Is Artificial Intelligence Business Ready? - Part 1

Artificial intelligence is one of today’s magic terms. However, due to the lack of a clear business definition, it is accompanied by many misconceptions.

And, a confusing and inflating landscape of solutions does not make things better.

We talk about Augmented Intelligence

In this article, artificial intelligence is narrowed down to its huge potential to augment and improve human decision making.

Decision making affects the near and mid-term future. Leveraging data to answer predictive (“what will happen”) and prescriptive (“how can we make it happen”) questions has a huge impact to take the right decision at the right time.

Do we move the right track?

Currently, artificial intelligence is in its infancy and experimentation phase. Companies raise tens of millions of dollars to build demonstrators or run pilot projects to identify use cases that might have business value.

However, this situation will change soon: Referring to IDC, worldwide spending on artificial intelligence will reach $77.6 billion in 2022, more than double the $35.8 billion forecast for 2019. A strong argument that AI is on the ra-dar of more and more enterprises.

This shows that AI must now leave the play-ground and turn in-to a mature business tool.

Are today’s AI solutions prepared? Do they really focus enough on AI’s success fac-tors in business beyond demonstrators and pilot projects?

A Simple Example

The current approach to answer predictive & prescriptive questions can be sketched as follows: Huge amounts of historical data are used to train a data model to answer a specific question with high accuracy.

Is this ATM transaction fraudulent? What is the demand for electric power within the next 24 hours? How does my best buyer look like? What is the best price?

A simple observation: A certain data model for a certain question. More questions require more data models.

Modern data-driven enterprises constantly increase their amount of data to cover more and more dimensions of their business environment to improve their decision making with more and more diversified data models.

Companies with a more or less static repertoire of data models will always get the same answers independent of their ever-changing business environment.

We are not convinced that such an approach is appropriate in our big data era.

Are we prepared? 

Agile AI Production Lines?

In a data-driven company, data models will become as numerous as word documents are today. It is obvious that we need agile AI production lines at our finger-tips to rapidly ge-ne-rate and use data models on demand:

Accepting that data scientists leverage technologies that are completely different from an enterprise’ production environment may have small impact on demonstrators and pilot projects.

For companies who constantly have to respond to their stakeholder needs, however, a cons-tant reimplementation of their data models runs the risk to take the wrong decision at the wrong time.

Continuous Integration?

When we demand for agile AI production lines, we also have to take into account that data is never at rest:

Existing algorithms that are at the heart of current data models have to be complemented by future algorithms and made accessible as components of our production lines as well.

The good news is that there is enough innovation out there and new algorithms will cons-tant-ly be generated. Therefore, we can shift our focus to their rapid integration.

Structured Business Process?

Artificial intelligence is no end in itself and must not be locked up in laboratories. It has to be integrated into every business workflow. And those, who are in charge to take action and decide, have to understand every machine generated result.

Without a common business process that covers the full range from identifying a business problem and its drivers to generating data models and retrieving answers, no AI solution will survive in business.

What comes next?

This article marks the beginning of a series of articles where we make clear, why it is time to rethink current AI solutions to get them ready for business.  

For those, who are interested to read more now, check out our website.