to consolidate a situation that tends to be frozen by managers who apply business models. There is a real difference between legal concepts and markets (1.1). It has to be taken into account by managers who practice in the field of artificial intelligence (1.2).

1.1 Legal concepts versus business models

Law and business models are used to create a situation that benefits someone. Managers have a business model in mind and may see contracts as a mean to enforce their business model, just as a car is a mean of transportation. A mechanical engineer does not conceive a car this way. There is nothing wrong in seeing a car as a mean of transportation. Anyone who buys or uses a car wishes to move more easily from a place to another. Let us say that your car breaks down and that you cannot go to work without it. Would you expect your mechanic to move you from home to work? No, you know that a car is a thing that needs to be fixed and that you have to figure out how to do stuff without it. Managers who deal with contracts may think that lawyers are here to blame someone and that a court will ultimately force this person to bear all the consequences of an event that has occurred since the contract has been signed. It makes sense since the manager has signed the contract to apply a business model to the company. Law is often seen as a tool that freezes a relationship. This however is not how law works.³ Law is best used to consolidate a situation in a changing context whereas a business model describes the elements that have to be used to succeed. Hence, a business model is a useful landmark while law is used to see what has changed. One can use law to freeze accounts or to establish that one has not contributed to a damage. One can also use law to improve something. When you are sick, you consult a healthcare practitioner to find relief not to undergo a postmortem examination! Spending time arguing about liability does not lead to a solution since a litigious issue is likely to change over time while parties are arguing. The dynamic aspect of law may not be taken into account by entrepreneurs.

Not taking into account the dynamic aspect of law can really be problematic when a business deals with artificial intelligence.

1.2 Artificial intelligence and law

An algorithm is a set of predefined instructions that is applied by a program to treat data. Training allows to notice and correct wrong results produced by the artificial intelligence following the algorithm. Let us imagine an artificial intelligence trained to recognise people. The artificial intelligence has no idea of what a face is but compares pictures. The artificial intelligence does not exactly know what your picture will look like when you will be in front of the camera that will take a picture of you to analyse it and tell that it is you. The light will always be slightly different. This is why an artificial intelligence is trained in different conditions. It has to recognise you without any assistance. Does one really look like one's biometric photo, by the way? Biometric proportions are made for machines not for humans. A machine does not interpret your picture as a human does. The standardisation of passports facilitates the conception of machines and improves their efficiency. Contracts are also standardised today since lawyers have noticed that it helps to reduce risk by using clauses that have not been conceived to fit the client needs but to have a predefined effect. A lawyer combines these clauses to meet his client's needs when drafting a contract. Artificial intelligence is now able to analyse standardised agreements to notice small but significant changes and to suggest corrections like a spellchecker. A lawyer traditionally charges per hour whereas the artificial intelligence checks a document within seconds. A manager who applies a business model that requires to make a large quantity of one-shot deals may appreciate to know that artificial intelligence may help him to draft contracts within minutes. I have written on this blog that artificial intelligence could help lawyers to handle well-known issues.⁴ Formalities that can be fully standardised will be automated by artificial intelligence. Lawyers who are drafting highly standardised contracts will experience disruption in their practice. An algorithm is conceived to treat a given type of data. The treatment is thus only efficient when the data provided match characteristics that underpin the algorithm. Relying on standardised forms increases the chances of matching these characteristics. Before artificial intelligence was seen as a hot topic, lawyers were already relying on standardised documents to make sure that they match characteristics that were supposed to reduce risk. Hence, artificial intelligence increases disruption in legal practice but does not cause it by itself. The combination of legal standardisation and artificial intelligence may well ignore practical details that are important but will not be integrated in the solution because they are not easy to standardise and treat by artificial intelligence.

It has been seen that managers applied business models that provided a static view of a situation that might cause difficulties because issues that could not be easily standardised would not be taken into account by artificial intelligence. Let us see what these issues are.

2 Disruption and standardisation

Disruption and standardisation can be seen from the perspective of relationships or markets (2.1). A different perspective opens a different path to value creation (2.2).

2.1 Relationships versus markets

I have learned during my law studies that legal agreements bind two or more people. I think that this is still true. I also know that people who are trying to solve an issue often see it from opposite perspectives and that this divergence is a source of a conflict. People want to be listened to and understood. This is quite different from applying a business model to succeed. A business model is useful when everyone agrees about what to do. It however becomes useless when the application of a business model leads to conflict. It is useless because business models focus on entities, be it companies or people, that operate on a market. A disruptive business model designs its own market. This is at the core of the Chan and Mauborgne's Blue Ocean Strategy.⁵ A customer who would not have thought about purchasing a given product may buy something that requires similar knowledge to product but targets different customers. A smartphone is a computer that can make phone calls rather than a phone and has successfully found its place on the mobile phone market. Lawyers can tell that everything does not start with a market but with a relationship between two people or entities. When two business partners agree on a strategy, they focus on the market. Lawyers know that entrepreneurs can act on a market in a way that hinders competition.⁶ Furthermore, a business can grow successfully on a specific market while the relationship between partners deteriorates. Market and relationship are two different issues that both have to be taken into account. Unfortunately, artificial intelligence usually increases standardisation that in turn increases the effect of a business model that only pays attention to a market.

Let us see that focusing mainly on a market opens only one path to value.

2.2 Paths to value creation

Berle and Means have highlighted the distinction between ownership and control in a famous book first published in 1932.⁷ They have shown that the shareholder value was at the heart of relationships within corporations. A theory of strategic organisation that relies mainly on the market tends to overlook the human part of a business. This is not only a question of theory or ethics. This is essentially a matter of value and has practical consequences. Artificial intelligence can be used to optimise the workflow between people by completing repetitive tasks that involve data treatment. A company sends data to another that produces contracts and provides a tool to sign them electronically. The client who wants to get contracts quickly drafted than receives them ready for signature by another person. This is what artificial intelligence changes. Traditionally, lawyers who assist a client during a deal-making process play a secondary role and often remain backstage. The main part is played by the parties to the agreement that is being negotiated or signed. This conception of deal making does not seem outdated to me. When one compares it to automated contract drafting, it appears that the company that provides the drafting and signature service play the title role not the parties to the agreement since the contract has to be standardised not to secure the benefit of a foreseeable situation but to enable automated contract drafting. Self-driving cars raise a similar question: should artificial intelligence be adapted to automobiles or automobiles to artificial intelligence? Anyone can tell what makes a house a home, even if there may be as many answers as respondents. Questions raised by artificial intelligence are not new but are asked from a new technical and entrepreneurial perspective. No one knows what Berle and Means would have written if they could use artificial intelligence in 1932. Nevertheless, their analysis of the shareholder's interest draws more upon personal relationships than on business models that are market-centric.

It has been seen that law dealt mostly with relationships whereas business models dealt mostly with markets. Choosing a different paradigm opens a different path to value creation.

In brief, deep changes are often hard to notice.

Go Top