‘Business stakeholders can no longer avoid the topic of artificial intelligence’

Artificial intelligence and machine learning offer companies great potential to increase their efficiency, develop more innovative products and make better decisions – and thus gain a competitive advantage. Christof Zogg, Head of Business Transformation at Swisscom, explains in this interview why AI is important for companies’ success and concerns more than just their IT departments.

Christof Zogg, why should companies engage with the topic of AI?

Primarily due to the huge potential for improved efficiency enabled by the latest developments in artificial intelligence (AI) in general and machine learning (ML) in particular. For example, the ever-improving language translation programs, the chatbots for the automated handling of simple customer queries and the so-called transformer models such as ChatGPT offer impressive performance in text analysis and synthesis. AI is already widely used as a software development tool: according to Microsoft, 58% of the software code checked into GitHub these days has been reviewed, commented on or generated by the integrated Copilot bot.

How can AI support sustained company success?

Artificial intelligence has the unique potential to give companies a knowledge advantage in their core business and processes. They can then potentially make better decisions or hire more suitable employees.

What specific steps must companies take to use AI to the benefit of their business?

The age of artificial intelligence has only just begun. Companies now need to gain practical experience through specific, but not too broadly defined, use cases. At the same time, I think it’s important that the business side doesn’t just leave the topic of AI/ML to the IT department. AI is too important and its implications for all forms of knowledge work are too far-reaching for it to be neglected by business teams and management. AI and ML can be very important strategic tools – and strategy is ultimately a matter for management.

What problems can be solved with artificial intelligence today?

We are currently in the stage of what is known as narrow artificial intelligence, in which machine learning algorithms can perform specific mental tasks faster, better or cheaper than humans. In particular, these include applications in the fields of computer vision – such as autonomous driving, facial recognition and medical image processing – as well as natural language processing – such as email filters, assistance systems, translation services and text analysis.

Have Swiss companies and business stakeholders already recognised the importance of AI?

According to a recent survey by Equinix, 80% of Swiss companies surveyed would like to use artificial intelligence. However, the majority of them doubt that their data and IT infrastructure is ready for this. So there’s still a lot to do. With new AI services such as ChatGPT and Bard from Google, however, business stakeholders have no choice anymore but to engage intensively with the topic of AI. After all, smart employees in their companies have been using these in their day-to-day work for some time now and are feeding the underlying ML models with business-critical data.

How can companies prepare for the use of artificial intelligence?

It’s important to distinguish here between using an existing ML model that a provider has already trained for its customers – and one that is useful for a specific company, industry or country. Only with the latter specialisation can you establish a lasting USP and thus an effective competitive advantage. As with all data applications, whether you want to create a dashboard or train an ML model, you first have to do your homework and build a reliable and automated data platform.

Which companies and industries are benefiting the most from AI? Are there sectors that are still making relatively little use of this potential?

In contrast to other trending IT topics such as the blockchain and the metaverse, AI/ML is universally applicable. From this perspective, all sectors have a lot of room for improvement. If I had to mention sectors where the benefits are particularly high, I would say pharmaceuticals, finance, health and education.

How can companies find the potential and possible new use cases within their own operations? What preparations are needed for this?

In order to find useful and productive initial use cases, you should gain an overview of all areas of AI/ML application and then consider all of your company’s units, roles and processes through the lens of this grid.

Companies have to acquire the relevant data but also train the model and interpret the results. What are the biggest challenges?

Whether a specific use case can be successfully implemented depends essentially on two factors: do I have enough of the data needed and is it of the right quality? With a reasonable amount of effort, can I train a model to handle the desired classification task or make forecasts with sufficient accuracy and validity?

Where can business stakeholders find answers to these questions? Do they even need to know in detail how AI is used or can they obtain this expertise by, for example, employing data scientists or bringing on board external service providers such as Swisscom?

As with other key technologies – such as cloud computing – I believe that every successful business stakeholder in the future will need a solid understanding of AI/ML – if only to be able to ask the right questions and better gauge the benefits and costs. Whether this expertise is obtained internally by means of in-house data engineers and scientists or through external consultants depends on the company’s preference.

How broadly will artificial intelligence impact companies? Should it be seen as ‘just’ an extension of core business or does the entire organisation need to be aligned with it – essentially a transformation to a data-driven business?

As with all major technical disruptions, artificial intelligence will affect the entire organisation. In the end, it won’t be enough to simply establish a new team in some business unit. But companies have to take the first step somewhere on their journey into the age of artificial intelligence – and a dedicated team could be a possible starting point.

How can companies counter internal scepticism and rejection of AI?

So far, we’ve mainly talked about the promising new world of AI, but obviously this new technology category isn’t without its problems either. Quite the opposite: important tech visionaries like Elon Musk, Max Tegmark and Geoffrey Hinton are actually seriously worried about the future of mankind. Stakeholders in Swiss companies don’t have to go that far with their considerations. But as with most technology trends, there will be inflated expectations and costly exaggerations – all the more reason for stakeholders to really get to grips with the topic.