GDI: Ms. Janik, in media, ChatGPT is all over the place. Is that more because this time a new technology is affecting the media people themselves, or do we really witness a revolution in society?
Marianne Janik: The curiosity is huge, also because many people have been experimenting with it since November, publishing and sharing their experiences. Within 5 days, ChatGPT had one million users, and in January, two months after its launch, it already had 100 million. No online service has grown that rapidly before. The new AI language models write articles, essays, letters and speeches. That alone is a reason for the media to engage with the new possibilities. But the implications are much more far-reaching. AI, with all its capabilities, will find its way into our private and professional everyday lives.
Auch wenn man selbst gar nichts mit Essays oder Reden zu tun hat?
Even then. Artificial intelligence reduces the complexity of tasks, can analyze data in large quantities, in different formats and from many sources, and prepare it for its human partners, making it usable and understandable. We will be relieved of tedious routine work and will be able to complete tasks and processes more efficiently and thus make better use of scarce human resources. This will change many things!
Microsoft is fully committed to this technology: ChatGPT has been integrated into the Bing search engine, integration into your Office package will follow. What role will AI play at Microsoft?
At Microsoft, we have been developing AI solutions for many years. Already today, AI elements can be found in almost all of our services. Think, for example, of the Narrator read-aloud feature for on-screen content or the live captioning and translation feature for Microsoft Teams. Now we're integrating AI into solutions like Bing, Microsoft 365, Dynamics 365 and our security products. This will unleash a wave of productivity boosts from powerful copilots. What's more, just a few months after launching the new AI models, more than 1,000 Microsoft customers and partners worldwide are developing their own AI applications and services. They are doing it with AI models such as ChatGPT (text/speech), Codex (programming code), and DALL-E (images), delivered through our Azure OpenAI service. Generative AI can shorten development processes by weeks or even months. It can, for example, write programming code – even via natural language input. At a time when many companies and government agencies are desperately seeking developers and programmers, this is a huge opportunity.
And where do you see AI being used by the retail industry in the short and medium term? Should AI write our shopping lists? Organize inventory management? Coach the management? Negotiate prices?
In principle, all these questions can be answered with "yes". However, the assessment of what AI "should" do for us and what makes sense in practice is still made by retailers and customer acceptance – In short, by human intelligence. We currently see many use cases in omnichannel, at the point of sale and in customer service.
Zum Beispiel?
In customer service, a typical use case is hyper-personalization through timely summarization of customer inquiries, facilitated search, and automated communication suggestions. Chatbots also enable a whole new level of customer service thanks to generative AI. With copilots in the supply chain, costly problems are identified at an early stage. In Dynamics 365, for example, we offer an AI-powered messaging module that proactively alerts you to external changes in weather, finances, and geographic developments. At the same time, an AI copilot can prepare contextual emails in Outlook to take action and alert affected partners to mitigate potential disruptions before they occur. Another example is AI-generated product descriptions for online stores, where this information is missing by the millions.
How do you see the timeframes for an AI rollout in retail? Are we talking about decades, as with the introduction of scanner checkouts – or about weeks, as with the update of a software already in use?
Currently, over a thousand companies are testing copilots for classic application software in Microsoft 365, and there is already the GitHub copilot for AI-supported programming. Such services and functions are then immediately available via the cloud. Employees in marketing, for example, enter a query for data analysis to the marketing copilot in Dynamics 365 in their own words. They then immediately receive a comprehensibly prepared evaluation with further results for segment analysis and for preparing a campaign. This means an enormous increase in efficiency for marketing as well as for IT and data analysts. Of course, there are projects that require more time. But we should not deceive ourselves: The power of AI models is growing exponentially, and this will accelerate their use. Therefore, “just wait and see” would not be a good advice. Dealing with the possibilities of AI is an issue for the executive suite; it must not remain a special discipline of IT experts.
So wie wir heute über ChatGPT reden, haben wir letztes Jahr über das Metaverse geredet – und vor zwei Jahren über Blockchain. Worüber reden wir nächstes Jahr?
I hope we'll talk then especially about the personal experiences we have with our copilots in everyday life, but also about the consequences and successes for the respective corporate strategy. So not so much a talk about AI, but lots of concrete projects, experiences and business cases in dealing with it.
Learn more about the future of retail at the 73rd International Retail Summit, which will be held at GDI from 7 – 8 September 2023. On the theme "Rethink Retail: Be Bold, Stay Focused, Earn Trust", leading retail experts and researchers will explain why the ability to focus is becoming the most important retail skill. Sign up now!