AI strategy and ChatGPT with Christopher König

"ChatGPT is a complete game changer."

Chatbots are pure passion for Christopher. He sees artificial intelligence as a must-have for companies in all sectors that should not be missing from any strategic planning. A conversation with the AI consultant about long-term success with AI.

Christopher, have you spoken to Siri or Alexa today?

No, I don't like them at all and don't use them for data protection reasons.

Which AI technologies do you rely on in your everyday life?

ChatGPT. I haven't actually had to work at all since the launch.

Like all of us here.

Joking aside. Artificial intelligence is already used in many things; Spotify, for example, is already well informed about me and designs my playlist. We first have to ask ourselves how we actually define AI. Machine learning is already AI, i.e. all self-learning systems.

Artificial intelligence has been a hot topic since the introduction of ChatGPT last year. How long have
have you been dealing with this topic?

My background is in economics, I did my master's degree in quantitative economics and dealt with innovation economics, among other things. There was a lot of AI involved and I dealt with questions about what is possible with the new machine learning models. My semester abroad in Berkeley certainly had an impact on me. At that time, there was no getting around the topic in America.

When was that?

I learned a lot about the mindset of the researchers there, who incidentally invented Apache Spark and later founded Databricks. When I was there, we examined the impact of technologies and AI on the labor market and opportunities for automation, among other things.

Was ChatGPT already an issue?

Language models weren't that good at the time. I would say about a year and a half before the ChatGPT launch, they were already more solid to use. Nevertheless, topics such as text recognition, speech recognition, recommendation systems and automated decision-making systems had long been a topic in America. The broad mass of German companies, which do not deal with digital topics by nature, did not yet have this on their radar. Data management topics were also barely present.

"With ChatGPT you can generate added value for yourself as a user."

What exactly fascinates you about AI?

It is very challenging, both technically and in terms of content. As an AI consultant, I have to understand every use case
and think my way into a specialist department, a business model or an entrepreneurial problem. I am also constantly getting to know new technologies and thinking about where I can make which adjustments to achieve good quality in the end. Every case is new and promotes flexible thinking.

In which areas or sectors do you think AI is particularly popular at the moment?

Especially the emergence of ChatGPT, and thus Retrieval Augmented Generation, has made AI accessible for many areas. Companies are combining their data with ChatGPT to inform. This is ideal for those who are not yet that far along in their data initiative. RAG is useful for anything to do with language. If there are areas in the company where people talk a lot, ChatGPT can provide support and take over tasks completely.

Another area is knowledge management, in particular making knowledge available. RAG systems are able to answer application-related questions relatively easily by accessing the available source systems. This is interesting for all industries. Every company has knowledge, some of which is documented and some of which is handled in a very diffuse manner.

Are companies asking for AI because it's hip right now?

Yes, AI becomes particularly tangible with ChatGPT. With ChatGPT, you can generate added value for yourself as a user. AI can also make predictions and automate decisions. This also includes machine learning models that are closer to traditional statistics. This gives companies access to the various AI topics.

What ideas do they have?

This varies greatly. Some companies look at the technologies in detail and come to us with a specific use case of what exactly they want to automate. But there are also those who say: "Hey, we've noticed that a lot is happening around us at the moment, what potential do you see for us to use AI?"

You then develop an AI strategy?

We look at the company's business model and processes. In doing so, we keep in mind where AI can be used to establish new processes or reduce costs. AI can also be a revenue-boosting tool, such as a simplified product search in the online store or improved customer service in terms of availability and low thresholds.

"AI can also be a revenue-boosting tool."

To what extent does your company have the ability to deal with the topic of AI confidently?

We introduce the company to the topic of AI, not theoretically, but in a very striking way. We like to bring examples with us. And then we ask a lot of questions: What are your challenges? What are your goals? What are your resources like? For example, a company may already have to cope with strong sales growth, so self-learning AI can help it make better predictions with a new model and new data. We have a number of use cases for the requirements engineering process.

Chatbots are currently particularly popular for use in customer service. How do companies implement them profitably?

Automating processes is an essential process that requires a lot of resources. It makes implementation easier if the technological and content-related setting is in place. You should also consider whether it needs to be a fully automated system or whether a supporting system, such as Co-Pilot from Microsoft, might be sufficient. It is difficult when a company commits to a use case and then realizes during implementation that it doesn't need it at all. Investing in AI is always worthwhile if it makes processes better or more efficient - which of course has to be reflected in costs or turnover. AI can be particularly useful for resource-intensive processes.

What are your experiences with chatbots as a user?

I've never really been a fan of chatbots, I have to be honest. Before ChatGPT, a chatbot had never helped me as a customer. That has now turned 180 degrees. I would prefer to work only with chatbots, rather than with prediction methods.

"I would prefer to only work with chatbots."

Whether chatbot or other automation: how do I choose the right use case?

It is better to choose a use case that has a relevant benefit even if the outcome is uncertain. But you can't say that in general, it always depends on the company.

Is there a typical project process?

That also varies again. Companies rarely come to us with perfect data. Data preparation in machine learning projects is often a large part of the work. In ChatGPT projects, on the other hand, it doesn't take up as much space. The way we interact with customers is very different. We generally work according to the CRISP-DM framework, a standard for data mining projects. Otherwise, we act in a knowledge-driven manner and then adapt the requirements again and again.

How long does an AI project usually take?

We complete an MVP within around three months. It takes us three to four days for an initial mock-up, on the basis of which a decision can be made on how to proceed. After the MVP comes the lengthy part. In productive operation, we protect ourselves against all eventualities and take ethical and legal measures. This can take another four to five months. During this usage phase, we collect a lot of feedback from users and go into refinement at the same time.

User feedback is fundamental here, isn't it?

Total. Here you can also see the difference between prediction and chatbot projects. With machine learning algorithms, we can optimize the model specifically for a KPI or metric. This is not possible with chatbots because ChatGPT cannot always deliver the same answers. It is much more difficult to measure how good each answer is. We are always dependent on user feedback to know how good the answer is.

How do you document this feedback?

As we are using LLM GPT-4 and not the OpenAI web app, we are developing a front end ourselves and building a feedback mechanism into it. It is similar to the feature that already exists in ChatGPT: thumbs up or thumbs down. Our development cycles are longer because we collect feedback. Faster tests would be feasible, for example by generating synthetic test questions and answers, keyword groundedness evaluation. But our approach means that we focus directly on the human factor and the business case. In this way, we avoid bypassing the problem when optimizing.

"We no longer just build the model logic; the model learns for itself what the best results are."

What else is needed to be successful with AI?

Certainly a certain openness to technology. Throughout the entire company. Change management is an important topic and the relevant departments should be involved, also to allay fears. In the short term, the integration of AI can cause friction, but it is worth persevering. It makes sense to find solutions and processes for abrupt changes in order to get employees on board.

What data should be available?

Data availability and data quality are important issues. However, the data does not have to be highly structured; unstructured data from SharePoint, PDF files or website content, for example, can also be processed. But there has to be some data. It then needs to be cleaned.

How should companies be strategically and technologically positioned in the coming years in order to be data-driven and AI-supported in a competitive future?

What I would really advise companies to do is to pick a use case and say: "We're just going to do it!" They should launch the whole thing as a secure test balloon that can also have a certain impact internally. So that the entire workforce is willing to work with such technologies. ChatGPT is a huge thing, it really is, and it should be used now.

Just start doing it? After all, these are investments that many people have to plan for.

Sure, but that's a strategic question. Most companies are so caught up in their day-to-day business that they don't deal with this kind of technology enough and don't see the potential. It is certainly a topic for the C-level. You have to sit down and think in a structured way about how AI can be integrated. ChatGPT is a complete game changer, comparable to the introduction of the PC. How do we deal with this in the long term? One or two projects are not enough, you have to think long-term. It's a technology that can turn entire industries upside down.

Can small and medium-sized enterprises benefit from AI just as much as large companies?

You can also build a solution with relatively little effort that quickly delivers added value. AI is an exciting topic, especially for companies that cannot afford many employees. The development costs are quickly amortized.

Which AI technology would you like to develop further for yourself?

I would like to develop my next chatbot with LangChain, an open source framework designed to build applications with LLMs. That could be exciting.

This article first appeared in a similar form in issue 01/24 of data! All issues and articles of our biannual magazine can be found here:‍

data! Magazine: Cloud Services, Data Analytics & AI | taod

‍